Process Geomorphology

I
Process Geomorphology
Frontispiece, G. K. Gilbert
standing by perched granite
boulder in Yosemite National Park,
1908
Second Edition
Dale F. Ritter
Southern Illinois University at Carbondale
mull
Wm. C. Brown Publishers
Dubuque, Iowa
web lltlill
group
Wm. Brown
C.
Wm. C. Brown Publishers, College Division
Chairman of the Board Lawrence E. Cremer President

Mark C. Falb James Romig Vice-President, Product Development
L.
President and Chief Executive David A. Corona Vice-President. Production and Design
Officer E. F. Jogerst Vice-President. Cost Analyst
Bob McLaughlin National Sales Manager
Catherine M. Faduska Director of Marketing Services
Craig S. Marty Director of Marketing Research
Marilyn A. Pheips Manager of Design
Eugenia M. Collins Production Editorial Manager
Book Team
Edward G. Jaffe Executive Editor
Lynne M. Meyers Associate Editor
Nova A. Maack Associate Editor
Mark Elliot Christianson Design Supervisor
Vickie Blosch Production Editor
Mary M. Heller Photo Research Editor
Vicki Krug Permissions Editor
Cover photo: © Kathleen Norris Cooke
Copyright © 1978. 1986 by Wm. C. Brown Publishers. All rights reserved
Library of Congress Catalog Card Number: 85-70934
ISBN 0-697-05047-5
No part of this publication may be reproduced, stored in a
retrieval system, or transmitted, in any form or b> any means,
electronic, mechanical, photocopying, recording, or otherwise,
without the prior written permission of the publisher.
Printed in the United States of America

10 9 8 7 6 5 4 3 2
To my family.
Contents
Preface xi
1
Process Geomorphology — An Introduction 1
Introduction 2
The Basics of Process Geomorphology 6
Summary 31
Suggested Readings 31
Climate and Internal Forces 33

Introduction 34
The Endogenic Effect 34
Climatic Geomorphology 50
Summary 62
Chemical Weathering and Soils 63

Introduction 64
Decomposition 65
Soils 84
Summary 107
VII
1
Contents
Physical Weathering, Mass Movement, and Slopes 109

Introduction 110
Physical Weathering 110
Physical Properties of Unconsolidated Debris 118
Mass Movements of Slope Material 126
Slope Profiles 144
Summary 152
The Drainage Basin — Development, Morphometry,

and Hydrology 153
Introduction 154
Slope Hydrology and Runoff Generation 156
Channels and the Drainage Network
Initiation of 161
Basin Hydrology 176
Basin Denudation 191
Summary 204
Fluvial Processes 205

Introduction 206
The River Channel 206
Sediment inChannels 21
The Quasi-Equilibrium Condition 222
Channel Patterns 232
Rivers, Equilibrium, and Time 245
Summary 252
Contents
Fluvial Landforms 255

Introduction 256
Floodplains 256
Fluvial Terraces 267
Piedmont Environment: Fans and Pediments 275
Deltas 294
Summary 301
8
Wind Processes and Landforms 303
Introduction 304
The Resisting Environment 304
The Driving Force 308
Wind Erosion 310
Wind Transportation and Deposition 314
Deposits and Features 317
Summary 332
Glaciers and Glacial Mechanics 335

Introduction 336
Glacial Origins and Types 336
The Mass Balance 341
The Movement of Glaciers 343
Ice Structures 358
Summary 361
10
Glacial Erosion, Deposition, and Landforms 363
Introduction 364
Erosional Processes and Features 364
Deposits and Depositionai Features 379
Summary 403
Contents
11
Periglacial Processes and Landforms 405
Introduction 406
Permafrost 407
415
Periglacial Processes
Landforms 421
Periglacial
Environmental and Engineering Considerations 434
Summary 442
12
Karst — Processes and Landforms 445
Introduction 446
The Processes and Their Controls 449
Karst Hydrology and Drainage Characteristics 455
Landforms 462
Surficial
Limestone Caves 474
Summary 479
13
Coastal Zones — Processes and Landforms 481
Introduction 482
Coastal Processes 483
Beaches 498
Shoreline Configurations and Landforms 508
Erosional Landforms and Rates 513
Depositional Shorelines 522
Summary 529
Bibliography 531
Credits 569
Index 573
Preface
Geomorphology has undergone a drastic change in scope and philosophy during

the last several decades. In the past, the discipline was primarily concerned
with the evolutionary development of landscapes under a wide variety of cli-
matic and geologic controls. More recently, geomorphologists have recognized

the need for an applied rather than a historical emphasis. This change in phi-
losophy has placed geomorphology at an interface with many other disciplines.
Today's geomorphologist must relate to problems that face hydrologists, en-
and many other types of earth scientists. The
gineers, pedologists, foresters,
bond that unites geology and geomorphology with so many apparently diverse
disciplines is the common need to understand the processes operating within
the Earth's surficial systems. Thus, although the historical aspect of land-
scapes remains important, it is absolutely essential for earth scientists to have
a basic understanding of surface mechanics and, in addition, of how those
process mechanics are reflected in the landforms they create. This edition of
Process Geomorphology, like its predecessor, is an attempt to satisfy those
needs. The prime purpose of the book remains as it was, to provide under-
graduate students with an introductory understanding of process mechanics
and how process leads to the genesis of landforms.
A wealth of new information concerning
surficial processes has emerged
since the first was completed, and many new techniques to analyze
edition
process have been developed. In most chapters new data and interpretations
have been assimilated within the format of the first edition. Some chapters,
however, have been changed significantly. This is especially true in the intro-
ductory chapters (chapters 1 and 2) and in the treatment of the drainage basin
(chapter 5) and coastal processes (chapter 13). A lengthy bibliography is again
presented so that students wishing to pursue a particular topic in greater depth
will find a ready nucleus of source material. Most references cited were pur-
posely selected from journals and books that will most likely be found in li-
braries of North American colleges and universities. There is less mathematical

treatment in the revised edition, and such an approach is used primarily to
clarify concepts that are particularly complex.
Preface xii
I wish to acknowledge the help and guidance I received from numerous

colleagues in the geomorphological discipline. I am especially indebted to
Steven P. Esling, Ronald C. Flemal, Thomas W. Gardner, Andre K. Lehre.
R. Craig Kochel, Frances J. Hein and Arthur N. Palmer who reviewed parts
or all of the revised text. Their constructive advice and criticism were instru-
mental in the completion of the text, and their conscientious efforts are deeply
appreciated. Shortcomings and errors in the book are, of course, mine.
D. F. R.
An Introduction
1
I. Introduction C. Thresholds and Complex
II. The Basics of Process Response
Geomorphology D. The Principle of Process
A. The Delicate Balance Linkage
B. Force/Process/Resistance E. The Time Framework
l. Driving Forces III. Summary
a. Climate IV. Suggested Readings
b. Gravity
c. Internal Heat
2. The Resisting
Framework
a. Lithology
b. Structure
Chapter 1
Introduction One of the remarkable aspects of planet Earth is the infinite variety of its
surface forms. probably safe to assume that as humans became aware of

It is
their physical environment, landscape was the first geologic characteristic they
noted. Familiar surface features guided their travels and established their ter-
ritorial boundaries. As time passed, people learned how best to utilize regional
characteristics for different purposes, such as agriculture, trade, and military
adventure. They also learned that some landforms possess certain peculiarities
that somehow, almost imperceptibly, set them apart from others. Gradually
these isolated observations grew into an organized collection of knowledge,
and a separate branch of science was born.
Geomorphology is best and most simply define d as the study of landfor ms.
Like most simplistic definitions, this does not do justice to a discipline that
can be exciting to even the uninspired and challenging to anyone who enjoys
science. Historically, landforms have been analyzed in a variety of ways be-
cause different students seek from the landscape different information and
different truths. For example, since people live on landforms, geographers may
justifiably be concerned with how landscapes affect human events. Engineers,
on the other hand, examine surface forms to select the best construction sites
or to control the physical environment in the most advantageous manner. While

engineers and geographers may look at the same landscape, they probably
never ask the same questions about it.
Landform data come from widely divergent disciplines. Synthesizing the

facts into a cohesive picture of the Earth's surface, therefore, becomes a mon-
umental task. The diverse nature of the data may explain the appearance of
subdisciplines such as dynamic geomorphology or climatic geomorphology
geomorphology has always had in finding
(Biidel 1968) as well as the difficulties
a definite academic home. Today in the United States, geomorphology is still
taught in both geology and geography departments, and the subject matter
becomes the responsibility of anyone who will properly adopt it. The stepchild
existence between geology and geography has created in the minds of some
the undeserved image that geomorphology is not clearly defined as a science
or based on scientific facts.
It is true that traditional geomorphology has been excessively descriptive.
Much emphasis in the past was given to placing landforms, both regional and
local, into some evolutionary model, so that the field was concerned primarily
with historical interpretations. In recent years, however, the discipline has be-
come more quantitative, and research has shifted to studies with a more prac-
tical value. Modern geomorphologists often deal with problems that link them
directly to other professionals working at the Earth's surface. Obviously, geo-
morphology is more than any definition can adequately express. Although it
has identity, its boundaries are ill-defined and more certainly ephemeral.
More important than a precise definition is the fact that geomorphology
is and probably always will be a field-oriented science. Map and photo anal-
yses are necessary first steps to good geomorphic work, and laboratory data
3 Process Geomorphology —An Introduction
support interpretations. But the real test of geomorphic validity is outdoors,

where all the evidence must be pieced together into a lucid picture showing
why landforms are the way we find them and why they are located where they
are. A prime requisite for a geomorphologist is to be a careful observer of
relevant field relationships. This trait cannot be easily taught, and truly out-
standing geomorphologists usually develop it by learning from their own mis-
takes. Geomorphic processes are remarkably subtle, and minor changes of basic
controls can result in an infinite array of landforms. Invariably, the person
with the greatest experience under varied conditions willmake the most viable
geomorphic interpretations. Thus a geomorphologist, like any other scientist,
must learn the trade. There are no shortcuts that produce geomorphic insight.
It must be acquired gradually through long field experience.
This book will concentrate on processes that create the features we see at
the surface of the Earth. Process can be defined as the ac tion-pro duced wh en
a forr^ induces a change, either chemical or physical, in thernaterialsor forms
at the Earth_^ s_surface. In simpler terms, process may be thought of as the
method by which one thing is produced from something else. It may not be
clear why this approach is more beneficial than using some other criterion,
such as climate or time, as a central theme. As we have said, geomorphology
stands at the interface between geology and many other disciplines that deal
with surficial phenomena. Today geomorphologists must be aware of the prob-
lems facing hydrologists, civil engineers, pedologists, foresters, urban plan-
ners, and other specialists. And since those scientists are working in an
environment underlain and partly controlled by the geologic fabric, they must
be concerned with geologic concepts and problems. It follows that there must
be a common interest uniting these apparently diverse fields, since they all
function in the same place at the same time. It is the universal need to un-
derstand processes that is basic to all surficial disciplines.
Understanding what process is also serves as a basic component of other

For example, we now know that application of our knowl-
scientific disciplines.
edge about geomorphic processes is basic in the field of environmental science.

Every surface environment is controlled by process. We have known for years
that human intervention into surface environments causes rapid changes in
processes (Gilbert 1917) and invariably requires adjustments in the environ-
ment itself.
A
good example of human influence on geomorphic processes is occurring
today San Diego County, California (Kuhn and Shepard 1983). Here the
in
effects of cyclic climate change are beginning to produce accelerated erosion

of the bluffs overlooking the Pacific Ocean. In the past several decades, wave
action has not been severe because the prevailing dry climate during that in-
terval created very few major storms. Beaches, shorelines, and sea cliffs were
relatively stable. This led to large-scale urbanization along the coast, and with
it excessive watering of lawns, irrigation, septic tanks, leach lines, and cess-
pools. The extensive use of water has caused a steady rise in the water table,
Figure 1.1. which is a prime culprit in slope failure (see chapter 4). Thus, as the climate
Large landslide and earthflow change has produced more precipitation and more erosive storm waves, the
along the California coast
sea cliffs have been primed for failure by human activities. Landslides and
other mass movements are now more numerous, and blocks of the coastal bluffs
(often supporting homes) are slipping downward into the ocean (fig. l.l).
Clearly, we are geomorphic catalysts. Therefore, prior to its inception, any
major surficial project requires a detailed understanding of geomorphic pro-
cesses in order to predict how those processes will respond to our activities
(Coates 1976).
Another discipline directly dependent on a knowledge of process is plan-
etary science (Baker 1981). There is little doubt that recognition of landforms
is a key factor in interpreting the surface domain of our sister planets (fig.
1.2). However, simple landform identification is not enough. Understanding
the genesis of those features requires knowledge of how processes function in

analogous Earth environments and, equally important, how processes might
function in conditions that are alien to anything known on Earth (for example.
Komar 1979).
•
•4
*dr.
w>
» * .
Figure 1.2.
Large landslides on Mars.
Compare with figure 1 1
Chapter 1 6
Finally, an understanding of process is critical in itself. Cause

geoscience
and effect are essential components in the events that document geologic his-
tory. Our we lack an ex-
reconstructions of history suffer, however, because
plicit understanding of what effects from particular causative
will arise
processes. We know, for example, that Holocene climate changes were severe
enough to upset the delicate geomorphic balance. What is confusing is the
diverse geomorphic responses resulting from the same climatic trends. We
must conclude we commonly understand the
that, for the very recent past,
cause better than the effect. How then can we
how processes
confidently infer
functioned 300 million years ago when, in fact, we cannot always predict their
responses to modern stimuli? Certainly our insight does not increase as we
contemplate rocks because what we see are not processes but the results of
processes. We oversimplify the system in order to make any interpretation at
all. There is nothing wrong with this practice as long as we admit that our
models are based on what we think about processes, not what we know. What
geology needs is a precisely defined understanding of modern processes; until
geomorphologists provide it, our reconstruction of past events will be at best
educated guesses.
The Basics of Process Assuming that our focus on process is a viable way to examine geomorphology,
Geomorphology we must identify those concepts that, when integrated, constitute the basic
principles of process geomorphology. They are listed here and discussed in
detail on the following pages.
-/ 1. A delicate balance or equilibrium exists between landforms and

processes. The character of this balance is revealed by considering
both factors as systems or parts of systems.
•y 2. The perceived balance between process and form is created by the
/ interaction of energy, force, and resistance.
/3. Changes in driving force and/or resistance may stress the system
beyond the defined limits of stability. When these limits of
equilibrium (thresholds) are exceeded, the system is temporarily in
disequilibrium and a major response may

The system will
occur.
develop a different equilibrium condition adjusted to the new force
or resistance controls, but it may establish the new balance in a
complex manner.
4. Various processes are linked in such a way that the effect of one
process may initiate the action of another.
5. Geomorphic analyses can be made over a variety of time intervals.

In process studies the time framework utilized has a direct bearing
on what conclusions can be made regarding the relationship
between process and form. Therefore, the time framework should
be determined by what type of geomorphic analysis is desired.
Process Geomorphology —An Introduction
Figure 1.3.
Interpretation of slope adjustment
to geology by G K. Gilbert
Equilibrium slope developed at a
is maintained at times b and c
The Delicate Balance

The idea that some form of balance or equilibrium exists between landforms
and the processes that create them is not new. It was clearly expressed by
G. K. Gilbert during the latter part of the nineteenth century in his classic
reports on the geology in the western United States (see Gilbert 1877). Es-
sentially, Gilbert believed that under any given climate and tectonic setting,
landforms reflect some unique accommodation between the dominant pro-
cesses and the local geology. He often used the terms "dynamic adjustment"
and "balanced condition" to describe this relationship. An example of Gil-
bert's perception of equilibrium is shown in figure 1.3. Here we see a series
of slopes that are adjusted to alternating weak and resistant rock layers. The
slopes developed on the different units are produced and maintained by the
interaction of geology and processes such as mass movement, sheet wash, and
river flow. Importantly, Gilbert believed that continuous erosion would not
change the slope angles as long as the processes and their climatic and tectonic
controls remained constant. Thus, the slopes at times b and c will be a mirror
image of the slopes at time a because the process types and rates have not
changed through time. If tectonic or climatic controls change, processes will
also change, and new slope characteristics will develop in an adjustment to
the altered processes.
In the first half of the twentieth century, Gilbert's ideas were pushed aside
when geomorphologists espoused the concept developed by W. M. Davis that
landscapes change continuously with time and progress through distinct stages
that can be identified by regional geomorphic characteristics. was not untilIt
after World War II that the equilibrium approach was number

revitalized in a
of papers reemphasizing the importance of the adjustment between process
and form (Horton 1945; Strahler 1950, 1952a; Leopold and Maddock 1953).
This shift in emphasis resulted in the dynamic equilibrium concept in which
J. T. Hack (1960b) essentially brought back Gilbert's approach as a philo-
sophical framework for geomorphic analyses. Dynamic equilibrium suggests

that elements of landscape rapidly adjust to the processes operating on the
geology, and thus process and form reveal a cause-and-effect relationship. The
forms within a landscape maintain their character as long as the fundamental
controls do not change.
Many workers believe that the balance between form and process is best
demonstrated by considering both factors as systems or parts of systems. A
system is simply a collection of related components. For example, suppose we
define a drainage basin as a system and consider its measurable parts to be
basin area, valley-side slopes, floodplains, and stream channels. The balance
Chapter 1 8
or equilibrium condition within our system is revealed by statistical relation-

ships between the various parameters; i.e., basin area may be directly related
to total channel length, etc.
The systems approach has become highly sophisticated (Chorley 1962;

Chorley and Kennedy 1971), and different types of systems have been iden-
tified and used in geomorphology (Schumm 1977). For our purposes, it is best
to consider landforms and processes as part of the same open system in which
energy and/or mass are continually added or removed. Any flux in energy or
mass requires that the processes and their statistically related landforms ad-
just to maintain balance in the system.
The systems approach has these advantages:
1. It emphasizes the intimate relationship between process and form.*

2. It stresses the multivariate nature of geomorphology.
3. It reveals that some forms may not be in balance because they owe
their character to relict conditions. Some glaciated regions, for
example, may have landforms that were adjusted to geomorphic
controls different from those of the present.
As stated above, equilibrium implies that landforms (and presumably

processes) exist in some type of unchanging condition. In theory this requires
that factors which ultimately control landforms and process (such as climate
and tectonics) must also remain unchanged. In reality changes do occur in the
controlling factors with time. Thus, the true meaning of equilibrium depends
on the time interval over which our balance is being considered. Schumm and
Lichty (1965) argued that different time intervals, which they called cyclic,
graded and steady, are critical to our understanding of process and landform
development, and the distinction of these is extremely important in our per-
ception of equilibrium. Indeed this insight was followed by the further sug-
gestion (Chorley and Kennedy 1971) that different kinds of equilibrium are
related to each particular interval of time (fig. 1.4). Static equilibriunus^that
which exists over the short steady-time interval (days or month s). In this
framework of time, landforms do not change, and therefore^ they are truly
timeîndependent. In steady-state equilibrium, landforms and/or processes are
considered over graded time, perhaps 100 to 1000 years (Schumm 1977). The
equilibrium demonstrated in this interval is one in which changes do occur,
but their offsetting effects tend to maintain the system in a constant average
condition (fig. 1.4). In contrast, dynamic equilibrium must bejxmsidered over
cyclic time, perhaps millions ot years (Schumm 1977). In this case, even though
fluctuations of variables occur, they are not offsetting and the average con-
dition of the system is progressively changing (fig. 1.4).
With the foregoing perspective of time, it is apparent why the concept of

equilibrium has been difficult to define or understand. Time is a major factor
in the sense of equilibrium, and effective use of the concept in geomorphology
demands that the time framework be specified. We will examine the time factor
in process geomorphology later in the chapter.
Process Geomorphology —An Introduction
Figure 1.4.
Different time intervals and
associated equilibrium in
geomorphic analyses. (A) Steady

time (static equilibrium). No
change in channel gradient over
short periods (B) Graded time
(steady state equilibrium).
Constant average channel
gradient with periodic fluctuations
above and below the average
condition Measurements made
(A) Steady time — static equilibrium during intervals of steady time
within the graded time period may
show no change in channel
gradient. (C) Cyclic time (dynamic
equilibrium) Gradual lowering of
the average channel gradient over
long time intervals. Intervals of
graded time and steady state
equilibrium exist within the cyclic
time scale (Adapted from
Schumm 1977)
(B) Graded time— steady state equilibnur
/\/\>
'"\y\, = Intervals of
A/\
graded time
10 6 10 7
Time (years)
(C) Cyclic time — dynamic equilibrium
Force/Process/Resistance
We know from Newtonian physics that only a small number of fundamental
quantities are needed to explain mechanics. One of the mechanical quantities
identified by Newton is force, which —on the basis of his laws of motion —can
be loosely defined as anything that changes or tends to change the state of
motion in a body. In more specific terms, Newton defined force as a function
of mass times acceleration,
F = ma;
since acceleration is a vector parameter having both magnitude and direction,
force also is a vector quantity.
Chapter 1 10
Table 1.1 Common systems of units used in mechanical analyses.
Units 3
Systems Length Mass Force Time
cgs centimeter gram dyne second

fps foot slug pound second
mks meter kilogram newton second
3
\ slug = 1 lb sec^ft
1
! dyne = 1 g cm sec 2 1 newton (Ni) = 10^ dynes
The measure of force is weight. Therefore, the standard units of force are
pounds, dynes, or newtons depending on what system of units is being used
(table 1.1).Another fundamental quantity, mass, is directly related to force
as can be seen in the equation on the preceding page. In fact, by substituting
weight and acceleration of gravity into that equation, it can be expressed as
W= mg or m = —W ,
where W is weight, m is mass, and g is the acceleration due to gravity. This

demonstrates the interchangeable relationship between force and mass in me-
chanical analyses.
Force is also related to energy, and in geomorphology we can think of
landforms and processes as resulting from the application of energy. Energy
is defined as the capacity for doing work. It can neither be destroyed nor cre-
ated, but it can exist in many forms and can be changed from one form into
another. Kinetic energy of an object is energy derived by virtue of its motion.
Potential energy stems from the position of an object. Any change in the ki-
netic or potential energy of a body is equal to the work done on that body to
produce the change. Therefore, units of energy are the same as the units in
measuring work.
For example, consider a 100-pound steel ball carried vertically to the top
of a building 30 feet high. The work required to lift the ball to a higher ele-
vation represents a change in potential energy due to the increase in elevation.
The amount of work is defined as the product of force and the displacement
of the body in the direction of the force such that
Work = Fs,
where F is force and 5 is distance of displacement. In the fps system of units

(table 1.1), the unit of work is the foot-pound. In other systems, work is ex-
7
pressed in units such as the erg (dyne-centimeter), the joule (10 ergs), or the
newton-meter.
Thus, in our example above
Work = PE
PE = Fs
PE = 100 lb X 30 ft = 3000 ft lb.
where PE is the gain in potential energy.

This discussion is meant to show that energy and force are not the same
but that they are related through the concept of work. Since energy is the
capacity for work and work is a function of force, mental substitution suggests
to us that ( 1 ) a major component of energy measurement and (2) force
force is
in mechanics can be thought of as the application of energy, a phenomenon

that we will refer to as driving force.
In process geomorphology landforms represent interaction between driving

forces and resistance. Driving forces in geomorphology are climate, gravity,
and other forces generated inside the Earth. Resistance is provided by the geo-
logic framework. The link between these two components is process. Thus, as
stated earlier, process may be considered as the method by which one thing
is produced from something else, and as the vehicle by which a quantity of
one system is transferred into, and participates in, the mechanics of another
system.
In general, processes are either exogenic or endogenic. Exogenic processes
o perate at or near the Earth's surface and are normally driven by gravity and
atmospheric forces. Endogenic processes are different because-ihe-eneFgy-that
inkiates the action is located inside the Earth. The processes themselves may
operate at the surface, but their energy source is usually well below the sur-
face. Both types of processes may sometimes be involved in the development
of the samelandform. For example, the shape of a volcanic cone is the product
of both endogenic volcanism and normal exogenic slope processes.
In sum, we suggest that geomorphology can be examined by using phys-
ical concepts that revolve around the application of force on surface materials.
In our model the effect of processes depends on how vigorously the forces drive
them and how strongly their action is resisted by the geological framework.
Process, in this sense, allows us to explain the incredible variety of landforms
at the Earth's surface.
Driving Forces Having suggested that energy is exerted on Earth materials

as a driving force,we should briefly examine the major forces in our systems.
Although each of these has been detailed after long and careful study, we will
treat them only briefly to fit our specific needs.
Climate Radiation emitted from the sun is the major source of energy needed
to drive exogenic processes. Radiation is expressed in terms of heat, a form of
energy possessed by molecules of matter because of their motion. Heat could
be expressed in normal units of energy, but it has historically been measured
in the special, more convenient units of calories or British thermal units (Btu).
These are simply measurements of the amount of heat required to raise the
temperature of a specified mass or weight of water one degree.
If an imaginary plane were placed at the outer limit of the atmosphere,
2
perpendicular to the incoming rays of sunlight, it would receive 2.0 cal/cm /
min of radiant energy over its entire surface. This value, called the solar con-
stant, represents the small fraction of the estimated 100,000 cal/cm 2 /min of
energy produced by the sun that survives the long journey to the Earth. The
2
solar constant averaged over the entire surface is only 0.5 cal/cm /min. As
Chapter 1 12
Table 1.2 Annual heat balance and the transfer of heat in different latitude
zones
Short-Wave Long-Wave Poleward Transport

Fraction Radiation Radiation ofHeat Across
Zones of Latitude of Total Absorbed Emitted Latitude Parallels
(degrees) Area (cal/cm 2 /min) (cal/cm 2 /min) (cal/min)
0-20 34 39 030
57 X 10' 5 (20°)
20-40 030 034 030

77 X 10 15 (40°)
40-60 022 23 030
50 X 10 15 (60°)
60-90 0.14 0.13 030

Weighted mean 0.30 30
From Handbook of Applied Hydrology by Ven T Chow Copyright <& 1964 by McGraw-Hill Inc Used with permission of
McGraw Hill Book Company
sunlight passes through the atmosphere, another segment of the radiation is
reflected back into space by clouds, particulate matter in the atmosphere, and
the Earth's surface. Therefore, the amount of energy absorbed in the system
(insolation) and actually available for work averages about 0.30 to 0.35 cal/
cm 2 /min over the entire globe.
The average insolation value varies greatly, however, with latitude (table
1.2) and with the seasons. Since the total heat budget does not change, the
earth-atmosphere system must return to space as much heat as it receives,
which it does in the form of long-wave, blackbody radiation. It is significant
that although absorbed radiation decreases with increasing latitude, heat loss
is fairly constant (table 1.2). This produces an obvious temperature differ-
ential between the equator and the poles, which drives a poleward transfer of
heat in the oceans and, even more, in the atmosphere. The transfer of solar
energy demands a series of complex processes that generate the various com-
ponents of our weather. These processes tend to establish reasonably well-de-
fined temperature and precipitation patterns for all portions of the Earth's
surface. The average weather conditions at any place, considered over a long
period of time, are called climate. Climate represents the net result o_f_how
solar energy is distributed in the earth-atmosphere system. In combination
with the surface material of a region, climate determines vegetation, weath-
ering and soil-forming processes, and the hydrology. Combined with gravity,
it controls glaciation, mass wasting, and fluvial processes.
Some is transferred by conduction into

of the heat absorbed at the surface
exposed Earth materials (rock, and into the air above the surface.
soil, etc.)
How much heat is conducted into each medium and how deeply it penetrates
depend on the physical properties of the air and surface matter. As Petterssen
(1964) points out, these conductive properties can be linked to several im-
portant facts about thermal distribution. First, the high heat capacity of water
makes large lakes and the oceans natural storage bins for heat. Second, the
temperature ranges over continents are significantly greater than over the
oceans. Third, the depth of heat penetration will be much larger in air or water
than in solid, denser materials.
Considered separately in terms of radiation and of conduction, the Earth's
surface and the atmosphere do not have balanced heat budgets. The surface
gains more heat than by these methods; the atmosphere loses more
it loses
than it gains. What balances the thermal ledger on a local scale is another
process of heat transfer, convection, which causes hotter and lighter air (or
water) to move toward zones of lower temperature. Air near the Earth's sur-
face is warmer and so will rise into the cooler zones higher in the atmosphere.
Some heat absorbed at the surface is used in the process of evaporation and
transferred into the atmosphere in a latent form, along with the ascending
water vapor. About 600 cal of heat is absorbed by air when one gram of water
is evaporated. It is subsequently released in the atmosphere as sensible heat
during condensation and precipitation. This heat usually is liberated at some
distance from the point of evaporation, and therefore the redistribution of heat
depends partly on air motions and the conditions needed to produce conden-
sation.
The combined effects of these factors help explain the great differences
between oceanic and continental climates. Continents heat and cool faster than
oceans, with more extreme variations in temperature. Since the thermal char-
acter of the surface controls the heating of the adjacent air, we can expect
temperatures over land and water to function the same way as the surface
itself. The seasonality of mid-latitude continental climates can also be under-
stood in terms of the relative rates of heating and cooling and the associated
pressure changes of the land-water settings.
On a larger scale, the inequality of heat with latitude (table 1.2) requires
a transfer of heat from the equatorial region to the poles. The precise me-
chanics of this transfer involves a series of complexly interrelated processes
that are a basic concern of meteorology. Variables include world circulation
patterns of air masses, vertical and horizontal pressure distribution, fronts,
rotation of the Earth, and the distribution of landmass and oceans. In general,
heat is transferred poleward by air motions, controlled by the average air pres-
sure and wind patterns of the Earth.
As figure 1.5 shows, at latitudes near 30° (north and south), high-pres-
sure centers dominate the circulation pattern and drive the persistent trade
winds at low latitudes. Where the trade winds converge, an equatorial trough
is created, and humid low-pressure air forms the rising limb of a giant con-
vection cell. the descending limb of the

The air moves poleward and sinks, as
convection cell, along the zone of high pressure at 30° latitude. Sinking high-
pressure air inhibits cloud formation and causes the predominance of desert
conditions in this region. Thus, at low latitudes most heat it transferred pole-
ward by cm2yectionu_circulalJQn, and the evaporation-precipitation process.
middle latitudes, heat migrates in association with cyclone and an-
In the
ticyclone wind circulation in a zone of unsteady westerly air flow ("prevailing
westerlies"). Here cold polar air meets the air moving poleward from lower
Chapter 1 14
North pole
Figure 1.5.
Prevailing wind patterns and
cellular air motions in the Northern
Hemisphere (Adapted from
Rossby 1941)
30° Pr evailinQ westerlies
/ /
~Horse latitudes -
Trade winds \2=^
Equator-
S '*«?**
Equatorial trough
latitudes. Instead of mixing, the different air masses remain clearly defined
and are turned into warm low-pressure masses rotating counterclockwise and
cool high-pressure masses rotating the opposite way. In the middle latitudes,
much of the poleward heat transfer is associated with atmospheric distur-
bances.
At higher latitudes, warmth is delivered poleward in the upper air by hor-
izontal mixing, called advection. when warm air rides over southward-flowing
polar air at the polar front.
Ocean currents also are important dispensers of heat. Warm ocean cur-
rents moving toward the poles flow under cold air and give off energy as latent
heat that ultimately warms the overlying atmosphere. The distribution of heat
is not complete in the sense that all regions attain equivalent temperatures. It
is effective, however, in stabilizing temperature conditions on a regional basis,

providing the Earth with a rather well-defined temperature pattern.
Precipitation is also controlled by the same processes that distribute tem-
perature, and the pattern of precipitation generally follows the major zones
of atmospheric circulation. Precipitation is greatest near the equatorial trough
and least near the 30° sinking limb of the low-latitude convection cell (fig.
1.6). Secondary peaks of precipitation occur in the middle latitudes with the
unsteady pressure cells of that region. Precipitation differs from temperature,
however, in that it can only occur when large air masses are cooled. There
must be some triggering mechanism that causes moisture-laden air to ascend
and be cooled by the decreasing temperatures at higher elevations. The prin-
cipal mechanisms that lift air masses and thereby initiate precipitation are
convection, orographic effects, and frontal mechanics. In convection, warm
air, being lighter than surrounding air, rises until condensation forms the fa-
miliar bulging shape of cumulus clouds. The orographic effect occurs when
air masses are forced to rise over high mountain ranges. Frequent rain occurs
on the windward side of the mountains, with a characteristic rain shadow on
the leeward side. Frontal activity involves the interaction of low-pressure and
high-pressure zones. A front is simply the line of contact between the moist,
warm air of the low-pressure cells and the cool, dry air associated with high-
pressure cells. At the frontal contact, warm air is forced to rise up and over
the cooler air and so is cooled; the result is precipitation.
1 5 Process Geomorphology —An Introduction
70
i i i i
Figure 1.6.
Mobile, Ala. Generalized world distribution of
60 average precipitation according to
latitude.
\ Asuncion, Paraguay
\ •
a> 50
£ 40 \ A
£ 20
\
10 \
_*,/ San Diego, Cal.
Iquique, Chile >».. _
• i i
80 60 40 20 20 40 60 80
North latitude South latitude
Gravity The second major driving force, gravity, manifests itself in a myriad
of both endogenic and exogenic geomorphic processes. Combined with the cli-
matic engine, gravity determines the rigor of fluvial power, mass wasting, gla-
ciation, tidal effects on coastal processes, and the movement of ground water.
Internally, gravity bears directly on the process of isostasy, which tends to
control the distribution of Earth materials of different densities, ultimately
powering regional uplift. Gravity is ubiquitous, affecting all substances. The
force of gravity is applied continuously in every system at, above, and beneath
the surface and so can never be completely ignored in any consideration of
process.
Sir Isaac Newton's classic work on the force of gravity was published in
1687, introducing his law of universal gravitation. Simply stated, the law says
that there exists between any two objects a mutual attractive force that is a
function of the two masses (mi and m 2 ), the distance separating them (r), and
the universal gravity constant (G):
„ _ mjmz
Thus, gravity attraction between two objects is an action-reaction phenom-

enon. Each body exerts a force on the other that is equal in magnitude but
oppositely directed along a straight line joining the two bodies. Our main in-
terest in gravity is how it affects geomorphology, especially surficial matter.
The gravitational force exerted on surface materials is measured in terms of

the amount of acceleration that the force imparts to any freely falling particle
having mass. It is normally expressed by the equation
GM
g
where M is the mass of the Earth. In most scientific work g is assumed to be

constant, having a value of 980 gals (a gal being a unit of gravity, which is
1 cm/sec/sec).
Chapter 1 16
From this equation, it is obvious that g in fact cannot be a constant, as

we normally assume, because it depends on several variable factors. The dis-
tance (r) changes because of topographic irregularities and because the Earth
is not a perfect sphere. The density of the Earth materials is not evenly dis-
tributed and so may vary along the line connecting the masses. In addition,
the rotation of the Earth introduces a counteracting force and causes a distinct
latitudinal variation in gravity. Therefore, g is not distributed regularly at the
Earth's surface. This fact is a justifiable concern of geophysicists because slight
variations in gravity have real significance, especially as an exploration tool.
However, the variation in gravity at the Earth's surface is so small compared
with the total magnitude that for most exogenic analyses g can reasonably be
considered to be constant, and this is normal practice in process analyses. On
the other hand, the minor variations that reflect internal density or mass dif-
ferences are extremely important in endogenic processes. We will discuss
gravity again in chapter 2 when considering isostatic adjustments as a geo-
morphic process.
Internal Heat Thermal energy is generated inside the Earth, primarily by

radioactive decay and secondarily by friction caused by earth tides and rock
deformation. The exact amount of heat available for geologic work is unknown
because thermal characteristics deep within the Earth must be estimated from
other physical properties (density, pressure, gravity, etc.) that have been de-
termined from analyses of seismic waves. Internal heat can be measured di-
rectly only in deep wells or mine shafts; any postulated thermal distribution
below the outer fringe of the crust is based on assumption, not observation.
Not only are we uncertain about the physical and chemical properties of sub-
crustal rocks, but hypotheses about temperature distribution tend to involve
us in consideration of how much heat the Earth obtained during its formation
and early Because of the ambiguities involved, estimates of thermal
history.
gradients within the Earth vary considerably (see Wyllie 1971). Temperatures
proposed for a depth of 1000 km, for example, differ by as much as 1500°C;
even at a relatively shallow depth of 100 km, estimated temperatures var\ by
approximately 600°C.
Regardless of the many problems inherent in determining the vertical dis-
tribution of temperature, it is a fact that the Earth transmits to the surface
about 2.4 X 10 20 cal each year of its internal heat. The total amount of heat
is minor compared with the heat received at the surface from solar radiation,
but it does indicate that heat, no matter what its origin or gradient, is being
transferred from place to place within the Earth. The mechanics of heat transfer
is significant since the energy distributed drives internal geological procc-
Like its atmospheric counterpart, internal heat is transferred by several
methods. Conduction is very slow because of the low conductivity of silicate
minerals, but nonetheless it is the dominant transfer mechanism in the crustal
layers. Convection as a method is still hypothetical since it cannot actualh be

seen, butits presence is supported by observed tectonic features, such as the
evidence for seafloor spreading, that are virtually inexplicable without some
type of convective overturn. Theoretically, convection is caused by tempera-

ture differences at depth (presumably in the mantle) that heat rocks locally
and thereby create a less dense mass. The hot, light rocks rise toward the
surface as cool, denser rocks are simultaneously sinking to replace the as-
cending mass. In this way, rock materials of different heat and density are
continuously exchanged, following the path of a large convective "cell." The
excessive heat at depth is carried along with the rising rock masses and re-
leased closer to the surface, efficiently transferring heat.
Measurements of heat reaching the Earth's surface are difficult and costly,
and often they are affected by secondary factors such as ground water, vari-
ations in conductivity, and recent volcanism. In addition, measurements are
not randomly spaced but tend to be concentrated in areas of some specific
interest so that large regions exist for which little or no data are available.
Nonetheless, the development of sophisticated instrumentation and the cur-
rent interest in ocean tectonics have produced a storehouse of information that
is beginning to yield a reasonable picture of surface heat flow. Except for local
abnormalities, heat emerges from all parts of the Earth in amazingly equal
amounts, with average continental and oceanic values differing by only 0.2 n
cal/cm 2 /sec (Wyllie 1971). is the major source of internal
If radioactivity
heat, the equality of heat flowfrom continents and ocean floors would require
an unusual distribution of radioactive minerals beneath the two environments
unless the thermal condition were balanced by a convective process. Such a
process may be demonstrated by examining heat flow for major physiographic
regions of the Earth, as presented by Lee and Uyeda (1965) and shown in
table 1.3. Note that heat flow from ocean ridges and trenches differs consid-
erably from average values for entire ocean basins; ridge crests are abnormally
high and trenches notably low. Heat may be actively rising under ocean ridges
as part of a convective overturn, while the low heat values beneath the trenches
presumably represent the descending limbs of the overturning cells. On con-
tinents, as one would expect, the lowest heat flow values occur in the very
stable shield areas and the highest in the most recent orogenic belts and their
associated regions of Cenozoic volcanism.
The transfer of internal heat plays a significant role in determining the
major topographic framework of Earth. Heat transfer drives processes be-
neath the surface causing uplift and deformation, distributes rock masses of
varying resistance, and controls the volume of ocean basins, thereby influ-
encing the position of sea level. Precisely how or if heat flow relates to gravity
distribution is debatable, but certainly the two forces combined represent a
major geomorphic element.
The Resisting Framework As pointed out earlier, landforms reflect a bal-
ance between the application of driving forces and the resistance of the ma-
terial being worked on. Having reviewed the salient features of the driving
forces in our systems, we should now examine the resisting elements, but ex-
actly how to do so is rather perplexing. It is tempting simply to state that the
Chapter 1 18
Table 1.3 Heat flow values from major geologic features
Average Heat Flow

Geologic Feature (Mcal/cm 2 /sec)
Land Features
1 Precambnan Shields 0.92
Australian Shield 1.02
Ukrainian Shield 0.69
Canadian Shield 088
S African Shield 1 03
Indian Shield 066
2 Post-Precambrian
Nonorogenic areas 1.54
Europe 1.67
Interior Lowlands, Australia 2 04
Interior Lowlands. N America 1 25
S Africa 1.36
3 Post-Precambrian
Orogenic areas a 1.48
Appalachian area 1 04
E. Australian highlands 203
Great Britain 1.31
Alpine system 209
Cordilleran system 1.73
Island arcs 1 36
4 Cenozoic volcanic areas 6 2 46
Ocean Features
1. Ocean Basins 1.28
Atlantic 1.13
Indian 1.34
Pacific 1.18
Mediterranean seas 1.20
Marginal seas 1 83
2 Ocean Ridges 1.82
Atlantic 1.48
Indian 1.57
Pacific 2.13
3 Ocean Trenches 0.99
4. Other ocean areas 1.71
From Lee and Uyeda Geophysical Monograph 8 1965 Copyrighted by Amerccan Geophysical Union Reprinted by
permission
Excluding Cenozoic volcanic areas
"Excluding geothermal areas
resistance in geomorphic systems is geology — the geologic affect on geomor-

phology is so pervasive and so varied that any brief review of its role in de-
termining process and form must be inadequate. A complete discussion of the
geological control of geomorphology would require an analysis of every pos-
sible geologic framework in every possible climatic and tectonic regime. Al-
though such an effort is impossible here, some general examples will show how
geological resistance manifests itself in landforms.
Lithology The resisting force in geomorphology is implemented through the

two major geologic variables, lithology and structure. The diverse origins of
rocks create lithologies at the surface that differ vastly in their chemical and
mineralogic compositions, textures, and internal strengths. In geomorpholog\
1 9 Process Geomorphology —An Introduction
Table 1.4 Weight percent of common elements in Earth's lithosphere.
Continental Crust Oceanic Crust Total Lithosphere
Ronov Ronov Ronov

Polder- Pakiser and Yaro- Polder- and Yaro- Polder- and Yaro-
vaart and Robin- shevsky vaart shevsky vaart shevsky
1955 son 1966 1969 1955 1969 1955 1969
Si0 2 594 57.8 61 9 46.6 48.7 55.2 593

Ti0 2 1.2 1.2 08 2.9 1 4 1.6 09
Al 2 3 155 15.2 15.6 150 165 153 159
Fe 2 3 2 3 2.3 2.6 3.8 2.3 2.8 25
FeO 50 55 3.9 8.0 6.2 5.8 4.5
MgO 4.2 5.6 3 1 78 68 5.2 4.0
CaO 6 7 7.5 5.7 11.9 12.3 8.8 72
Na 2 3.1 3.0 3.1 2.9 2.6 2.9 3.0
K2 2.3 20 2.9 1.0 04 1.9 24
Table 1.5 Abundance of rock and mineral types in the Earth's crust.
%of
Crustal Modal
Rocks Volume Minerals %
Quartz 12
Sands 1.7 K-feldspar 12
Clays and shales 4.2 Plagioclase 39
Carbonates 2.0 Micas 5
Granites, gneiss Amphibole 5
and crystalline schist 36.9 Pyroxene 11
Granodionte and Olivine 3
diorite 11.2
Clay 4.6
Syenite 04 Calcite and
dolomite 2.0
Basalt, gabbro, Magnetite 1.5
amphibolite, eclogite 42.5
Pendotite, dunite 02 Others 49
Total 99.1 Total 100.0
Modified from Ronov and Yaroshevsky. Geophysical Monograph 13. pp 37-57. 1969 Copyrighted by American Geophysical
Union Reprinted by permission
we are concerned with the modern resisting framework, regardless of its his-
tory. It is important to gain an overall picture of the crustal and surface rock
distributions as they presently exist.
Table 1 .4 synthesizes several estimates of the bulk chemical composition
of the Earth's lithosphere. As expected, the chemistry of continental crust is
higher in silica and K2 than that of oceanic crust, and lower in CaO, MgO,
and total iron. Such a chemical distribution can be converted into reasonable
estimates of the volume-percentage of common rock types and their modal
mineral composition (table 1.5). The significance of these analyses is to em-
phasize that the resisting framework in geomorphology basically entails only
T2
Chapter 1 20
Table 1.6 Volume percentage and chemical composition of silicic and mafic
crust in the United States.
Western Provinces Silicic Mafic Western Eastern

California coastal Si0 2 600 57 .1
region 75 25 Ti0 2 1.1 13
Sierra Nevada 50 50 Al 2 3 15 1 152
Pacific NW (coastal) 28 6 71 4 Fe 2 3 23 23
Columbia Plateau 223 111 FeO 49 57
Basin and Range 66 7 33.3 MgO 45 56
Colorado Plateau 625 37 5 CaO 63 75
Rocky Mountains 62 5 37 5 Na 2 30 30
K2 20 2 1
Average 564 436
Eastern Provinces Total 99 2 99 8
Interior Plains
and Highlands 400 60.0
Coastal Plain 572 428
Appalachian Highlands
and Superior Upland 375 625
Average 426 574
Total United States 463 536
From Pakiser and Robinson Geophysical Monograph 10 pp 620-26 1966 Copyrighted by American Geophysical Union
Reprinted by permission
two igneous and metamorphic rock suites, and approximately ten mineral va-
rieties. The crust consists primarily of a silicic assemblage (granites, gneisses,
schists, granodiorites, and diorites) that makes up 48 percent of the crustal
volume and a mafic association that constitutes about 43 percent. Obviously

the silicic group is plutonic or metamorphic in origin and is dominantly con-
tinental; the mafic types are overwhelmingly volcanic and rooted beneath the
oceans.
The crust beneath the conterminous United States, however, is more mafic
than one might guess (table 1.6). Pakiser and Robinson (1966) point out that,
based on seismic velocities, the total U.S. crust is 54 percent mafic by volume
(55 percent by weight). In addition, they show that the mafic content is con-
siderably greater in the provinces of the eastern United States. In general, the
eastern regions have a crust that is predominantly mafic, and the western pro-
vinces a crust that is mostly silicic.
If Pakiser and Robinson are correct, it is even more interesting to examine

the igneous rocks exposed at the surface in the Appalachian and Cordilleran
regions (table 1.7). In the Appalachians, where the crust is predominant!)
mafic (as noted in table 1.6), the surface igneous rocks are overwhelmingly
calc-alkalic, plutonic rocks. Of the rocks of this type indicated in table 1.7
(84.5 percent of the total), 96 percent of the plutons are granites. In contrast,
the igneous rocks exposed in the Cordilleran system are mainly extrusive (63.6
percent), and of these 77 percent are basaltic or andesitic in composition. The
thick mafic crust in the eastern United States supports a surface rock assem-
blage that is dominantly granitic, in contrast to a silicic crust supporting mafic
surface rocks in the west.
Table 1.7 Area of different igneous rocks exposed in fhe Appalachian and
Cordilleran regions of the United States (percent).
Cordilleran Appalachians
Plutonic Rocks
Calc-alkalic Rocks 33 6 84 5
(granite, granodionte, quartz monzonite,
quartz diorite, diorite, gabbro, anorthosite)
Alkalic Rocks 04 neg
(syenite, monzonite, others)
Ultramafic Rocks 05 neg
(periodotite. pyroxemte)
Hypabyssal Intrusives
Calc-alkalic Rocks 1 5 74
(porphyries, quartz diabase, diabase)
Alkalic Rocks 03 02
(porphyries)
Extrusive Rocks
Calc-alkalic Rocks 63 4 79
(basalt, dacite, andesite, rhyolite)
Alkalic Rocks 02 neg
(trachyte, latite, phonolite. others)
Total 99 9 100
Adapted from Daly. P (ed Igneous Rocks and the Depths

. ) ol the Earth © 1933 McGraw-Hill Inc Used with permission of
McGraw-Hill Book Company
Table 1.8 Rocks exposed at the surface of the North American continent
(expressed as % of area).
Gilluly 1969 Blatt and Jones 1975
Sedimentary 61 5 52
Volcanic 82 11
Plutonic 3.8 6
Metamorphic and total PG 265 31
InNorth America, sedimentary rocks make up most of the exposed ma-

even though they are only a minor constituent of the total
terials (table 1.8)
crustal volume. Their ultimate source, however, is older igneous, metamor-
phic, and sedimentary rocks, and so their chemistry and mineralogy reflect
changes induced by exogenic geomorphic processes. Geomorphology, there-
fore, becomes an important link in the rock cycle.
The wide areal distribution of sedimentary rocks undoubtedly causes a
surface mineral composition different from that shown in table 1.5. At the
surface, quartz and feldspars are dominant and probably exist in equal amounts
(feldspar 30 percent, quartz 28 percent); calcite and dolomite increase to about
9 percent; and clay minerals and micas become much more significant, rising
to approximately 18 percent of the surface material (Leopold et al. 1964).
Chapter 1 22
In any given climate each rock type will respond to the processes of weath-
ering and erosion in a different manner and at a different rate. With time and
tectonic stability, high-standing landmasses commonly will be underlain by-
resistant rocks,and low-standing regions will be formed from rocks that are
more susceptible to weathering and erosional attacks. These effects of differ-
ential weathering and erosion in landscape development are stressed in every
introductory course in the basics of geology. In fact, we are conditioned early
in our geological training to view regional topography as a mirror of gross
lithology, tectonics, and geologic history. For example, the concept of physio-
graphic provinces stresses this approach, causing us to think of geological con-
trols in geomorphology as regional phenomena. It is worthwhile to emphasize,
however, that geomorphic processes will accentuate lithologic differences on
many scales. Mega-scaled differentiation produces regional features such as
mountains and plains (fig. 1 .7), and can be utilized in an erosional topography
as a first approximation of the gross lithologic distribution. Within any large
region of similar rock type, small lithologic discrepancies will also surrender
to geomorphic processes and appear at the surface as minor landform devia-
tions. These tiny blips in the general landscape provide critical information
about geological history and exert important controls on subsequent geo-
morphic developments (figs. 1.8 and 1.9). Even on a microscopic scale, lith-
ologic variations may have a distinct effect on the style of weathering (fig.
1.10), ultimately causing subtle topographic differences (Eggler et al. 1969).

It seems certain that even long periods of erosion cannot completely erase the
Figure 1.7.
Mountains and surrounding plains,
looking west-southwest from a
point about 16 km northeast of
Boulder at an elevation of 2,152
meters Boulder County, Colo.,
ca 1934
23 Process Geomorphology—An Introduction
Figure 1.8.
A topographic irregularity caused
% —
by differences in lithology West
Spanish Peak, from the northwest
Dikes cutting flat-lying Eocene
strata. Spanish Peaks quadrangle,
Huerfano County, Colo
Figure 1.9.
Variations in lithology as
evidenced in cuestas formed by
hard sandstones north of Galisteo
Creek, N.M. The rocks in
^^^fcfc.
succession from left to right are
Mansamo red beds, Morrison,
Dakota, and Mancos (Galisteo
Creek and the Santa Fe Railroad
in foreground)
'• **,.
m.
Chapter 1 24
influence of minor lithologic abnormalities from the landscape (Flint 1963).

although their appearance may be greatly subdued.
Lithologic diversity must be considered on a variety of levels. Large areas
underlain by crystalline rocks or sedimentary rocks may develop a distinct
regional character, but smaller variations within the region are revealed in
subtle topographic changes that often provide significant geologic and geo-
morphic information. The geomorphologist must be able to "read" these subtle
topographic modifications in order to present a coherent interpretation of his-
tory and process.
Structure Geologic structures that influence landforms also range in mag-

nitude from large, areawide tectonic styles to minor features that exert only
local control (Lattman 1968). Structural influence is readily apparent only
when the rocks and climate involved are conducive to differential weathering
and erosion. In depositional environments, structures may be buried by thick
accumulations of sediment that mask the surface expression of the underlying
structure. Comparably, the internal structure may not be immediately evident
in erosional topography formed in areas with distinctly similar lithology. such
as shields or crystalline mountain cores, but minor structures still may pro-
duce a discernible topographic control (Flint 1963). Spacing of joints, for ex-
Figure 1.10.
Disintegration of granitic boulders ample, is recognized as a prime factor in the development of the longitudinal
due to expansion of biotite grains "staircase" profiles that characterize glaciated valleys in mountains held up
Boulders in terrace gravel near
Red Lodge. Mont
25 Process Gee-morphology —An Introduction
by rocks of uniform lithology. The most likely lithologic environment to dis-

play structural control is a sedimentary sequence with alternating resistant
and nonresistant units, such as the Valley and Ridge province of the Appa-
lachian Mountains. There resistant sandstone and conglomerate layers form
ridges that are separated by intervening valleys underlain by easily eroded
shales and limestones. The regional topography reveals the pervading struc-
ture of plunging anticlines and synclines because the ridges cross the coun-
tryside in a sinuous pattern that shows the character of the underlying folds.
Thresholds and Complex Response

The third basic principle of process geomorphology involves the threshold con-
cept. It may or may not be apparent to you that any concept proposing equi-
librium inherently implies a contrasting state of disequilibrium. If variations
in controlling factors demand a response within the system, there must be a
period of readjustment during which process and form are out of equilibrium.
Landslides, subsi d ence, an dgulley erosion are examples of disequilibrium
gene rated when the variablesoTTorce and^/ oTj^ sistance are afteredso the y
can no longer maintain a balanced relati ons hip (fig. 1.11). They represent
events that occur as systems attempt to reestablish a new equilibrium. Such
events can happen suddenly or can proceed toward equilibrium over a long
period of time, depending on how great the disequilibrium is and how much
energy is involved.
Schumm (1973) recognized that if a system in equilibrium can be defined
by real parameters, it follows that there must be parameter values that rep-
resent the limits of the balanced condition. If these limiting values are ex-
ceeded, the system enters a condition of disequilibrium. The limits of
equilibrium_arg__ciitic^ljandlti ons called th resholds (Schumm 1973).
If parameter v alues are pushed to the limiting co n dition by va riations of
external controlling factors, the threshold is known as an e xtrinsic thresho ld.
Examples are numerous in nature; geologists will be most familiar with

threshold velocity in streams, at which sediment movement begins. The change
in external variables (in this case the force) causes instability of the channel
Figure 1.11.
Threshold relationship between
gullied and ungulhed valley floors
in several drainage basins of
northwest Colorado.
0.01
I I
10 20 30 40 50
Drainage area (mi )
\
Chapter 1 26
sediment. Other examples are found in responses to the fluctuating climate

that characterized much of the Pleistocene epoch. A jrior£_siihlle type of
threshold, however, is the intrinsic threshold, where instability aurLf-a-iUire of
a system occurs even though external variables remain relatively-constant.
The threshold conditions develop in response to gradual, often imperceptible,
changes within the system. In many cases the threshold represents a deteri-
oration of resistance rather than an increase in driving forces. For example, a
region characterized by periodic heavy rains may have stable slopes for a long
time, but continuous freeze andthaw or other soil-forming processes gradually
reduce the cohesion of slope material. Eventually one storm, no more severe
than thousands that have preceded it, triggers slope failure.
A special typ e of threshojcL.calkda^ggo worp/?/c threshold (Schumm 1973,
1980), essentially refgrsjo the-SlabiLit y of a la ndjormjtself. Originally Schumm
(1973) considered this to be a type of intrinsic threshold; i.e. progressive al-
teration of the landform, especially the slope factor, eventually reaches a sta-
bility limit and demands a rapid adjustment in the character of the feature.
The significance of the concept is that abrupt changes may be a normal part
of landscape development and do not always require a change in external con-
trols to precipitate the event. Schumm (1980) has since expanded the concept
to include externally generated changes in landforms.
It can be demonstrated that a threshold response often occurs as a_series
of reactions called complex response. The sequence of events happens because
all processes and components of a system may not reach the threshold con-
dition at the same time. This phenomenon was demonstrated experimentally
by Schumm and Parker (1973) in a study of an artificial drainage basin. An
induced base level decline at the mouth of the basin caused downcutting of
the trunk river at that point and the formation of terraces. At the same time,
however, tributary channels were unaffected and remained in their equilib-
rium state. With time, the site of channel incision migrated progressively up-
stream until the base level of each tributary was lowered and channel
entrenchment ensued. The tributary incision, however, provided so much sed-
iment to the trunk river that aggradation began at the basin mouth because
the stream was incapable of transporting the increased load derived from en-
trenchment of the tributary channels. Clearly, the processes functioning in
different parts of the systemic basin were out of phase.

It is not difficult to imagine the same sequence of events occurring during
major glacial stages when sea level declines dramatically. The effect of that
base level decline will be initially felt at the mouths of major rivers such as
the Mississippi. Tributaries in that huge basin may not experience the ex-
pected incision until long after the initiating event.
Actually, natural complex responses similar to the experimental study have
been documented. For example, the changes in fluvial systems in response to
hydraulic mining for gold on the west flank of the Sierra Nevada produced a
sequence that was generally the same but diametrically opposed in detail to
the Schumm and Parker study. Gilbert (1917) was able to show that as mine
tailings were released from the mountain deposits, the coarse fraction grad-
ually invaded the channels of the major downstream rivers as sand and gravel
bedload. The rivers, unable to transport such an overwhelming load, adjusted

by drastically raising their channel bottoms as the material was deposited. As
each segment filled, the gradient increased so that the river acquired the ca-
pacity needed to transport the sediment farther down the valley. The rise in
channel level stopped at different times in each segment of each river, de-
pending on the distance from the source and the amount of load. For example,
the channel of the Yuba River at Marysville (Cal.) rose about 6 m (19.1 ft)
between 1849 and 1905, when it reached its highest level. The Sacramento
River at Sacramento (Cal.) elevated about 3 m (10.8 ft), attaining its highest
level in 1897. It is interesting to note that both rivers continued to fill even
after mining ceased in 1884. This happened because the upstream reaches, no
longer receiving great volumes of sediment, had excess energy on their steep-
ened gradients and therefore entrenched their channels. The sediment from
the entrenching process was transported to the lower river segments. Thus,
part of the fluvial system was filling while, at the same time, other parts were
entrenching.
The complex response observed by Gilbert differs from that of Schumm
and Parker because (1) the force initiating the threshold occurred near the
basin divides rather than at the trunk stream mouth, (2) the driving force was
associated with human activity, and (3) the initial response was aggradation,
the site of which proceeded downvalley.
Thresholds in geomorphology were first demonstrated in fluvial systems.
They have since been recognized in almost every aspect of the discipline (for
examples see Coates and Vitek 1980), and innovative methods are being de-
vised to indicate when a system is tending toward a threshold condition (Bull
1979, 1980). In addition, it is increasingly clear that the systemic stress that
produces instability is commonly generated by human activity. Thus, the
widespread applicability of the concept has prompted the suggestion that it
should be the primary working model of geomorphology (Coates and Vitek

1980). It is difficult to argue with that suggestion because it is when stability
limits are exceeded that things begin to happen, and many deleterious events
at the surface may be nothing more than nature's way of responding when a
threshold is passed. With that consideration, it becomes critical for geomor-
phologists to define threshold values for every environment and for all con-
ceivable combinations of process and geology. Such information would be
extremely important for future land management and could be the foundation
for identifying natural hazards and predicting imminent disasters.
The Principle of Process Linkage

Complex adjustments to altered conditions often involve a chain reaction of
responses that we will call process linkage. Process linkage_esse ntially ope r-
ates on the domino principl e; it means that the changes that occu r in one pro-
cess -COlJaQdiorrn_ duri ng an aHjiistrnent__perinH often initiate subsequent
respo nses in totally diflHerrrprocessesand/or landforms. Linkage works be-
cause a driving force can transfer from one process type to another as its effect
filters through a system, or it can even shift to processes operating in totally
different systems. Thus, a myriad of different processes can be involved in the
response to a single threshold-inducing force.

Chapter 1 28
EXPLANATION
Town ^ River surveillance site
10 15 20 Kilometers (stream gage)
Bridge
Figure 1.12.
Map showing location of river
A recent case history exemplifies how process linkage works. On May 18,
surveillance sites in the lower 1980, Mount St. Helens in southwestern Washington was blown apart by a
Toutle and Cowlitz river systems violent volcanic eruption. The widespread effects of the eruption have been
(From Lombard et al US
Geological Survey Circular 850,
documented in a series of short papers published as U.S. Geological Survey
1981) Circular 850. The initial process response occurred during the eruption as a
massive debris avalanche that deposited enormous volumes of rock. ice. and
other debris in the upper 17 miles of the North Fork Toutle River valley (fig.
1.12). The deposits are up to 600 feet thick at places. Physical, chemical, and
biological characteristics of lakes close to the eruption were drastically al-
tered, and benthic faunas in the adjacent rivers were destroyed.
Immediately following the avalanche, snow and ice that had melted during
the eruption provided enough water to generate a mudflow in the same valley.
In addition to environmental damage, the mudflow deposited about 25,000
acre feet of sediment in the Cowlitz River channel (fig. 1.12). This elevated
the channel floor and decreased its cross-sectional area, making the valley
bottom prone to more frequent flooding (fig. 1.13). Furthermore, a significant
volume of sediment reached the Columbia River, where it created a shoal area
that blocked the channel used for shipping (see Lombard et al. 1981).
The Mount St. Helens catastrophe involved a number of process links
which demonstrate that the location of the dominant response shifted pro-
was
gressively downstream. In addition, the single driving force in this case
Idealized stream channel, Figure 1.13.

Toutle River Channel bottom and surface
elevation of the lower Toutle River
prior to and mudflows of
after the
May and post-
18, 1980. Pre-
eruption water surfaces are based
upon a flow of 38,000 cubic feet
9 per second- (From Lombard et al
U.S. Geological Survey Circular
850, 1981)
-, 100-
3
>
Q 90
o
*; 5
<D o
> 80 u
o
<u
T3
U 70
(1)
LU
123456789
Distance, in thousands of feet upstream of Cowlitz River
produced inside the Earth. The results of volcanic action, however, were quickly
transferred into slope processes (debris avalanche and mudflows) and from
there into fluvial, hydrologic, and lake systems and processes. Physical changes
even altered the biological balance.
The remarkable ability of process action to shift from one form to another
in response to a single impetus is a critical ingredient in process geomor-
phology. It often provides the only explanation for isolated or apparently un-
provoked geomorphic actions.
The Time Framework

we saw that geomorphology can be considered over dif-
Earlier in the chapter
which Schumm and Lichty (1965) called cylic, graded,
ferent time intervals,
and steady. The question before us now is which of these time spans is most
conducive to demonstrating the mechanics of process geomorphology?
To visualize the time factor more clearly, consider a hypothetical drainage
basin and the component subsystems (rivers) within it. If we observe a single
cross section of any river, we can describe the channel morphology at that
point by a group of parameters such as width, depth, slope, and shape. Any
single flow event in the river may bring about temporary scouring or filling,
Chapter 1 30
which causes immediate changes of the channel parameters but does not affect
them permanently; they will return to their previous state when the flow event
passes. Measurements taken an hour or day apart will show different values
for the variables, but they will always be internally consistent and apparently
adjusted to their external controls. Significantly, observed sediment and water
discharges through the cross section are dependent variables when viewed in
this time sense because they are modified by changes in the channel config-
uration. For example, if the river scours its bed, the moving load will increase
and the channel area will expand, allowing a greater volume of water to be
held within the banks. That is, both water and sediment discharge are tem-
porarily affected by the channel changes. However, we are observing the river
for only a very short period of time, and any equilibrium that we define is
almost instantaneous (or "steady" in the terminology of Schumm and Lichty).
Since no permanent changes can be expected on this temporal scale, we can
justifiably say that processes and landforms are time-independent when con-
sidered in this sense (see fig. 1.4).
Over longer periods, the steady-time channel measurements will vary with
mean values of sediment and water discharge; the statistical relationship be-
tween the external and internal variables defines the equilibrium state for
graded time (fig. 1.4). On this time scale (unlike steady time), sediment and
water discharge function as independent variables to determine the morpho-
logic and flow characteristics of the stream reach. In graded time, changes in
external variables due to modifications of climate or base level may require a

new set of equilibrium conditions within the channel. If and when thresholds
are exceeded, some significant change in the river is necessary and (again in
contrast to steady time) the adjustment will be permanent unless still another
change occurs in the external controls. Commonly the response takes the form
of channel incision or aggradation, pattern adjustments, or modification of
sinuosity.
The point is that the system is temporarily out of balance, and there exists
a certain time interval during which the river is approaching a new equilib-
rium state. The time involved is intermediate between long-term cyclic time
and instantaneous steady time. Absolute time in years is not the important
element in recognizing graded time as a valid geomorphic concept. Its signif-
icance lies in the fact that, once equilibrium is achieved, disruption of the
balance will be counteracted by each subsystem's ability to reestablish a new
equilibrium quickly (in the geologic sense); graded time thus does not involve
continuous and progressive change in the landscape. In addition, all parts of
the regional system may not be affected simultaneously or in precisely the
same way.
If we consider our basin over cyclic time, the balance (or temporary im-
balance) seen in steady or graded time becomes irrelevant. The inexorable loss
of sediment and energy from the basin suggests that the system must be con-
tinuously approaching, but perhaps never attaining, an ultimate equilibrium
condition. In this time framework, landforms within the basin should be pro-
gressively losing relief in phase with the deteriorating systemic energy.
Clearly a hierarchy of time exists; how one is able to interpret geomor-

phology depends, in a real sense, on which time scale is used. In process geo-
morphology, analyses are best made on a steady- or graded-time scale, assuming
landforms and processes to be time-independent or only temporarily time-de-
pendent phenomena.
In this chapter we have attempted to demonstrate that the process approach Summary
in geomorphology is more useful than other approaches because it relates better
to a variety of disciplines that examine phenomena occurring at the Earth's
surface. A set of basic principles constitute the framework of process geo-
morphology. Processes and the resulting landforms will be analyzed as bal-
ances between the driving forces (climate, gravity, etc.) and the resistance
offered by the geologic framework that makes up the Earth's surface. Pro-
cesses will be considered on short time scales rather than over geologically
significant intervals.
The following references will help you understand the approach used in this Suggested Readings
book.
Chorley, R. J., and Kennedy, B. A. 1971. Physical geography. London:

Prentice-Hall International.
Coates, D. R., and Vitek, J. D., eds. 1980. Thresholds in geomorphology.
London: Allen and Unwin Ltd.
Hack, J. T. 1960. Interpretation of erosional topography in humid
temperate regions. Am. Jour. Sci. Bradley Vol. 258-A: 80-97.
Schumm, S. A. 1973. Geomorphic thresholds and complex response in
drainage systems. In Fluvial geomorphology, edited by M. Morisawa,
pp. 299-310. S.U.N.Y., Binghamton: Pubs, in Geomorphology, 4th
Ann. Mtg.
. New York: John Wiley & Sons.
1977. The fluvial system.
Schumm, and Lichty, R. W. 1965. Time, space, and causality
S. A., in
geomorphology. Am. Jour. Sci. 263:110-19.

Climate and Internal Forces
2
I. Introduction B. Climate Change and
II. The Endogenic Effect Geomorphic Response
A. Epeirogeny 1. Sea Level Fluctuation
B. Orogeny and Tectonic 2. Geologic and Vegetal

Geomorphology Screens
C. Volcanism IV. Summary
III. Climatic Geomorphology V. Suggested Readings
A. Climate, Process, and
Landforms
33
Chapter 2 34
Introduction One of the major principles of process geomorphology is that spasmodic dis-
ruptions of equilibrium are significant components of surficial mechanics. In
many cases the cause of instability is external to the geomorphic system being
affected. In this chapter we will examine the primary external controls on

geomorphic systems.
In theory, exo genic processes uni mpeded by opposing forces will gradually
£ejducjLlh_eJan3scape to a rather dull, featureless surface with only minor to-
pographic inxg^Jlaritle^îrreffupnts sameness. However, as the Earth's sur-
face does possess relief, and has done so throughout geologic time, the Earth
must be constructed in such a way that exogenic processes are not always
preeminent. The surface is not a static environment but a locale where de-
nudational processes have been repeatedly counteracted as ne w mass is ele -
vated or crjeated_by_gndogen ic forces Each influx of mass not onlv jjrings-aith
.
it esc alated relief but also new potential energy thatjs ava ilable to ac celerate
or change the character of the exogenic processes. The ntroduction of new i
massa nd energy occurs primarily through volc anic activity and vertical uplift
.of the jand _surlaceT
endogenic input, we know from the glacial and
In addition to pulses of
Quaternary history that the climatic regime is not a
interglacial episodes of
static phenomenon Thus, although the equilibrium conHition in geomorphic
.
systems adjusted to climate, p hanges in cl imate ar e equally significant be-

is
cause th ey serve as spark plugs of te mporajxiBM&b jj'ty .
The importance of major climate change and significant endogenic input

is that they tend to occur as irregular events separated by long periods of rel-
ative constancy. This explains why these factors are able to create conditions
whereby threshold responses in geomorphic systems are required. Further-
more, periods of climate change and endogenic action are the ingredients
needed to decipher the sequential nature of geomorphology. Therefore, the
major external controls are the bases for Quaternary geomorphology, another
subarea of the discipline, in which the determination of history rather than
process is a common goal of geomorphic investigation. In this book we will
occasionally examine various aspects of how geomorphology may be used in
Quaternary studies. However, we will take these excursions into Quaternary
geomorphology cautiously so that we do not subvert our primary process theme.
The Endogenic Effect The fact that endogenic forces are an integral part of what we observe at the
Earth's surface is beyond question. Volcanic activity and seismicity are un-
mistakable manifestations of how internal energy and force can affect land-
forms and processes. Often, however, exertion of internal force results in equally
important though less dramatic geomorphic effects. For example, figure 2.1
shows that mostjand surface exists at twjjjg ersistent leêjs^jiicjiôrrespond

t o the continents and the oce an_floors. The EartjVs hypsometric curve (fig. 2.1)
also indicates the physiographic importance of continents and the ocean floor;
surfaces at elevations between sea level and 1 km account for 20.8 percent of
.
35 Climate and Internal Forces
Frequency \ Cumulative frequency

(percent of area) Figure 2.1.
(percent of surface area of world)
Distribution of areas of the Earth
20 40 60 80
standing at different elevations.
(A) Distribution frequency (B) The
Earth's hypsometric curve
200 300 400

Cumulative area higher than each level
(B)
the world's total area, and ocean bottoms between 3 and 6 km in depth com-
prise about 52.4 percent of the total surface area. The average elevation of
continents is about 0.875 km, and the average depth of the ocean floor is ap-
proximately 3.729 km.
The rocks th at make up continents are drastically different in miaexalogic
and cjiejmic^lxcunpxmtion from those that form the ocean basins. They are
ric her in silica and less dense than the ocean rocks so that c ontinental crust
,
is thic ker and stands higher than the o cean flnnrs Exactly how this monu-
mental difference between oceanic and continental crust developed is not cer-
tain because theories concerning the origin of continents lack hard data. The
reconstruction of events that led to the formation of protocontinents depends
entirely on indirect evidence of the Earth's early history. Nonetheless, the
maintenance of high continents and low ocean basins represents a balance per-
petuated by endogenic mechanics. In addition, as figure 2.2 shows, with the
exception of Antarctica, larger continents stand at higher elevations a fact ,
that must reflect some primary systemic operation. Our task, then, is to de-
cipher what internal process is responsible for this topographic phenomenon.
Epeirogeny
The processes by which the Earth's crust is deformed are collectively known
as diastrophism or tectonism. a nd are commonly divided into two distinct tec-
tonic styles. \[Orofenic) processes culminate in the forma tioj)_j3fj>tructural
^mountains Such mountain systems are typified by intense disruption of the
.
included rocks due to jolding a nd overthrusti ng; the effects are usually local-
ized in narrow, elongatedbelts. Tn contrast, ypeiroênic (processes cause uplift
or depressi on on a re gional scale and proceed without internal-disr uption of
original rock structures. Response to driving forces is rather passive in this
type of deformation. Although gentle tilting of strata may accompany the ver-
tical displacements, folding and thrusting are absent during the movement.
Chapter 2 36
Figure 2.2.
Relationship between area and
elevation of the continents
Asia
20 30 40 50 60 70 80100
Area (10 6 km 2 )
Mountain systems, however, are also affected by epeirogeny after the orogeny
that formed them has ended. Such vertical displacement of rocks and surfaces
is driven by the fundamental gravitational force.
In chapter we briefly examined the force of gravity and the various fac-
1
tors that determine its effect on a body at any location. The net force on any
mass is the vector sum of all gravitational attractions acting on it. Each body,
therefore, possesses a discrete amount of potential energy because mutual at-
tractions can be transferred into a kinetic form that is capable of doing work.
Unfortunately, all the attracting elements influencing any particular mass are
not applied in the same direction; thus, precise calculations of gravity should
be resolved into separate components operating in the orthogonal (.v. y, z)
directions. This complicated procedure can be simplified by viewing gravit)
as an energy field consisting of horizons of equal potential (t) in which the
attractive force is defined by
F = grad U.
In this field, F is everywhere normal to a series of surfaces, each of which

includes only points with equal potential and therefore a constant value of I
In such a model the value of F can be expressed in terms of energy, and its
magnitude, perpendicular to the equipotential surfaces, is
dU Gm
-2
dr
Geoich^,,
Figure 2.3.
Generalized relationship between
the spheroid and the geoid along
a line between east Africa and the
Pacific Ocean near Australia
Spheroid
where dil is the change in potential over the distance dr. By integration,
where U goes to zero when r is infinity.
In this model, sea level becomes an extremely important equipotential

surface in the Earth's gravitational field. Even though the surface may be
slightly distorted because of local factors, the inflections are small in ampli-
tude compared with the radius of the Earth and limited in areal extent.
The sea level equipotential surface is called the geoid, which on land is
usually defined by the water level in a series of imaginary canals cut through
the solid mass. The Earth's surface topography is referred to the geoid because
any elevation isdetermined by extending upward a succession of planes that
are parallel to sea level. At any point to be measured, a surveying instrument
is set tangent to the geoid, with the tangential line being the perpendicular to
a vertical plumb line over the site. Like any body, however, the direction of
the plumb bob itself is the vector sum of all the gravitational forces acting on
it and so may not be perfectly normal to the Earth's center of gravity. To
resolve this complication, geodesists utilize a second surface called the spheroid,
which is a mathematical representation of sea level with all irregular influ-
ences removed. Essentially, the spheroid is the hypothetical sea level surface
of an Earth with no lateral variations in density or topography and with a
vertical change in density that is uniform from the center of gravity to the
surface. With such a mass distribution, gravity would vary consistently from
pole to equator, and its theoretical sea level values can be easily calculated.
The differences between the predicted values of gravity, calculated under the
above assumptions, and the actual measured values are called gravity anom-
ali es; they ind icate the depar ture of t h e geoid from the spheroid (fig. 2.3).
Because few~gravity measurements are made at sea level, most observa-

tions must be reduced into separate components indicating the portion of the
measurement produced by mass and the portion due to distance. Corrections
for each must then be made before a gravity measurement can be compared
with the spheroidal value. There is a general tendency for anomalies produced
Chapter 2 38
by mass (Bouguer anomalies) to be strongly negative in mountains and in-

creasinglymore negative with higher elevations, demonstrating a most im-
portant principle in geomorphology — that s urface to pography is somehow
relaXecLlo the int ernal dis tribution o f mass!
The is influenced by the distribution of mass within
idea that topography
the Earth is was expressed by Leonardo da Vinci, and its concrete
not new. It
formulation as a hypothesis arose from analyses of data obtained in the mid-

nineteenth-century land surveys in India. C. E. Dutton, working in the Col-
orado Plateau, introduced the term is ostasy to define the internal process in-
volved. In essence, the result of isostatic adjustment is a condition in wh ich
largjêjcvate d regions such as continents or mo untain r anges are com pensa ted
.by amass deficiency in the crustal rocks benejjJhjtherrL The process oflsostasy
requires that atsome depth beneath~seaTevel the pressure exerted by overlying
columns of rock will be the same, regardless of how high the various columns
stand above sea level. Mountains, ocean basins, shields, etc., are balanced with
regard to the total mass overlying each area at some internal level called the
Lj,epth of compensation. This isostatic equilibrium is probably controlled by
lateral variations in rock density (Hsu 1965; McGinnis 1966) and maintained
by internal adjustments that are not clearly understood, and in fact, the me-
chanics may differ depending on the scale of topography being supported at
the depth of compensation (Officer and Drake 1982).
A correction of measured gravity due to isostasy can be calculated, and
when this is subtracted from the Bouguer anomaly, a residual value called the
isostat ic^ flflom o/^jemains. If th e isostatic an omal y is zero, the system is in
perfect bal ance.. We know, however, that the equilibrium cMdTHonTs~eaîty

upset so that negative or positive anomalies are not unusual. A negat ivgjso-
static anomaly indicates deficiency of mas s_at the loca lity, and the surface
s hould have a tendency t gjjs_g_b ecause more matter must be added _at_depth
t o establi shth e equi hbjliirrLState. Similarly, positive isostatic anomalies should
portend sinking since they indicate an excess of mass beneath the surface.
Because most topographic blocks, local or regional, are not perfectly
equalized with respect to one another, vertical movements of crustal segments
are inherent in the attempt to establish equilibrium. When isostatic balance
is disrupted by erosion, thick sediment deposition (example. Lake Mead), or
tectonics, a counteraction by isostasy is required to restore equilibrium. It is
also known that the accumulation of massive glaciers is accompanied by

depression of the surface; conversely, when the weight is removed as the ice
disappears, the surface will rise to reestablish the isostatic balance. This re-
sponse to glacial and interglacial conditions is called glacio-isostasy and rep-
resents one of the most important geomorphic processes in high latitudes of
the world (Andrews 1974).
The significance of this is that isostasy is the endogenic process that causes
epeirogenic diastrophism and is responsible for maintaining the topographic
relationship between large blocks of the Earth's crust. The relief between ocean
basins and continents probably reflects the isostatic balance established be-
cause the crustal thickness and density of rocks underlying the two areas are
different. The same process, however, functions on scales smaller than conti-
nents or ocean basins.
The fact that isostasy works is uniquely important in geomorphology be-
cause its balancing act requires vertical motion of the Earth's surface. When
the moveme nt u pward, isostasy produces potent ial en ergy that is availabl e
is
for_use in exogenic processes Geologists intuitively recognize this effect be-

.
cause enormous ra tes of depo sition hav e been observed in the geologic record
d uring and following a m
ajormou ntain-building episodeTwhat is difficult to
ascertain from these events, however, is how much uplift of the surface is as-
sociated with the diastrophism. The enigma is due to the fact that the geologic
time scale lacks the precision needed to calculate a meaningful uplift rate,
and the geologic evidence is often ambiguous. For example, assume that a
shallow-water marine deposit containing fossils of Late Pliocene age is now
exposed at an elevation of 3000 m. We may state confidently that 3000 m of
uplift has occurred since the Late Pliocene, even though we cannot be certain
how far below sea level the animals lived or whether the Pliocene sea level
was exactly the same as sea level today. More troublesome is the question of
how to translate the assumed 3000 m uplift in terms of rate with units of
velocity such as m/yr. Two problems are inherent in such a calculation. First,
how many absolute years are represented by the Late Pliocene age of the ma-
rine fossils? Is the deposit 2 m.y. old or perhaps 5 m.y. (for example, see Yeats
1978)? Obviously the calculated rate of uplift depends on one's interpretation
of an inadequately known time interval. For the second problem, assume that
an absolute age of 2.2 m.y. is given for the deposit by radiometric dating of
interbedded volcanics. Here again large plus-or-minus errors must be consid-
ered. But assuming the date is accurate, the calculated rate of uplift then be-
comes 3000 m in 2.2 m.y. or .00136 m/yr. Although the numbers may seem
precise, they say very little about the actual rate of uplift because there is no
way of knowing whether the vertical motion was continuous over the entire
time interval or whether it occurred in one catastrophic spurt sometime be-
tween the Late Pliocene and the present. The calculated rate assumes a con-
stant and continuous uplift for 2.2 m.y. and so is a minimum value. If the
entire uplift was accomplished during a limited segment of the total time, the
rate may have been much higher.
In spite of the problems discussed above, it may be possible to determine
precise uplift rates if the environmental setting is proper and the time span
being considered is relatively short. This is especially true where glacio-iso-
static uplift has affected coastal regions (Andrews 1974; Ten Brink 1974; Hil-
laire-Marcel and Fairbridge 1978). For example, Ten Brink (1974) was able
by correlating the ,4 C age of marine
to derive detailed uplift curves (fig. 2.4)
shell samples with the elevation of the strandlines in which they were pre-
served. Strandlines are essentially old shorelines formed by bodies of water,
such as a lake or ocean, that are now elevated above the present water level.
In Ten Brink's study, the oldest strandlines, dated by the related fossils, are
higher than subsequent strandlines that formed as the Greenland coast emerged
during deglaciation. The curve shown in figure 2.4 is a clear indication of how
uplift proceeded in the postglacial adjustment.
Chapter 2 40
Figure 2.4.
Upliftcurve along the Greenland
coast based on 14 C dates of
140
marine fossils in emerged
strandlines
120
100
2000 4000 6000 8000

Radiocarbon years bp
Meaningful uplift rates are important in geomorphology because they de-

termine whether a surficial system can remain in equilibrium during the uplift
event. If rates of uplift exceed by far the prevailing rates of denudation, the
system will cross a threshold and enter into disequilibrium. How long it will
take to establish a new balance depends on how radical the difference is be-
tween the rates of uplift and of denudation. Schumm (1963c) discusses the
complications involved in making such an analysis because although uplift
rates are significantly higher than denudation rates, the uplifts occur in short,
spasmodic bursts rather than as long, continuous events.
Because of the problems inherent in determining a precise uplift rate from
long-term data, table 2.1 is presented to demonstrate rates of vertical dis-
placement measured in the modern setting or based on data from late glacial
and postglacial times. The data presented were chosen at random and should
not be considered complete. Some of the represented movement is caused by
isostatic adjustment to the unloading of ice and/or water or the loading of
sediment and water. Other motion is generated in regions that presumably are
experiencing the effects of active tectonism. On the basis of this limited sample,
it is interesting to note that the rates of vertical displacements are high re-
gardless of the tectonic environment. For example, the uplifts in Fennoscandia
Table 2.1 Rates of vertical d splacement.
Area Time Rate (cm/1000 yr)
Fennoscandia '
Recent ( + 1100
)
Hudson Bay12 '

Recent ( + 1700
)
Lake Superior111 Recent ( + )500

Lake Mead 13 '
Modern (-) 1200
Lake Bonneville' 4 '
Late glacial- -Modern (+) 1200-60 3
East Coast U.S. ,5) Modern ( + )200 - (-)500
California Coast Ranges 16 '
Late glacial- -Recent (+) 500-800
Alaska' 7 Modern ( + )2400
'
Greenland 18 '
Modern (+) 1000-1700
(1)Gutenberg 1941 (2)Walcol1 1972 (3) Longwell 1960 (4) Crittenden 1963 (5) Fairbndge and Newman 1968 (6) Bandy and
Manncovich 1973 (7) St Amand 1957 (8) Ten Brink 1974
^Represents declining rates during last 20,000 years
and central North America, undoubtedly isostatic rebounds initiated by de-

glaciation, are within an order of magnitude of those in the California and
Alaskan regions, which are probably undergoing active deformation. It also
appears that all vertical movement is considerably greater than maximum rates
of denudation, which will be discussed later. The values presented in table 2.1
are also complicated by the fact that rates for isostatic rebound decline pro-
gressively as the region gets closer to equilibrium. They will be very high im-
mediately following the removal of the excess weight of ice or water but may
be very low when equilibrium is nearly reestablished (see Gutenberg 1941;
Crittenden 1963). Schumm (1963c) suggests that a reasonable average rate
of uplift is about 750 cm/ 1000 yr.
The significant geomorphic aspects of isostatic uplift can be summarized

briefly:
1. Almost all regions within the continents are in or near some form
of isostatic equilibrium.
2. Structural features and initial relief are formed by vertical
epeirogenicmovements associated with density variations in the
basement rocks. Presumably postorogenic uplift in mountains is
also related to isostatic compensation.
3. Redistribution of mass at the surface by erosion and deposition,
glacier development, thrusting, etc., requires vertical movement of
the underlying rocks to reestablish isostatic equilibrium.
4. The driving force behind isostasy is gravity, which is responsive to
a heterogenous distribution of rock density. The cause of density
variations and the precise mechanics of isostatic compensation are
very poorly understood, but the process and its effect on
geomorphic systems are real.
5. Rates of uplift are normally high compared with rates of

denudation.
.
Chapter 2 42
Orogeny and Tectonic Geomorphology

Scientists have pondered the origin of mountains ever since they first recog-
nized that rocks in mountain belts were structurally different from those in
other areas. The intense folding and overthrusting displayed within moun-
tainous regions led geologists to realize that significant crustal shortening was
involved in their formation, but learning what caused the deformation was
hindered by our ignorance of the Earth's interior. Initially, a progressively
cooling and shrinking Earth was suggested to explain the needed congres-
sional stress, but this idea was rejected after the perception of continuous ad-
by radioactivity and the acceptance of a cold origin for
dition of internal heat
the Earth. Other proposals met an equally unsatisfactory fate. The advent of
the plate tectonic theory, however, forced geologists to reevaluate mountain
building in light of the new global model. A detailed discussion of plate tec-
tonics isbeyond the scope of this book, and the basic concepts of the plate
model can be found in any textbook of physical geology. For our purposes it
is sufficient to say that mountains and ocean features such as island arcs and
trenches are intimately associated with the seismicity and volcanism found at
plate margins where the lithosphere is being actively consumed.
The effect of orog eny oneeomorphic process is less jaagihlp thanjj^L
produced by epeirogeny be£aiis^jJie_j^t^fôuntairi-bjii lding occu rs_below
sea level. Thus, it is not the rocks or region squeezed in the mountain-building
evenTfhat shows geomorphic response during the application of orgenic forces.
As indicated earlier, these regions are usuall y-affect ed aftgr orogen y is ove r,
when isostasy raises the thick pile-oXJow^-density and highly deformed rocks.
NonetKeTessTge^morphic systems adjacent to the orogenic zone may be thrown
into disequilibrium during the orogeny. The driving forces_that cause the
changes are found_jn_sgcondary phenomena that are ancillary to the orogenic
force. The most prominent of these are seism icity_and volcanism.
Seismic activity transfers internal energy into exogenic process in several
ways. First, earthquake motion increases driving force while simultaneously

reducing the resisting strength of materials. This commonly results in sudden
slope failures (landslides) that represent threshold responses to seismic ac-
tivity in regions close to an active orogenic zone (see Garwood et al. 1979).
Second, movement along faults that cause the earthqu akes_sometimes dis-
places the surface;upward displacement produces the sarjie~g£ojnor phic re -
sponses asjhose occ urring in ep eirogenic movement, but they a re usually less
jdramaiic_-
The fact that deformation leaves an imprint on landscapes is one of the
oldest tenets in geomorphology. Initially, regional tectonics were used to ex-

plain diversity of character in large-scale topography. For example, geomor-
phologists recognized that the block mountains and intervening basins of the
Basin and Range province in the western United States, and the sinuous val-
leys and ridges in the Folded Appalachians of Pennsylvania were reflections
of significantly different tectonic styles. Such observations are interesting in
themselves, but the utility derived in demonstrating the relationships is pri-
marily physiographic and does little to further our understanding of process.

fe<^
More recently, the relationship between form and tectonics has become the
basis for much more detailed and sophisticated geomorphic studies and has
led to the development of a special branch of the discipline known as tectonic
geomorphology. Tectonic geomorphology as now practiced deals with ho wjec-
tonic activity affects jSrocesses and morphology in geomorphic systems and,
conversely, how technically controlledTandf orms can be used" to assess Qua -
t ernary tectonic activity (Bu and McFadden 1977). Most studies have been
ll
'directed at understanding the geomorphic response to vertical and/or hori-

zontal displacement along faults and the tilting or warping associated with
broad many studies landforms are used to provide in-
uplifts. Significantly, in
sight into the styleand rate of tectonic processes, therefore revealing a shift
in emphasis from the explanation of form to the analysis of what form tells
us about process.
Tectonic geomorphology has been used for a variety of scientific purposes,
and studies range in scale from contrasting the tectonic style of Earth, Venus,
and Mars (fig. 2.5; Arvidson and Guinness 1982) to the analysis of a single
landform such as an alluvial fan (Keller et al. 1982). The approach has tre-
mendous implications in environmental planning (Schowengerdt and Glass
1983), especially with regard to landform stability (Bull and McFadden 1977),
seismicity and earthquake risk (Bull 1974), and prediction of fault-movement
Figure 2.5.
periodicity (Wallace 1977, 1978; Buckman and Anderson 1979^Na'sh 1980;
Tectonic features on Mars. Linear
Colmanand Watson 1983). depressions on left of photo are
interpreted as large grabens.
Chapter 2 44
Geomorphic response to faulting constitutes one of the basic components

in tectonic geomorphology. Faulting produces adjustments in process and form
that can be used as clues to present and past tectonic activity and, additionally,
to the rates and magnitudes of displacements. For example. Bull and Mc-
Fadden ( 977) were able to characterize the activity of Quaternary tectonism
1
in subareas associated with the Garlock fault in southern California. The de-
gree of tectonic activity was recognized on the basis of two geomorphic pa-
rameters: ( ) the mountain-front sinuosity, which is the ratio of the length
1
measured along the junction of the mountain and piedmont to the total straight-
line length of the mountain front, and (2) the ratio between valley floor width
and valley height in major river valleys measured at a given distance upstream
from the mountain front.
Theoretically, a tectonically inactive region shows mountain fronts with
high sinuosity because continuing erosion and deposition would progressively
shift the position of the mountain front away from the range-bounding fault.
Valleys in the mountains should be wide and shallow under tectonic stability
because lateral erosion controlled by a stable base level would dominate the
system. Conversely, in tectonically active areas the mountain-front sinuosities
are low and the mountain valleys are narrow and deep. This follows because
repeated vertical movements preserve the coincidence between the bounding
fault and the mountain front, and the rivers upstream from the front are spurred
into downcutting on the upthrown side of the fault. The basis for this expected
response has been established by countless observations that river gorges are
related to enhanced stream power brought on by rapid, fault-related uplift
(for example, see Jackson et al. 1982).
Horizontal displacement on faults can also be documented by variations

in geomorphic landforms. Keller et al. (1982), for example, recognized nu-
merous tectonically produced landforms along a branch of the San Andreas
fault in the Indio Hills of southern California. Offset drainages, stream di-
versions, and an alluvial fan-pediment complex demonstrated approximately
0.7 km of horizontal displacement. Other features were recognized as prod-
ucts of vertical displacement along the fault. Perhaps the greatest importance
of this study is the fact that Keller and his colleagues were able to explain
almost all of the fault-related features by simple shearing associated with a
bend in the trace of the fault. Therefore, landforms here were used as primary
evidence to explain the mechanics of local deformation.
Evidence of regional, large-scale tilting and warping are also present in
the elements of landscapes, especially in the profiles of planar surfaces or stream

channels. Marine terraces and wave-cut platforms (fig. 2.6) are excellent in-
dicators of these tectonic movements because they originate as nearly hori-
zontal surfaces related to sea level (Bradley and Griggs 1976: Pillans 1983).
Thus, any divergence from their original state is a manifestation of tectonic
influence. In both studies referred to above, the researchers were able to dem-
onstrate distinct shore-parallel warping of the coastline based on differential
uplift of the platforms and terraces. In addition, seaward tilting of the planar
features and associated river terraces indicates that the tectonic motions are
domal type regional uplifts.
Longitudinal profiles of river terraces, alluvial valley floors, and stream

channels are also sensitive indicators of regional tectonism. For example, re-
peated geodetic surveys have shown clearly that two areas in Louisiana and
Mississippi are sites of rapid modern uplift. Burnett and Schumm (1983) ex-
amined the characteristics of fluvial features crossing these uplifts (fig. 2.7)
and found considerable geomorphic evidence to substantiate the geodetic find-
ings. Terrace and valley floor profiles show pronounced convexities where they
cross the axes of the uplifts, being abnormally gentle upstream from the axes
and oversteepened downstream from the axes (fig. 2.7). In addition, river
reaches upstream and downstream from uplift axes develop markedly dif-
ferent channel patterns, depths, and sinuosities, which are probably controlled
in part by the orientation of the underlying bedrock structure (Gardner 1973).
Precisely how much response will occur in modern rivers experiencing neo-
tectonism depends on the river size. Larger rivers may possess enough energy
to keep degradational pace with the rate of uplift, and therefore, they may
experience little change. Their valley floors and terraces, however, may show
considerable change in character at the uplift axes. Nonetheless fluvial fea-
,
Figure 2.6.
tures clearly refle ct neotectonism an d should be considered as prime criteri a
Uplifted marine terraces (wave-cut
in tectonic geomorp hojogy"
'
platforms) along the central

California coast
Chapter 2 46
Terrace profiles
Figure 2.7.
(A) Longitudinal profiles of
Floodplain profile
terraces, floodplam, and a Projected channel profile
projected channel profile of the
Pearl River, a major river crossing
the Wiggins uplift (B) Longitudinal
valley profiles of streams crossing
the Monroe uplift Macon Ridge, a
remnant of a deformed 2
Pleistocene Mississippi River
terrace, is also shown, MSL. mean
sea level
40 30 20 10 90 60 30 30 60
(A) Valley distance from mouth (km) (B) Valley distance from uplift axis (km)
The selected cases just presented are only a few examples of how the re-
lationship among landform, process, and tectonics can be employed to gain
insight in numerous geological and environmental investigations. The practice
of tectonic geomorphology will certainly increase in the future, and every geo-
morphologist should become familiar with its basic premises and uses (for an
excellent review see Bull 1984).
Volcanism
Although some (but not all) volcanic activity is associated with orogenic events,
volcanism is such a major input of endogenic force that it deserves separate
treatment. Volcanism is nothing more than a surface manifestation of the in-
ternal processes that create and mobilize magma. Although volcanoes are
spectacular in eruption and unique in topographic form, describing examples
of active volcanoes is not necessary here. Most physical geology texts do this,
and excellent treatises on volcanoes and volcanic landforms are available (Bul-
lard 1962; Oilier 1969; Green and Short 1971; Macdonald 1972). It may be
pertinent, however, to examine briefly how internal variables control the type
of volcanic eruption and the ensuing topographic form.
The violence of a volcanic eruption is determined mainly by the compo-
sition of the magmaandjjie^momuof gasjjT_JL_£\_affecting the viscosity of
the magmaTthese factors influence the observed differences between conti-
nental and oceanic volcanism. Highly fluid, basaltic magmas are produced in
both continental and oceanic environments. JJiejriore visc ous, higlv silica lavas
and rhyolites arejêngrajl y restr icted to
that crystallize as andesites, dacites,
where most viscous Java
tinents or margln^TTstaTtrfarcs. In orogenic zones,
occurs, the eruptinglna^ma~atso contains more gas then do the true oceanic
types. The combined effect of higher gas content and greater viscosity creates
a tendency toward more explosive eruptions. This trend is magnified in con-
tinental volcanoes where yiolent explosions are commonplace, and solid vol-
d {e£hra
canic ejecta. called_[££/rra.
; . is the dominant extrusive material rather than
flowing lava.
The major volcanic landforms consist of three types: plateaus okfpfetJg^

cones, and calderas. Lav a plains and plateaus are extreme lyfla t surfaces, both
continental and oceanic, that have been_ aggraded by overlappi ng, flow s of flniH
lava with a mafic composition. Although the nature of the vents is not clear,
they probably are fissure type sjii stributed over wide areas or a series of u n-
connected pipe vents. The dominant characteristic of lava plains and plateaus
is the enormous volume ofjava extruded and spread nypr a vq st surface area.
For example, Columbia River Plain of Washington and Oregon, areas
in the
greater th an 2 00,000 km 2 are known, and in the Deccan Plain of India, more
than_50QjlQ0. km 2 Oceanic plains can be even more extensive (Kuno 1969).
.
Commonly the total thickness of the extrusive rocks exceeds 2000 _m great —
enough to bury a mountainous terrain. In fact, in the Columbia River Plain
peaks of buried mountains (called steptoes) may project through the flat sur-
face as isolated "islands" of older rock.
The surfaces of lava plains and plateaus show minor perturbations where
broad, low cones rise 30-60 m above the general level. However, individual
flows, 2 to 50 m thick, can extend for hundreds of kilometers from the cones,
blending imperceptibly into lava issued from fissured vents to create a surface
that normally slopes less than 1 °. Lava plains are almost always composed of
basaltic rocks because silicic magma tends to be too viscous for the long dis-
tance of flow necessary to create a planar topography.
Some do underlie flat plains, but these rocks did not
silica-rich volcanics
crystallize from lava and the surface extends over much smaller areas. Silicic
plains develop when incandescent volcanic glass is erupted within dense clouds
of gas capable of flowing for considerable distances. The welded tu ffs (ignim-
brites) resulting from the ash flows are com mon in t he cordilleran of the United
States, especially in the southwest. Field evidence there suggests that the ash
was vented from both linear and arcuate fissures and central pipes.
The second major volcanic landform occurs as the topographic expression
surrounding a s i ngle vent These surfaces, c alled cones, have a variety of shapes
.
depending on the type of volcano and the predominant mode of eruption. Shield
volcanoes devel op cones that are typical of t hose seen in Hawaii apii-Iceland.
The sides of these" cones slope trom 21_tQ_lP_° and m er ge gradually, a t_their
base, with the adjacent oc fean_floor (fig. 2.8). The massive Hawaiian shields
rise at least 4800 m above their bases, and some (Mauna Loa, Mauna Kea)
have an absolute relief greater than 9 km, making them among the largest
topographic mountains of the world.
Shield volcanoes are always built up from fluid, basaltic magma; tephra
is only a minor part of the erupted material, and explosive eruptions are rare.
The magma in moves surfaceward in rift zones that gen-
the Hawaiian chain
erally parallel the Hawaiian Ridge (topographic high in the central Pacific).
The rift zones usually contain hundreds of fissures that the ascending magma
uses as vents for eruption. The Hawaiian rifts extend discontinuously for 3500
km across the center of the Pacific plate to a point where they abruptly bend
to join the Emperor Chain. This linear group of seamounts, similar in all as-
pects to the Hawaiian volcanoes, continues northward for another 2500 km.
Chapter 2 48
The Hawaiian eruption rates, estimated between 0.05 and 0.1 km a year, 1
(Moore 1970; Swanson 1972) are the greatest known on Earth. There seems
to be little question that the primary Hawaiian magma originates in the upper
mantle (Eaton and Murata 1960) and, furthermore, that no genetic relation-
ship exists between the volcanoes and the age or structure of the adjacent sea
floor (Dalrymple et al. 1973). The plate-interior location of the island chain
precludes an orogenic origin for the volcanic activity. This apparent lack of
coincidence between plate tectonics and volcanism has prompted the hypoth-
Figure 2.8. esis that volcanic action here occurs over a roughly circular hot spot in the
Profile and crater area of shield
volcanoes, Hawaii Summit
upper mantle, about 300 km in diameter. Volcanism occurs as the Pacific plate
caldera of Kilauea volcano in rides over the heated region, and successive generations of volcanoes are car-
foreground. Mauna Kea in
ried away from the heat source.
background shows gentle slopes
typical of a shield volcano.
'A
J
The Icelandic shields commonly are smaller in all dimensions than the
Hawaiian volcanoes. Although they originate in a plate-margin environment,
they are not related to orogenic events. The magma probably rises along the
mid-Atlantic ridge in tensional fractures associated with the forces that move
the oceanic plates apart.
Most other volcanoes take the form of a cojnijosite co ne, so named be-
cause the cone is constructed of inl£rj)edded lava flows anrjja yers of t ephia-
The fl ows and the tephra most ly cinder or ash, but the
a re usually bloc ky
characteristics of components may vary with the viscosity of the ascending
all
magma. Composite cones, like the one in figure 2.9, are more peaked than
shield volcanoes, commonly ris^mTseveral-thousand meters ab ove a narrow Figure 2.9.
circular base. Most develop from a single pipe vent. Side slopes are typically Steep slopes of composite
volcanoes Mount Rainier, Wash
in foreground
t
Chapter 2 50
high, r anging fromlO° to 35°, because the laj^kjofjjuidityjjulie^^

to a more localized deposition. Most of the world's continental volcanoes are
"oTThis typeTin North America famous examples are Mt. Rainier (Wash-
ington), Mt. Shasta (California), and Mount St. Helens (Washington).
The major landform associated with volcanic activity is known as a
third
caldera. Calderas are large depressions in volcanic regions that result where
j;ruption spews forth large q uantities of material, creatin g an empty sp acejn

the und erlying magm a chamb er. Thi s result s_in_an inwajj_gQJla_pse of the uppgr
part of Tn ejaalcjurKTcone, often along fractures that develop following the
eruption. The caldera depression, therefore, has a larger diameter than the
original crater, and in fact, the primary distinction between a normal volcanic
crater and a caldera is size . Depressions lar ger than_ 1.6 km (1 mile) in di-
a meterare u sually accepted as being ca lderas (Macdonald 1972) because they
almost certainly result from subsidence rather than incremental construction
of a cone.
The largest calderas are associated with volcanoes that produce tephra
sheets. For example, the caldera basin buried under the thick ignimbrite sheets
of the Yellowstone Plateau is 70 km by 45 km. In addition, some of the most
well-developed calderas, such as the one occupied by Crater Lake. Oregon,
occur on the summits of composite cones, and their origin is directly associated
with the occurrence of widespread ash falls.
Volcanic activity produces a myriad of smaller landforms that, like the

major forms, also relate to the viscosity of the involved magma. The interested
reader is referred to Rittmann (1962) for discussions about the classification
and origin of these forms.
Climatic Climate, Process, and Landforms

Geomorphology Because of climate's vast effect on many surface phenomena, it was inevitable
that scientistswould attempt to group climates into some useful classification.
Temperature and precipitation commonly are the prime variables used to de-
fine climatic boundaries, although vegetation and soil assemblages may serve
as auxiliary factors. The most successful classification is an empirical system

devised by Koppen in which five major climate groups are distinguished on
the basis of observed temperature and precipitation values. Each major group
can be divided into progressively smaller units with local variations such as
seasonality. The Koppen approach has the great advantage of relying on mea-
sured data; climates can be defined by precise physical characteristics, and
maps based on this system can provide rather detailed information about local
or regional climates. The Koppen classification, however, tells us nothing about
what causes the climate and therefore provides no genetic information. More
recent attempts at classification employ variables that indicate the origin of
climatic zones rather than simply placing them in numerical pigeonholes
(Strahler 1965). Geomorphologists should be more concerned with how well
energy in the climatic regime is utilized in geomorphic work than with how
the climate is classified. In other words, we need to know which landforms will
develop most efficiently under any prevailing climate.
The relationship between landforms and climate is the basis for a major
philosophic approach in geomorphology known as climatic geomorphology,
which has been most forcefully championed by European scientists (Tricart
and Cailleux 1972; Biidel 1982). Essentially, the underlying premise of this
approach is that geomorphic mechanics vary in type and rate according to the
particular climatic zone in which they function. If that assumption is correct,
landforms produced from these mechanics will be different from region to re-
gion and will reflect the dominant climate.
Biidel (1982) suggests that modern processes in mid-latitudes are too weak
to remove features that formed prior to the Holocene. As a result, 95 percent
of the topography found in those regions developed under a former, more rig-
orous climate, and landforms there are not related to modern processes. Thus,
the framework of climatic geomorphology becomes evident only if we consider
landforms that developed from processes operating in climatic zones estab-
lished during the Holocene. If that basic rule is followed, landforms will be
uniform over large areas called morphoclimatic zones, and their character
willchange drastically as climate varies from one zone to another. Forms that
do not fit in a specific zone are considered as relict features that developed
under climatic conditions that no longer exist in the region.

Although most geomorphologists recognize some validity in the tenets of
climatic geomorphology, the approach suffers because we do not yet under-
stand how landforms are affected by climate changes of different magnitude
and/or frequency (Stoddart 1969). In addition, the approach has no process
base, and its practitioners have not applied experimental methods to climo-
genetic problems (Derbyshire 1976). Thus, the descriptive and qualitative as-
sociations between climate and topography proposed in climatic geomorphology
have simply not been adequately tested.
In spite of such problems, there is little doubt that some relationship exists
between the type of dominant process in a region and the prevailing climate.
This relationship has been expressed by Peltier ( 1950), who designated certain
areas as morphogenetic regions. In these regions, dominant processes can be
ascertained and related to a specific climatic regime that is broadly defined
by its prevailing temperature and precipitation. This concept has been fol-
lowed by others, although the climatic variables used to distinguish the regime
are sometimes different. L. Wilson ( 1 968) objects to the term "morphogenetic
regions" because it usually has been used to refer to a process and not to a
tangible region of the earth. He suggests that the relationship between climate
and process be called climate-process systems, and the relationship between
climate, process, and landforms be called morphogenetic systems. The term
"morphogenetic region" would be restricted to the case where an actual region
is being considered.
Regardless of semantics, it is geomorphology to establish the
valuable in
relative importance of a particular process under a set of temperature-precip-

itation conditions. The concept is still useful even when it is recognized that
landscapes are controlled by factors other than climate and that the details of
morphogenetic systems are poorly understood. It serves as a viable first ap-
proximation of what landforms we can expect to find in any environment. As
Chapter 2 52
Mean annual precipitation (inches)

Figure 2.10.
20 40 60 80 100
Six possible climate-process
systems as suggested by Wilson
(1968) Each set of temperature- Selva
precipitation values tends to drive
processes that function most
under those climatic
efficiently
conditions
I I
1000 1500 2000

(millimeters)
Biidel (1982) would suggest, it is helpful in detecting relict features, forms

that developed under an earlier and different climate and have not yet ad-
justed to the present climatic regime. For example, moraines in central Illinois
reflect a morphogenetic system that no longer exists; the landforms have not
completely readjusted to coincide with modern conditions.

Figure 2.10 is a diagrammatic representation of six climate-process sys-
to a set of paramount processes. The diagram
tems that ideally relate climate
was derived by combining the relationships of several specific processes and
climates; the arbitrary system names are meant to be descriptive of the cli-
matic type. For instance (other factors being equal), we can expect that a
region with a mean annual temperature of 0°C and mean annual precipitation
of 500 mm will function as a periglacial climate-process system, and those
processes that function most efficiently in such a climate will prevail. Each
climate-process system in the figure can be converted into a morphogenetic
system by defining the landforms that most commonly result from the pro-
is done in table 2.2.
cesses involved, as
minor heterogeneity of geology may produce subtle variations
In reality,
in the topographic outcome. In addition, relict landforms, and seasonally
changing climates may complicate the expected form. Nonetheless, the mor-
phogenetic approach does provide a reasonable framework for the analysis of
landforms. Absolute servility to this approach, however, leads us away from
the study of processes, since the method passes quickly from climate to form.
This nonchalant leap from climate to landform does not mean that we un-
derstand the intermediate phase of process.
In addition to isolating the predominant processes, it is important to un-
derstand how climatic variables influence process mechanics. In most ca
solar energy must be converted into another form of energy to do geomorphic

work. For example, although precipitation releases stored heat, it is not the
temperature of raindrops that drives hydrologic processes. Evaporation and
Table 2.2 Simple morphogenetic systems and their landscape characteristics
System Name Equivalent Dominant Landscape

Koppen Geomorphic Characteristics' 7
3
Climates Processes"
Glacial EF Icecap Glaciation Glacial scour

Nivation Alpine topography
Wind action Moraines, kames, eskers
(freeze-thaw)
Penglacial ET Tundra Frost action Patterned ground

EM Solifluction Solifluction slopes, lobes,
D-c Humid Running water terraces
microthermal Outwash plains
Arid BW Desert Desiccation Dunes, salt pans (playas)

Wind action Deflation basins
Running water Cavernous weathering
Angular slopes, arroyos
Semiand BS Steppe Running water Pediments, fans

(subhum d) Cwa Tropical Weathering (especially Angular slopes with
savanna mechanical) coarse debris
Rapid mass movements Badlands
Humid Cf Humid Running water Smooth slopes, soil

temperat e mesothermal Weathering (especially covered
D-a chemical) Ridges and valleys
Creep (and other mass Stream deposits extensive
movements)
Selva At Tropical Chemical weathering Steep slopes, knife-edge

Am Monsoonal Mass movements ridges
Running water Deep soils (laterites

included)
Reefs
Copyright © 1968 by Dowden Hutchinson & Ross Inc
Equivalents are approximate and not exhaustive

^Processes are listed m order of relative importance to landscape (nol in order of absolute magnitude) List is abbreviated
c Both erosional and depositional forms are included List is neither comprehensive nor definitive merely suggestive
rising air provide water vapor with potential energy that is transformed to
kinetic energy when rain begins to fall. A maze of reactions changes energy
into various forms that ultimately drive the different geomorphic processes.
Climate Change and Geomorphic Response

Regardless of the obscure relationships existing among climate, process, and
landforms, the fact that climate changes is an irrefutable cornerstone in geo-
morphology. In modern times, for example, we know from direct observation
that average air temperatures in the Northern Hemisphere increased from the
late1800s to 1940 and since then have been steadily decreasing (Budyko 1977;
Lockwood 1979). Additionally, overwhelming evidence of past climate change
has been documented in such diverse fields as pedology (soil science), paly-
nology (study of pollen), archeology, paleontology, oceanography, and many

others. The most dramatic manifestation of climate change is, of course, that
Chapter 2 54
several widely spaced episodes of prolonged glaciation have occurred in the

Earth's history. The most recent of these, encompassed by the Pleistocene
epoch, was characterized by distinct alternations of glacial and interglacial
stages when climate was patently different from the present. Thus, the ephem-
eral nature of climate is beyond question. The reasons for climate change,
however, are open to considerable debate because they are tucked away in the
complex maze of interactions that control climate itself.
Detailed examination of the dynamics of climate change are well beyond
the purposes of this book and are best left to the climatologists. In fact, some
climatic variations are presently inexplicable. In general, however, the vari-
ables that trigger climate change exist in several major groups; significantly,
the variables in each category seem to function over different time intervals.
First, variations in the composition of the atmosphere may alter the amounts
of solar radiation reaching or escaping the Earth's surface. This is especially
truewhen changes in the C0 2 content alter the Earth's greenhouse effect or
when recurring volcanism introduces large volumes of fine-grained ejecta into
the atmosphere. Variations in atmospheric factors can produce tangible cli-
mate change in a matter of years or decades.

Second, astronomical motions may produce changes in the pattern and
intensity of solar radiation. For example, parameters such as the tilt of the
Earth's axis and the orbital path around the Sun vary from maximum to min-
imum values in specific intervals of time. Periodically, a number of such fac-
tors coincide to produce minimum solar radiation, enough decrease in radiation
to induce an episode of glaciation. The astronomical motions have a period-

icity such that their effect on climate would occur in a time span ranging from
20,000 years to 100,000 years.
Third, many scientists believe that a cause-and-effect relationship exists
between climate change and variations in the elevation and distribution of con-
tinents. The concept implies that the major cooling needed to produce glacia-
tion occurred when landmasses were high. It also has been suggested that the
distribution of landmass and ocean basins associated with the mechanics of
polar wandering and plate tectonics is important in controlling climate. Land-
mass movements probably influence climate over cyclic time intervals, per-
haps millions of years.
In geomorphology the cause of climate change is less important than the
adjustments demanded within surficial systems in response to the change. Space
prevents us from examining the effect of climate in every geomorphic system.
It may be instructive, however, to consider several examples of threshold-pro-
ducing phenomena that serve as intermediary links between the variables of
climate change and the process response in geomorphic systems.
Sea Level Fluctuation One major intermediary factor related to climate

change and responsible for generating geomorphic adjustments is the phe-
nomenon of sea level change. As stressed earlier, sea level is a most important
horizon in geomorphology. In addition to serving as the datum for relative
gravity analyses, sea level represents the ultimate base level for rivers draining
the land and the theoretical end point of continental erosion. We know from
Thousands of years before present

Figure 2.11.
s
MSLO
CD
50
40
1 1 1 1
35
/
1
/
/
1 1 1
30
iiii
s.
25
\
i
\
i i i
20
IIII
15
W
10
*
Late-Ouaternary fluctuations of
sea level, from compilation of
published and unpublished
radiocarbon dates and other
geologic evidence Dotted curve
estimated from minimal data
Solid curve shows approximate
E
/
\
\ mean of dates compiled Dashed
§"
100 v curve slightly modified from
Curray(1960, 1961) Probable
- / V /
fluctuations since 5000 B p not
- /' shown here
150
Mid-Wisconsin Late Wisconsin Holocene Modern

transgression regression transgression
geologic history that sea level is not constant relative to the adjacent land-
masses. This relative position has shifted through time because (1) the con-
tinents are subject to isostatic uplift or depression and (2) the water level itself
may rise or fall, an adjustment known as eustatic change. Eustatic sea level
change occurs on a worldwide basis in contrast to isostatic movements, which
are determined by the local or regional gravity environment.
Although there are a number of ways to produce eustatic sea level change,
the accumulation and removal of glaciers is by far the most obvious and ef-
fective. Glaciers are interruptions of the hydrologic cycle because water locked
up in ice represents precipitation that is not being returned to the ocean. As
a result, when the volume of ice held on the land increases during a period of
glaciation, the oceanvolume shrinks and its surface is lowered. Conversely, if
all modern were melted, sea level would rise by an amount comen-
glaciers
surate with the volume of water released from storage. Clearly, the growth
and wastage of glaciers attendant to glacial and interglacial stages of the Qua-
ternary are major progenitors of eustatic sea level change.
Exactly how much eustatic fluctuation occurred in Pleistocene cycles is
difficult to ascertain because all evidence used to make such estimates is sub-
ject to error (see Flint 1971). The most reliable evidence is the presence of
near shore (littoral) marine fossils that are datable by radiocarbon and are
now covered by about 100 m of ocean. These indicate that the shoreline was
approximately 1 00 m lower at the time of the last glacial maximum. Similarly,
emerged strandlines bearing datable shells are now preserved well above present
sea level and, if isostasy is neglible, they mark sea level during interglacial
times. The use of such evidence has allowed the construction of curves showing
the rise in sea level since the last glacial maximum (fig. 2.1 1). These curves,
corroborated in numerous studies, indicate a reasonable steady increase in
ocean level from about 18,000 years b.p. (before present) to about 5000 years
B.P. when sea level apparently stabilized near its present configuration.
It is ironic that the most accepted evidence of eustasy is also the evidence
used to document isostasy. This presents Pleistocene geologists with a nasty
interpretive problem because vertical change in relative sea level is produced
by the mechanics of two different processes that are both based on the same
Chapter 2 56
REEF Vila Vllb

Figure 2.12.
I
Late Quaternary paleosea levels

based on estimates from New
Guinea and elsewhere
-120
Years X 1000. before present
New Guinea reef crests, dated, each dot is from reef heights on traverse.
T New Guinea reef complexes, low sea level maxima, undated (Chappell. 1974)
New Guinea reef complexes, dated (Veeh and Chappell. 1970)
] Compilation of Steinen. et al. (1973) — primarily Barbados data

Sea level minimum not known
Diagrammatic paleo sealevel trace
• Sea level estimate using premature emergence correction
evidence. Therefore, sea level change attributable to eustasy cannot be ac-

curately determined without knowledge of isostatic movement during the same
interval. In addition, the depth measurements of preserved fossils are often
made suspect because of subsidence of the sea floor under the weight of water
added during the interglacial stage (Clark and Bloom 1979). The problems
associated with interpretation of sea level change are amazingly complex (see
Morner 1980 for discussions), but detailed studies in widespread locations are
beginning to unravel the complicating factors (Bloom 1967, 1980; Bloom et al.
1974; Bender et al. 1979; Cronin 1981; Cronin et al. 1981; and many others).
Notwithstanding the interpretive problems, numerous studies, especially
those in Barbados and New Guinea (Matthews 1973; Steinen et al. 1973; Bloom
et al. 1974; Chappell 1974), have resulted in a reasonable curve of eustatic
sea level for the last 150,000 years (fig. 2.12). The data, based on 230Th/ :;u U
dates of coral reef fossils, show that the highest sea level in that period was
about 6 m higher than present sea level during the Sangamon interglacial stage.
Other high sea level stands occurred during interstadials of the Wisconsin stage,
but none of these was as high as the present level. The lowest sea level, prob-
ably somewhat over 100 m lower than present, occurred during the last glacial
maximum.
Sea level fluctuations have a direct influence on a number of shoreline
processes and features (beaches, barrier islands, etc.) These will be discussed
Years before present

Figure 2.13.
40,000 20.000
Lower Mississippi Valley
chronological concepts (A) as
proposed by Fisk (1944), and
by Saucier (1968)
(B) as applied
to sea level fluctuation curve
proposed by Curray (1965)
1 1 1 1
BHH M| ^M ^^H *"
S
Early Wisconsin ^^. Late Wisconsin
100
glaciation / \ glaciation /
I \ Woodfordian /
— \ 1 \ substage (?) J
200 c
1-5 \ Altonian
J Farmdalian \
\ substage (?) 1 substage (?)\
300
co £
c \
400 V
500
V,
Deposition Valley alluviation culminating r Valley alluviation
of Prairie with the formation of certain culminating with the
formation braided-stream surfaces such formation of braided-
as Macon Ridge and possibly stream surfaces
some meander belts
Primary entrenchment Secondary entrenchment Development of

of the alluvial valley of the alluvial valley modern meander belts
(B)
in a later chapter. Sea level change also permeates into the fluvial system and
through it into other systems. For example, in coastal regions river terraces
commonly result when alternating cutting and filling are initiated by fluc-
tuating sea level. Theoretically, entrenchment should accompany glacial ex-
pansion (when sea level is decreasing), while filling would take place during
the waning phase of the glacial cycle (when sea level is rising). Fisk (1944)
interpreted the depositional terrace sequence in the lower Mississippi Valley
as being related directly to waxing and waning glaciations and their effects
on eustatic sea level change. Although the details of aggradation and degra-
dation were modified as more data became available, the relationship between
sea level change and depositional terraces in the lower Mississippi basin is
generally recognized as being real. Figure 2.13 shows these interpretations.

Chapter 2 58
Figure 2.14.
Location of the Pomme de Terre
River in relation to the
Missouri-Mississippi river
drainage basin, and the limits of
continental glaciation
/ Study area
Wisconsin glaciation 500 Km
Pre-Wisconsin glaciation
The downcutting produced during glaciation may be gradually propa-

gated upstream, causing a similar erosional response in each tributary basin.
At that time, slopes may be regraded to new levels and groundwater tables
may be lowered. Thus, by process linkage, one external change in climate be-
gins a chain reaction of adjustments throughout the geomorphic regime. How
far responses will be propagated upstream in a system as large as the Missis-
sippi basin depends on the erosive power of the river and how long conditions
remain stable in the interior reaches. We know, for example, that some trib-
utaries in central Missouri (fig. 2.14) were apparently not affected by the lower
Mississippi sequence recognized by Fisk (Brakenridge 1981). In fact, Brak-
enridge (1981) suggests that the terrace and floodplain sequence along the
Pomme de Terre river shows no clear correlation with either sea level history
or glacial chronology in the upper Missouri-Mississippi basin. Aggradation

occurred during glacial, nonglacial, and interstadial periods. Terrace forma-
tion by river downcutting proceeded in the late Pleistocene, as well as in the
Holocene when major glaciation and eustatic base level changes were absent.
Presumably this indicates that alternating filling and cutting can be directly
induced by climate change without glaciation or base level change. This sup-
ports the contention by Knox (1976) that Holocene downcutting might be
caused when variations in atmospheric circulation produce conditions that lead
to an increase in river energy. Downcutting, therefore, would represent a re-
sponse to a threshold of stream power (Bull 1979). In the following section
we examine how such responses might occur and whether the relationship be-
tween climate and threshold response is as straightforward as it appears.
Figure 2.15.
Curves illustrating the effect of
temperature on the relation
between mean annual runoff and
mean annual precipitation
1.0 10 100
Mean annual runoff, in inches
Geologic and Vegetal Screens The preceding discussion reveals that it may
be more difficult to explain the climatic effect on landforms located far from
the site of sea level change. We can use the same landform as above to ex-
emplify this point. River terraces in the interior of continents have historically
been attributed to the fluctuating climate accompanying glacial and intergla-
cial conditions. The swing from arid to humid as glaciation begins (and vice
versa) affects not only the prevailing discharge but also the amount and type
of sediment delivered to the rivers. Such changes in fundamental river controls
logically produce the trenching and filling. Pleistocene and Holocene climates
in North America are reasonably well known. In general terms, average tem-
peratures in the Pleistocene were probably about 7°C (= 13°F) lower than
present during glacial stages and 3°C (« 5°F) higher than present during
interglacials. Average annual precipitation was as much as 25 cm (10 in.)
greater during glaciation than now and 12 cm (« 5 in.) lower during the
warmer periods. These values vary depending on latitude and the technique
used to make the estimate (Flint 1971; Budyko 1977).
Knowing this, it is ironic that little agreement exists as to which climate
produced the filling and which caused the trenching in the process of terrace
formation. To complicate matters, different interpretations of the cause and
effect relationship in nonglaciated regions of the United States seem to be
supported by field evidence (for a discussion see Flint 1971, pp. 304-307).
The ambiguity revealed here relates to the second intermediary link between
climate change and response that rests in how climatic ingredients are filtered
into geomorphic systems through geologic and vegetal screens.
A
major climatic influence in geomorphic systems occurs because tem-
perature and precipitation are fundamental controls on mean annual runoff
and the magnitude of erosion. As you might expect, runoff increases with higher
annual precipitation, but at constant precipitation the runoff will decrease with
higher temperature because evapotranspiration is known to be greater in
warmer regions (fig. 2.15). The amount of sediment yielded from basin slopes
is an indication of erosion and is also a function of temperature and precipi-
tation (fig. 2.16).
.
Chapter 2 60
1000
Figure 2.16.
Curves illustrating the effect of
temperature on the relation
between mean annual sediment
yield and mean annual CD
precipitation >* CD
11
E £
•o
<D
5
=>
tn cr
_ in
co ,_
3 <D
C CL
c „
CD£
c o
CD
10 20 30 40 50 60 70
Mean annual (effective) precipitation (inches)
Figure 2.17.
~ 300
c
;
Average annual sediment yield as

it varies with effective 250
precipitation and vegetation
5
CD
200
CD
.1 150
T3
CD
CO
"co
100
c
CO
CD
D)
C3
50
20 40 60 80 100 120 140
Effective precipitation (cm)
The curves shown in figure 2.16 were derived by using data from basins
in the western United States averaging 3900 km 2 (1500 mi 2 ) in area and may
not be suitable for other settings. Significantly, however, they show that the
influence of climate on erosion is filtered through a vegetal screen. For ex-
ample, figure 2.17 represents the 50° F curve, where effective precipitation is
the mean annual precipitation adjusted for that prevailing temperature

(Langbein and Schumm 1958). Under the stated conditions, maximum sed-
iment yield occurs at approximately 30 cm (12 in.) of effective precipitation
because the density and type of vegetation developed at that precipitation and
temperature is most conducive to water-related erosion. Where precipitation
is lower than 30 cm, not enough water is Above
available to erode the slopes.
30 cm of precipitation, the vegetal cover becomes more dense and changes
Temperature — Precipitation Figure 2.18.

Hypothetical flow chart showing
how climatic variables exert an
influence on rivers A change in
climate alters sediment

concentration, sediment size, or
load type, requiring a response by
the river system Responses vary
depending on local conditions
Hydrology
Sediment concentration
Sediment size
Load type
Erosion and deposition Channel adjustment Systems

Slope change Hydraulic geometry Pattern change
Cut and fill Sinuosity Drainage networks
Terrace formation Meander wavelength
from desert brush to grasses. The ubiquitous root systems tend to fix sediment
on the slopes, and therefore, with increased precipitation sediment yield be-
comes progressively lower until it nearly stabilizes under a forest cover.
The Langbein-Schumm analysis will be discussed further in chapter 5.
For now, it is important for you to recognize that climatic change does not
affect geomorphic systems directly but is passed through geological (rock type,
and vegetal screens that determine how much water and sediment get
soils)
two factors require threshold adjust-

to a river channel. Variations in those
ments that can occur in a number of different ways (fig. 2.18).
Equally important is the fact that the same climatic change may prompt
entirely different sediment yields and, therefore, different geomorphic re-
sponses. For example, considering figure 2.17 again, assume that a 15-cm de-
crease in precipitation occurs in a particular drainage basin that had an effective
precipitation of 45 cm prior to the change. The reduced precipitation will re-
sult in a greatly increased sediment yield. However, the same 15-cm decrease
in a basin having a 35-cm annual precipitation prior to the change will produce
a major reduction in sediment yield. Theoretically, then, the same climate
change may result in cutting by one river and filling by another because the
type and amount of sediment yielded during adjustment to the new climate is
oppositely directed. What this tells us is that the effect of climate change may
be highly dependent on the antecedent values of temperature and precipita-
tion. If that is a true statement, knowledge of preexisting climate may be as
important in understanding how systems respond to climate change as knowing
the magnitude of the change itself.
Chapter 2 62
In sum, the relationship among climate, process, and landform is not easily
determined because the effect of change is sidetracked into ancillary factors.
The adjustments of these factors in the new climate provide variable condi-
tions of load and water which spur responses that are not predictable under
the present state of our knowledge. This suggests that we do not have clear-
cut relationships between climate and landforms because we are far from un-
derstanding the climatic geomorphology scheme.
Summary In this chapter we briefly examined climate and endogenic factors as major
external controls on geomorphic systems. The endogenic influence occurs pri-
marily through the addition of mass and energy by volcanism and tectonic
activity.The most important tectonic processes are those producing vertical
movements of the Earth's surface. One of these is the isostatic adjustment
required when the internal mass balance is upset. Other vertical movements,
associated with faulting and warping, are integral parts of a subdiscipline
known as tectonicgeomorphology in which the relationship among tectonics,
processes, and landforms is utilized in a variety of geologic and environmental
studies. In most cases, vertical displacements induce threshold conditions and
responses in the effected surficial systems.
Climatic geomorphology explores the relationship between prevailing cli-
mates and the landforms expected to form under those conditions. Of greater
significance is the fact that climate change is a threshold-producing phenom-
enon, although the cause-and-effect relationship may pass through interme-
diary factors such as eustatic sea level fluctuations or vegetal and geological
screens.
Suggested Readings The following references provide greater detail concerning the concepts dis-
cussed in this chapter.
Andrews, J. T, ed. 1974. Glacial isostasy. Stroudsburg, Penn.: Dowden.

Hutchinson and Ross.
Bloom, A. L.; Broecker, W. S.; Chappell, J. M. A.; Matthews, R. K.; and
Mesolella, K. J. 1974. Quaternary sea level fluctuations on a tectonic
coast: New
230
Th 234 U dates from the Huon Peninsula, New Guinea.
Quat. Res. 4:185-205.
Bull, W. B. 1984. Tectonic geomorphology. Jour. Geol. Educ. 32:310-24.
Derbyshire, E., ed. 1976. Geomorphology and climate. London: John Wilej
& Sons, Ltd.
Keller, E. A.; Bonkowski, M. S.; Korsch, R. J.; and Shleman. R. J. 1982.
Tectonic geomorphology of the San Andreas fault zone in the southern
Indio Hills, Coachella Valley, California. Geol. Soc. America Bull.
93:46-56.
Macdonald, G. 1972. Volcanoes. Englewood Cliffs, N.J.: Prentice-Hall.
Oilier, C. D. 1981. Tectonics and landforms. New York: Longman.
Chemical Weathering and Soils
/> er t
3
I. Introduction 3. Estimates Based on
II. Decomposition Chemical Analyses
A. Processes of Decomposition III. Soils
1. Oxidation and A. The Soil Profile
Reduction B. Soil Classification
2. Solution C. Pedogenic Controls and
3. Hydrolysis Regimes
4. Ion Exchange 1 . Podzolization
B. Mobility 2. Laterization
1. Leaching 3. Calcification
2. pH D. Geomorphic Significance of
3. The Eh Factor Soils
4. Fixation and IV. Summary
Retardation V. Suggested Readings
5. Chelation
C. The Degree and Rate of
Decomposition
1. Mineral Stability
2. Secondary Minerals
a. Clay Minerals
b. Hydrous Oxides
63
Chapter 3 64
Introduction We are now ready to focus on the exogenic processes that mold the geomorphic
framework into recognizable topographic forms. The first step is to recall that
most of the Earth's surface is not composed of solid rock but is underlain by
the unconsolidated remains of thoroughly altered rock. The fresh rocks and
minerals that once occupied the outermost position reached their present con-
dition of decay through a complex of interacting physical, chemical, and bi-
ological processes, collectively called weathering. Weathering progressively
alters the original lithologic character until what finally remains in the space
of the former rock is an unconsolidated mass consisting of (1) new minerals
created by the weathering processes, (2) minerals that resisted destruction,
and (3) organic debris added to the weathered zone.
Since every mineral species has, by definition, a unique chemical com-
position or atomic arrangement, it is not surprising that each type resists or
responds to weathering in a special way. Considering the wide variety of cli-
mates driving the processes and the almost endless array of rock structures
and mineral types, the precise mechanics of weathering could easily be beyond
our comprehension. Nature, however, has simplified our task because, as in-
dicated earlier, the bulk of the Earth's crust composed of a surprisingly

is
small number of mineral varieties with an equally limited chemistry. With this
advantage, we can explore weathering efficiently even though the systems in-
volved are exceedingly complex.

Weathering is usually divided into separate domains of chemical pro-
cesses {decomposition) and physical processes {disintegration). The distinc-
tion between the two is real because the processes of disintegration involve no
chemical reactions but simply produce smaller particles from larger ones.
Nonetheless, the two realms of weathering operate simultaneously and, in fact,
each may directly affect the character and rate of the other. For example,
breaking a large rock into smaller particles increases the total surface area
and thereby accelerates the chemical attack on the material. Conversely, ex-
pansion of minerals by chemical processes may exert enough internal stress
to hasten the disintegration of the rock. Realistically, then, the physical and
chemical functions of weathering may be so intimately intertwined that to
consider them as unique processes is mostly a matter of convenience. In this
chapter we will deal with only the chemical and biological aspects of weath-
which are dominant in the development of soils. Disintegration will be
ering,
examined in the next chapter when we consider the stability of slopes, because
many of the processes that break rocks apart are also important in the erosion
of unconsolidated slope material by mass wasting.
A soil is the residuum that resultsfrom the application of weathering over
an extended period of time. Given the proper conditions, a distinct layering,
called the soil profile, will develop in the residual material, and its charac-
teristics will directly reflect the weathering and soil-forming processes. Al-
though both geologists and pedologists use soil profiles as a basis for
65 Chemical Weathering and Soils
classification, other scientists use the term "soil" in different ways. To an en-
gineer, for example, a soil consists of any accumulation of unconsolidated de-
bris; it may include material such as alluvial deposits that have not experienced
the effects of weathering. Agricultural experts may be interested in only the
upper part of a soil profile, which is the segment that supports vegetation.
Clearly, the definition and depend on what information
classification of soil
an investigator wishes to gain from and on what ways such data will be used.
it
The farmer, of course, wants increased crop production, and the engineer re-
quires knowledge of physical properties, such as bearing strength, to design
buildings and other constructions judiciously. The geologist, however, utilizes
soils and other weathering phenomena as clues to the intricacies of geologic
history and the relative age of unconsolidated deposits. Although most soils
are young in the geological sense, some profiles have been preserved in the
older record, and these provide critical evidence about environmental condi-
tions at the time of formation. Mineral compositions of many sedimentary
rocks reflect the combined tectonic and climatic conditions during their origin,
and a correct interpretation of such rocks requires a knowledge of weathering
and soil-forming processes. Landforms are often recognized as relict features
because their soil character is inconsistent with the prevailing modern climate.
Many other examples could show that we utilize an understanding of soils in
many aspects of geology. The study of chemical weathering and soils is not
simply an adventure into esoteric geomorphology, for a working knowledge of
weathering processes is essential to any scientist interested in the surface en-
vironment.
Rocks and minerals are usually not in equilibrium with conditions that exist Decomposition
at or near the Earth's surface. As a result, decomposition can be viewed as a
group of processes that attempt to create substances that are more nearly stable
in that environment. This march to equilibrium is accomplished by alteration
of original materials and/or production of new mineral types.

The most important agent of the weathering regime is water introduced
as rain. Because rainwater is usually mildly acidic, a significant part of chem-
icalweathering can be visualized as a process whereby minerals assimilate
hydrogen ions and/or water and release cations to the soil liquid. Compli-
cating this simple model are two facts: (1) water entering the ground is not
chemically pure but contains a variety of ions captured from the atmosphere
and introduced from surface materials; and (2) organic processes, involving
metabolism of microorganisms and decay of vegetal matter, add gases and
organic acids to the system. These organic functions are of such importance
that some authors (Carroll 1970) suggest that chemical weathering proceeds
in two stages. The first stage, driven primarily by inorganic processes, is called
geochemical weathering and produces rotten rocks or saprolites. The second

stage, called pedochemical weathering, leads to the formation of soils from
the saprolitic material; it is chiefly a biologically controlled phenomenon.
Chapter 3 66
Figure 3.1. OH-

Ion exchange and chemical O-H
bonding take place as surface of
orthoclase feldspar comes In Si-OH
contact with a solution containing
OH-
dissociated H + and OH" ions. O-H
Orthoclase Orthoclase
AI-OH OH-
feldspar feldspar
A-H H+
H KOH
l_j-/
Exchange OH
OH O-H
OH-
OH
As rainwater percolates into exposed rock material, the first important

response commonly is the breaking apart of the structures of the parent min-
erals. As water surrounds a mineral or penetrates its structure through micro-
openings and cleavages, chemical reactions occur that tend to disrupt the min-
atomic arrangement. Jenny (1950) pointed out that atoms along
eral's orderly
exposed mineral surfaces are not satisfied electrically and so may attract the
dipolar water molecules. If the attraction causes water to dissociate into H+
and OH", these will bond to the exposed ions of the mineral, as in figure 3.1.
Hydrogen is then in the proper position to replace mineral cations, thereby

releasing them to the surrounding fluid. This process may have a profound
effect on the pH of the liquid, as hydrogen is progressively depleted and hy-
droxyls are concentrated. Experiments observing the change in pH when dif-
ferent minerals are pulverized in distilled water and carbonic acid (Stevens
and Carron 1948) show that an equilibrium pH is eventually attained. Pre-
sumably this condition is reached when the number of cations removed is bal-
anced by an equal number reentering the mineral structure. These abrasion
pH values are significant in that they help determine the fluid's effectiveness
in attacking other minerals. The values of abrasion pH for common minerals
(table 3.1) demonstrate that minerals with alkali or alkaline earth elements
in their composition tend to produce high pH liquids, while other minerals
yield acidic fluids.
The original mineral is not necessarily completely destroyed. The process
may proceed layer by layer from the external surface, or it may break the
mineral into many small pieces, each retaining the structure of the original
material. Weathering of minerals often occurs at sites of excess energy on the
mineral surface (Berner and Holdren 1977, 1979), and the position of these
sites is probably determined by the internal properties of the mineral itself.
As a result, the mineral surface is not attacked uniformly but, instead, de-
velopsnumerous widely spaced etch pits (fig. 3.2) where weathering is active
(Parham 1969; Keller 1976, 1978; Berner and Holdren 1977, 1979). Thus,
the end product of weathering may be directly influenced by the manner in
which the minerals are destroyed. Some silicate minerals, for example, break
into molecular chains that are easily recombined with available cations to form
layered clay minerals.
Table 3.1 Abrasion pH values for common minerals.
Mineral Composition Abrasion pH
Kaolinite AI 2 Si 2 5 (OH)4 5, 6, 7
Boehmite AIO(OH) 6, 7
Gibbsite AI(OH) 3 6,7
Gypsum CaS0 4 -2H 2 6
Hematite Fe 2 3 6
Montmonllonite (AI 2 Mg 3 )Si 4 O, (OH) 2 nH 2 •
6, 7
Quartz Si0 2 6, 7
Zircon ZrSi0 4 6, 7
Chlorite (Mg,Fe) 2 AI 4 Si 2 O 10 (OH) 4 7,8
Muscovite KAI 3 Si 3 O, (OH) 2 7, 8
Biotite K(Mg,Fe) 3 AISi 3 O, (OH) 2 8,9
Calcite CaC0 3 8
Anorthite CaAI 2 Si 2 8 8
Orthoclase KAISi 3 8 8
Albite NaAISi 3 8 9, 10
Dolomite CaMg(C0 3 )2 9, 10
Augite Ca(Mg,Fe,AI)(AI,Si) 2 6 10
Hornblende Ca 2 Na(Mg,Fe +2 )4
(AI,Fe +3 ,Ti) 3 10
Si 6 22 (0,OH) 2
Olivine (Mg,Fe) 2 Si0 4 10, 11
Data from Stevens and Carron 1S
MkJ Er *^
Figure 3.2.
Scanning electron microscope
photograph of hornblende grain
&/
Mv< y,'
f -
r
t
showing weathering of etch pits
that are elongated along cleavage
planes.
V ^* *m
t
\ &3
If s / jf '
*
WlJSk < dv***

I 4F
w ...
m w *Ktj^
y
Bfl
t
V f K K
/
P. z' f •
%. Mp If
Chapter 3 68
The second important response in chemical weathering relates to how easily

ions are released from the parent structure and how mobile or immobile they
are after their liberation. Considerations here are whether released ions or
groups of ions are readily fabricated into new, stable minerals, or whether they
can be removed in solution by percolating soil fluids. The ultimate fate of re-
leased particles depends on what they are and on the characteristics of the
fluid into which they are released. The equilibrium state, for example, can only
be attained in a closed system, for continuous addition and removal of water
surrounding the decomposing mineral will alter the pH. carry some of the
new elements that can combine chemically with

released ions away, or provide
those escaping the mineral. The final result is determined by a myriad of in-
terreactions, depending on how the original structure breaks apart and how
mobile the ionic or molecular particles are under the physical and chemical
constraints within the weathering zone.
Processes of Decomposition
As just suggested, the magnitude and direction of chemical weathering de-
pend on the composition and structure of the minerals being attacked, how
they break apart, and how mobile the constituents are in the weathering en-
vironment. Various combinations of mineral types and chemical processes can
result in a variety of final products. The common chemical reactions involved
in decomposition are oxidation and reduction, solution, hydrolysis, and ion
exchange. Each process plays a particular role in the overall scheme of chem-
ical alteration, although all function simultaneously in the weathering zone.
Oxidation and Reduction Oxidation occurs when an element loses electrons

to an oxygen ion. The process tends to occur spontaneously above the water
table where atmospheric oxygen is readily available: therefore, most elements
at the Earth's surface exist in an oxidized state. Below the water table the
environment is generally reducing (Loughnan 1969): however, high concen-
trations of organic matter may cause local reducing conditions to occur above
the water table.
The ease of oxidation depends on the redox potential (Eh), the magnitude
of which is controlled by the abundance of organic matter and the accessibility
of free oxygen. In most soils. Eh values range from —350 to +700 (fig. 3.3).
sufficient to keep the majority of common elements in their oxidized state.

However, some elements, such as aluminum, change from a reduced to an
The most common elements af-
oxidized form with considerable difficulty.
fected by fluctuations of Eh are iron, manganese, titanium, and sulfur. Iron is
easily oxidized to the ferric (Fe

+3
) state:
2Fe +2 + 4HCO.r + Vi0 2 + 2H 2

— Fe 0,
:
- 4H : CO.,.
Its appearance as crustations on grains or as reddish-brown stains along frac-

tures usually signifies the first tangible evidence of decomposition. The redox
phenomenon exerts a significant control on the mobility of certain elements
as they are affected by decomposition. This role of the EH factor will be dis-
cussed later in the chapter.
Figure 3.3.
Eh-pH characteristics of soils Eh
values in millivolts
Solution The process of solution is critical in chemical weathering because

when atoms are dissolved from a mineral, the structure becomes unstable.
However, the precise way in which a mineral collapses varies with its crys-
talline structure and the mobility of its constituent ions. When an atom is
removed from its parent mineral, it may remain in solution and be taken com-
pletely out of the system by the downward moving fluids. This depends on the
concentrations of reactable ions in the fluid and on other chemical character-
istics of the medium.
Solution is usually exemplified by the reaction of calcite and acid:
(calcite) (carbonic acid)
CaC0 3 + H 2 C0 3 -^ Ca +2 + 2(HC0 3 )".

The role of C0 2 in the solution of calcite is critical and is especially significant
in the development of karst topography. The C0 2 factor will be discussed in
chapter 12 as we examine the processes that form karst regions. Here we simply
emphasize that most common elements and minerals are soluble to some
degree in normal groundwaters, where pH values usually range from 4 to 9
(fig. 3.4). In the weathering of silicate minerals, note that silica is soluble
under all normal groundwater conditions. The degree of solubility is rather

low (as 6 ppm) when the silica is contained in quartz (Morey et al. 1962) but
increases considerably (« 1 15 ppm) in amorphous silica (Morey et al. 1964).
Aluminum oxides are virtually insoluble under normal groundwater condi-

tions, and iron in the ferric state can be dissolved only by rather acidic fluids.
It is not surprising, therefore, that mature weathering profiles in humid cli-
mates should be characterized by the presence of abundant ferric iron and

aluminum. Other cations are readily soluble, and their removal to the water
table concentrates the least soluble constituents in the weathered zone.
Chapter 3 70
Figure 3.4.
The relationship of solubility of
common substances to various /
/ \
\ \ /M f
^
B
pH conditions. Ti0 2 Al 2 0, \ Fe(OH) 2 Ca(OH)J
,
Fe(OH) 3 CaC0
9/
Ti(OH)\
w 7 /
/ 1/ 3
•
r
O
le
o
CO
\>/
/Si0 2
IH
A-Mg(OH) 2
1
\
I
3 4 5 6 7 10 11 12 13 14
pH
Sol. Sol. (^ Insol.
Insol.
J
Hydrolysis The reaction described on the preceding page between mineral

elements and the hydrogen ion of dissociated water is called hydrolysis. Chem-
ically, hydrolysis involves a reaction between a salt and water to produce an
acid and a base; it is probably the most important mechanism in breaking
apart structures of the silicate minerals. During the process, metallic cations
are separated from the mineral structure and replaced by H+ , which is held
in the original aluminosilicate complex (fig. 3.1). Most of the replaced cations
are soluble in natural waters.
In addition to freeing cations, hydrolysis usually produces H 4 Si0 4 HCOr ,
and OH~, all of which can be considered to be in solution (Birkeland 1974).

In the common case where water and acid attack an aluminosilicate mineral,
the reaction leaves the H + embraced in segments of the original structure, and
these are subsequently recombined into clay minerals. The hydrolysis of or-
thoclase feldspar shown here and in figure 3.1 demonstrates this response:
(orthoclase) (kaolinite)
2KAlSi 3 8 + 2H + 9H +
2
0- ,H 4 Al 2 Si 2 9 + 4H 4 SiO, + 2K + .
The removal of the metallic cation will proceed as long as free hydrogen
ions are available, easily replaced cations are present, and the solvent has not
reached saturation with respect to the ion being liberated. The continuous in-
troduction of fresh water and the development of organic acids will assure a
ready source of H+
Since OH~ is carried downward to the groundwater table
.
by percolating water, it seems unlikely that highly alkaline water can ever be
produced in open systems with constantly moving water. Therefore, the normal
product of a continuously leached system is a residue in which all mobile cat-
ions have been freed from the original mineral. After the cations are released,
they either remain in solution or attach to the surfaces of minerals held in
colloidal suspension. Eventually they are carried to the water table and deliv-
ered to the regional streams.
Clearly, the effect of hydrolysis decreases as clays depleted of cations be-
come the dominant aluminosilicate in the weathering zone. In fact, as more
clays form, they commonly become colloidally suspended in the fluid and may
+
adsorb H to their surfaces. This may cause the liquid to become more acidic
rather than more alkaline as hydrolysis proceeds.
Ion Exchange Ion exchange is the substitution of ions in solution for those
held by mineral grains. Although all minerals possess some capability for ion
exchange, the process is most effective in clay minerals. The ions to be ex-
changed are held on the surfaces of clays because unsatisfied charges, exposed
hydroxyl groups, and isomorphic substitutions such as Al +3 for Si +4 have given
the clays an overall negative charge. Cations are held at the mineral surfaces
by adsorption. In soils, "adsorption" usually refers to the attraction of ions
and water molecules to the surfaces of colloidal clays in an attempt to neu-
tralize their negative charges. The adsorbed ions are not held too tightly and,
therefore, are susceptible to replacement by or exchange with other cations.
Each clay species has a different propensity for adsorbing cations, called
its cation exchange capacity (c.e.c), which is expressed as the number of mil-
liequivalents per 100 grams of clay. When the adsorbed ion is hydrogen, the
c.e.c. directly influences the pH value that the clay assumes. Kaolinite, for
example, takes on a pH of 4-5 under complete adsorption, while H + -mont-
morillonite attains a pH value as low as 3. Colloidal suspensions of these clays
create acids that are capable of attacking other minerals. In soils, colloids of
organic compounds produce similar acids because the c.e.c. of organic matter
is usually rather high, ranging from 150 to 500 (Birkeland 1974).
It should also be noted that cations other than H + may be adsorbed by
clays and that the prevailing environment controls which cation types are more
likely to be adsorbed. For example, in humid regions colloidal clays will adsorb
H^ and Ca* more readily than Mg + Na + or K + In well-drained arid soils,
+
, , .
Ca ++ and Mg ++ are usually the most prominent exchangeable ions, and H +

is the least common. In poorly drained arid soils, the number of adsorbed so-
dium ions often equals or exceeds the calcium (Buckman and Brady 1960).
Ion exchange is governed by the composition and pH of the interstitial
water as well as the type of ion in the exchangeable position. In general, strongly
acidic water allows H+ to replace metal cations of the parent minerals, but
this tendency changes as the water becomes neutral. At higher pH values, the
+
mineral cations may remain in the exchangeable position or, in fact, H may
be replaced by metallic cations. In soils, the pH of the soil-water mixture is
an indicator of the number of cations held in the exchange position by clays.

Chapter 3 72
This characteristic is commonly referred to as a percentage of base saturation,

meaning the percentage of exchange sites occupied by cations other than hy-
drogen. The higher
the percentage of base saturation, the higher the pH of
+
the soil-water complex because less is held by the clays. H
Mobility
The extent to which chemical weathering will alter the parent mineralogy de-
pends largely on the relative mobilities of the constituent ions. Some ions are
easily removed (high mobility) from the weathering system under normal
groundwater conditions, while others are relatively difficult to remove (high
immobility). The presence of highly mobile ions in a mature weathering zone
indicates that some factor is impeding the transfer of the ion from the system.
Orthoclase feldspar, for instance, will hydrolyze to kaolinite (as shown earlier)
if all the potassium is lost in the process. If some potassium is retained, the
clay product will be illite rather than kaolinite, as shown in the following re-
action:
(orthoclase) (illite)
3KAlSi 3 8
+
+ 2H + 12H z O -. KAl 3 Si 3 0,o(OH) 2 + 6H 4 Si0 4 + 2K + .
Accordingly, then, since K+

its immobile behavior in the for-
is a mobile ion,
mation and preservation of must be linked to the prevailing character of
illite
+
the fluid or to an incomplete breakdown of the orthoclase that traps the K
in molecules of the original structure. In either case, the easily removed po-
tassium is rendered immobile and remains in the weathered zone.

The relative mobilities of common cations are as follows, in order of de-
creasing mobility:
(Ca +2 , Mg +2 Na + > K + >

, ) Fe +2 > Si
+4
> Ti +4 > Fe +3 > AT 3
.
Such a mobility distribution is probably related to a parameter known as the

ionic potential, which is expressed as the ratio of the valence (Z) to the ionic
radius (r). In general, Z/r is a useful first approximation of mobility because
very mobile ions have Z/r values less than 3; those forming immobile precip-
itates have values between and those forming soluble, complex an-
3 and 9.5;
ions are usually greater than 9.5 (fig. 3.5). However, as each cation can be
immobilized by external factors, those factors in a sense greatly influence how
effectively the major processes of chemical weathering will work.
For example, Shoji and his colleagues (1981) found a different mobility
sequence than the normal one just shown. They concluded that the volcanic
glass parent rock and the secondary minerals derived from its weathering ex-
erted a strong control on the ion mobilities. Nonetheless, if all processes do
function without hindrance, the mobile ions will be depleted from the system
and the immobile ions will be progressively concentrated. The major external
factors that control mobility include leaching, pH, Eh, fixation and retarda-
tion, and chelation.
Figure 3.5.
Grouping of some common
elements according to their ionic
radius (r), ionic charge (Z), and
ionic potential (Zr). (From Gordon
et al 1958)
Soluble complex anions

&S*&~ i i
3 4 5 6 7
Ionic charge (Z)
Leaching The most important factor influencing ion mobility is the amount
of water leaching the weathering zone. How frequently rainwater flushes
through the system, and in what quantity, depends mostly on the climate but
also on the permeability of the material. The significance of leaching is sev-
eralfold: (1) it removes in solution the constituents that have been separated
from minerals by hydrolysis and ion exchange and, by providing new hy-
drogen, allows these processes continuously to alter the original material to-
ward an ultimate degraded condition; (2) it directly affects the pH of fluids
surrounding minerals and thereby helps determine which elements will remain
in solution; and (3) it provides the mechanism by which dissolved ions and
clays are transferred from higher levels to lower levels in the weathering zone.
This introduces new components to each level and so may engender the chem-
ical environment needed new minerals.
to precipitate
In general, the absence of continual leaching makes the weathering zone
function like a closed system. Since ions removed from the parent minerals
are held in the surrounding fluid, the exchange processes will continue only
until an equilibrium condition is established between the ions in the fluid and
those in the mineral. For all practical purposes, further chemical weathering
is impossible beyond this stage, and mobile ions and easily decomposed min-
erals will remain in the system. In this way, the retardation of leaching, usu-
ally due to insufficient rainfall, is a prime factor causing immobility of ions
in a weathering system. It is not surprising, then, that mobile constituents are
abundant in arid-climate soils. On the other hand, in humid regions where
leaching is continuous, the same mobile substances are removed entirely. Even
Chapter 3 74
Hematite
Figure 3.6. Boehmite and Goethite Others
Characteristic mineralogy of
weathering zone tormed on
kaolinitic sandstone under intense
chemical weathering. Note
decrease in quartz and increase
in stable clays at top of profile. Zone of
(From Loughnan and Bayliss Concretion
1961 American Mineralogist
.
12
46 211)
^
Gibbsite
15
/ /^
21 I
iQ.
S 24 1 1 Zone of
J I Kaolinite fluctuating
27
30
y j water table
y<
33
36
C
/ ^r
^^-^J
Quartz
39 1 Parent rock
42 r s^^l^^H^^H^^H^ÎHHiBBI
Composition
very resistant parent materials such as quartz can be chemically altered to a

drastic extent if high-volume leaching continues over a long period of time.
Loughnan and Bayliss (1961) reported a truly amazing example of how a ka-
olinitic sandstone with 90 percent quartz was altered under a tropical climate
to a residuum that contained only 5 percent quartz (fig. 3.6).
pH In addition to its previously discussed influences. H + concentration (pH)

also plays an important role in the solubility of most common elements (fig.
3.4). Stated in another way, the mobility of ions pH. is partly controlled by
Although rainwater reaching the surface is slightly acidic, processes operating
beneath the surface can considerably alter the original pH value. Thus. pH is
not an independent variable in weathering but may, in fact, be dependent on
the type and degree of inorganic and organic processes within the weathering
zone.
In geochemical weathering the pH is influenced by leaching, by the c.e.c.
of residual minerals, and by the composition and structure of the parent min-
erals(Loughnan 1969). The effect of leaching and cation exchange capacity
can be shown diagrammatically by considering the progressive weathering of
a basalt, shown in figure 3.7. With continued leaching, the pH is related al-
most inversely to the type of clay produced during various stages of the weath-
ering. Assuming complete mobility of Ca, Mg, and Na, the first clay developed
Figure 3.7.
Relationship of cation exchange
capacity (c.e.c.) and pH with
continuous weathering of a basalt
Transition from basalt to bauxite
represents gradual change in
mineralogy with increased
leaching.
Basalt Montmorillonite
— Kaolinite
""3
Bauxite
Increasing leaching
is an H + -montmorillonite with a high c.e.c. value (70-100 meq/100 g). The

high c.e.c. allows the clay to adsorb many of the hydrogen ions made available
by the continuous leaching, and the colloidal acid formed causes a pronounced
drop pH. As leaching continues, however, montmorillonite is degraded to
in
kaolinite, which becomes the dominant clay mineral. The much lower c.e.c.
of kaolinite permits less hydrogen adsorption, and the pH rises. The final
product, bauxite, can adsorb virtually no H + and the pH climbs to an almost
,
neutral condition. Obviously, pH is dependent on the other factors.

It is important to remember that organic substances also form colloidal
suspensions with very high cation exchange capacities. Although the values
vary with climate and the type of organic matter, it is not uncommon for these
humic acids to lower the pH below 4.
The Eh Factor As suggested earlier, the redox potential exerts a control on

ion mobility. The effect of Eh on ion mobility is due to the fact that oxidation-
reduction reactions are reversible. When soils become waterlogged, free oxygen
is excluded and strongly reducing anaerobic bacterial action begins. This gen-
erates lowered Eh values that, in turn, cause some elements to revert to their
reduced forms. In some cases, ease of reduction is also dependent on pH (Con-
nell and Patrick 1968). Such transitions have a direct influence on the mobility
of certain elements that may be relatively insoluble in one form and easily
dissolved in the other. Removal of iron oxides attached to clays, for example,
can result from only a minor lowering of Eh, which transforms ferric iron to
the more soluble ferrous type (Carroll 1958). Apparently, a fluctuating water
table can have important ramifications on the redox potential (Eh) and with
it the relative mobility of important ions.
Chapter 3 76
The significance of Eh in decomposition and ion mobility can be briefly

stated as follows:
+2
1. Ferrous iron (Fe ) commonly binds silica tetrahedra in the
structures of silicate minerals. The oxidation of the iron to the
ferric state requires impossible internal adjustments, and the lattice
structure is destroyed (Carroll 1970).
2. The by-products of the oxidation process may facilitate the
decomposition of other, more stable, minerals. For example, the
oxidation of pyrite produces sulfuric acid, as shown:
4FeS 2 + 14H 2 + 150 2 ^4Fe(OH) 3 + 8H S0 4

2 .
This acid lowers pH and will react with any nearby mineral suscep-
tible to attack by acid. Groundwater in regions near zones of sulfide
mineralization is usually very acidic.
3. As indicated before, the solubility or insolubility of some elements
in groundwater having a normal pH is directly controlled by
whether the elements exist in oxidized or reduced form. This is
especially true for iron and titanium.
Fixation and Retardation Some cations, especially potassium, have a ten-

dency to be retained or fixed in the weathered zone, so that their mobility is
considerably lower than might be expected. The fixation of potassium may

account for its low concentration in seawater and the widespread distribution
of illite in sedimentary rocks. The precise mechanism involved is poorly under-
stood, but it is probably related to the fact that potassium silicates (muscovite.
K-feldspars) are more resistant to chemical weathering than are other min-
erals with similar lattice structures. It is known that potassium is most readily
fixed in clay minerals with expandable lattices such as chlorite, illite, micas,
vermiculite, and montmorillonite. Why this should be or why potassium is so
much more susceptible to this phenomenon than other ions is not clear, but
Wear and White (1951) suggest that the unique size of the K + ion provides
the most stable structure when combined with oxygen in layered silicates.
Ion mobility may also be slowed or retarded without complete fixation.
Wollast (1967) demonstrated that the initial stage of orthoclase hydrolysis is
accompanied by the creation of a thin surface layer of amorphous Al(OH).,
and silica. Although these substances are transferred from the coating layer
to the surrounding liquid, the fluid is quickly saturated with respect to alu-
minum. Silica, and presumably other ions, continue to diffuse through the layer,
but the rate diminishes because the distance from fresh feldspar to the fluid
increases continuously as the aluminum-enriched sheath thickens. The re-
tarding coat is not formed at pH values less than 5 because Al(OH)^ is soluble
under those conditions.
Chelation The process of chelation represents one of the most dramatic ef-
fects on the mobility of ions. Lehman (1963) defines chelation as "the equi-
librium reaction between a metal ion and a complexing agent, characterized
by the formation of more than one bond between the metal and a molecule of
the complexing agent and resulting in the formation of a ring structure in-
corporating the metal ion." To those of us unversed in organic chemistry, this
means that metallic immobile under normal conditions
ions that are extremely
can be mobilized by reacting with complexing agents and be vertically trans-
ported as part of the compound.
Most complexing agents involved in chelation are organic compounds
nurtured in soils by alteration of humus into a plant acid called fulvic acid
(Wright and Schnitzer 1963), although other organic processes also produce
chelating complexes. Lichens, for example, secrete such materials (Schatz
1963). The complexing agent ethylenediaminetetraacetate (EDTA) is by far
Figure 3.8.
the most completely understood chelator, and its structure (fig. 3.8) shows how
Structure of complexing agent
the metallic ion is bonded and held.
EDTA. Metallic ion (M) is bonded
Although the chemistry of the chelating process is far from clear, its im- within structure (Reproduced
portance in weathering has been recognized for several decades (Schatz et al. from Soil Science Society of
America Proceedings, Vol 27,
1954). During this time several excellent studies have demonstrated that EDTA 1963, p 169, by permission of the
and other chelators may play a dominant role in mobilizing iron and alu- Soil Science Society of America)
minum under pH conditions that are not conducive to Fe and Al solution (At-
kinson and Wright 1957; Wright and Schnitzer 1963; Schalscha et al. 1967).
Using several different chelating agents Schalscha and his co-workers (1967)
were able to extract iron from a variety of minerals. Significantly, their study
revealed no correlation between pH and the amount of liberated iron. EDTA
released more iron from magnetite and hematite, for example, than did hy-
drochloric acid, even though the HC1 solution was more acidic. However, in
other studies (Tan 1980) fulvic acid was capable of extracting ten times more
silica and six times more aluminum when pH was decreased from 7 to 2.5.
In soils, chelating agents become soluble in water as they are oxidized

(Wright and Schnitzer 1963). Iron and aluminum are locked in the ring struc-
ture and carried downward with the percolating water. The downward move-
ment continues until the entire mass is flocculated because of small changes
in ionic content of the soil water, or until the complex is broken by microbial
action. In either case further downward movement and the iron
is curtailed,
and aluminum are redeposited.
Even before soils begin to develop, chelation can be important in the al-
teration of the bare rock. Jackson and Keller (1970) showed that the presence
of the lichen Stereocaulon vulcani accelerated the chemical weathering of
volcanic rocks in Hawaii. Rocks covered by lichen growth had a thicker
weathering crust than lichen-free rocks, and the crust was enriched in iron
and depleted in titanium, silicon, and calcium. Rocks devoid of lichens were
measurably less affected by chemical weathering.
The Degree and Rate of Decomposition

To utilize soils in geological studies, it is important to know what the end prod-
ucts will be in a completely altered system. Geologists are, therefore, con-
cerned about what material will remain when no more chemical reactions are
possible, given the constraints of climate, vegetation, and rock type. If such a
condition can ever be reached, a steady state will have been established be-
tween the driving forces (decomposition) and the resisting materials, and al-
though the weathering zone may get progressively thicker, the upper parts of
Chapter 3 78
Table 3.2 Weathering stability of the common silicate minerals.
Olivine. Anorthite Least Stable

Pyroxenes, Ca-Na Plagioclase i \
Amphiboles. Na-Ca Plagioclase

Biotite. Albite
K-Feldspars
Muscovite
Quartz Most Stable
After Goldich 1938 © by The University of Chicago Press Used with permission
the profile are degraded to their ultimate form. In thoroughly leached soils,
all minerals remaining presumably will be stable and all mobile ions will be
gone. Thus, we have some idea as to what final products are expectable in
if
any climate, we can estimate the degree of weathering (how far the material
has progressed toward the steady state) by observing the mineral and chem-
ical composition of the soil.
Mineral Stability In considering the degree of weathering, a logical first
question to ask is what minerals will be most rapidly destroyed in a thoroughly

leached open system and, conversely, which types will remain if a steady state
is attained. Perhaps the first analytical attempt to answer that question was
provided by Goldich in 1938. In a detailed study of the weathering of several
varieties of igneous and metamorphic rocks, he suggested that mineral sta-
bility of the rock-forming silicates is directly related to their order of crys-
tallization as determined earlier by Bowen (1928). Quartz, being the last to
crystallize, forms under the lowest temperature conditions and therefore should
be most stable in the surface environment. High-temperature minerals such
as olivine and pyroxene are least stable and weather most rapidly. With the
exception of muscovite and the plagioclase feldspars, the Goldich stability se-
ries (table 3.2) reflects the number of oxygen atoms shared in the silicate
structure. Quartz and orthoclase, for example, have four shared oxygens in
their structure, making their internal bonding strength much higher than that
of a mineral like olivine, which shares no oxygens. Keller (1954) in fact sug-
gested that bonding energy may provide a useful basis on which to estimate
mineral stability.
Since Goldich's work, other guides to mineral stability have been devel-
oped (for example, Pettijohn 1941; Reiche 1943), and studies using the var-
ious techniques have led to the significant conclusion that the extent of chemical
weathering can generally be estimated from the mineral assemblage con-
tained in a soil as compared to the minerals of the parent material. Further-
more, this conclusion provides a basis for the determination of relative ages
of different soils, which is a major concern of most Quaternary geomorphol-
ogists.
In detail, the correlation between the differentmethods used to determine

mineral stability is far from and the factors that control mineral sta-
perfect,
bility are not clearly defined. For example, the relationship between stability
and silicate structure recognized by Goldich is not as strong as we often think.

Plagioclase feldspars show considerable variation in stability depending on the
mineral variety, even though they all possess the same framework structure.
Micas show the same type of discrepancy. In addition, zircon appears to be
an extremely persistent mineral under weathering, yet it is structurally the
same as olivine, a notably unstable mineral species. Despite these apparent
difficulties, the use of mineral stability estimates is still an acceptable tech-
nique to gauge the rate and degree of decomposition.
Secondary Minerals The minerals just discussed are components of the orig-
inal rock, commonly referred to as primary minerals. It is well to remember,
however, that the processes of chemical weathering play a dual role. In ad-
dition to destroyingprimary minerals, they also create new minerals by re-
combining or reprecipitating materials liberated from the parent rocks. These
secondary minerals, born within the weathered zone, are distinctly more stable
than their primary ancestors because they reach equilibrium in the temper-
ature-pressure environment of the soil rather than in the magmatic or meta-
morphic conditions that created the original crystals. The most common
secondary products of weathering are clay minerals and amorphous hydrous
oxides of iron, aluminum, silica, and titanium.
Clay Minerals The advent of X-ray diffraction and electron microscopy

provided scientists the capability to examine the actual lattice structures of
clay minerals. As a result, clay mineralogy has developed into a separate sci-
entific discipline, the details of which are well beyond the purposes of this
book. Nonetheless, clay minerals are important indicators of the degree and
character of weathering, and they are significant components in the physical
and chemical attributes of soil. As such they deserve mention, albeit brief and
oversimplified.
Clay minerals are aluminum silicates in which silica tetrahedra and alu-
minum octahedra are bonded together into a layered atomic structure. A silica
tetrahedron has one silicon atom surrounded by four oxygen atoms; an alu-
minum octahedron consists of a single aluminum atom bonded to six oxygen
atoms. In clays the individual tetrahedrons and octahedrons are linked to-
gether in planes, forming distinct sheets or layers typified by either a tetra-
hedral structure (silica sheet) or an octahedral structure (alumina sheet).
Most clays are either a 1:1 layer silicate or a 2:1 layer silicate (table 3.3).
A 1:1 structure has as its fundamental building block one layer of silica tet-
rahedra and one layer of aluminum octahedra. In 2:1 mineral structures, one
aluminum octrahedral layer is positioned between two layers of silica tetra-
hedra. Other clay varieties do form, but they either represent combinations
of the basic types (mixed-layer) or originate under very special weathering
conditions (chain silicates).
Chapter 3 80
Table 3.3 Classification of the clay minerals.
Structure Common Minerals
cz 3 11 Kaolinite
Dickite
Silica sheet Nacrite
Crystal unit
o ">
Alumina sheet Halloysite (4H 2 0)
to ro Little or no internal adsorption Allophane

"to = Silica sheet
w > Crystal unit
E
o Alumina sheet
Lattice structure of kaolinite
Mica group
2:1 Muscovite, 2M
Mite
Silica sheet Glauconite
Alumina sheet Crystal unit Biotite
Vermiculite
o <» S Silica sheet
"D <J Distanc
Internal adsorptive surfaces Smectite group
« £ variable
Montmorillonite
Silica sheet
Many other types
Alumina sheet - Crystal unit
group
\ Silica sheet
Chlorite
Chlorite
Chamosite
Lattice structure of montmorillonite
Reprinted with permission of Macmillan Publishing Company from The Nature and Properties of Soils. 6th ed . by Harry O Buckman and Nyle C Brady Copyright © 1960 by
Macmillan Publishing Company
The mineral kaolinite represents the 1:1 clay mineral group and has the
greatest importance in soils. As shown in table 3.3, kaolinite is constructed of
crystal units consisting of one tetrahedral sheet (silica sheet) and one octa-
hedral sheet (alumina sheet). These two layers are held together by ions that
are mutually shared by aluminum and silicon atoms in the separate sheets.
Importantly, the bonding between adjacent, two-sheet crystal units is also very
strong. The bonding strength between the units is significant because it keeps
the lattice spacing fixed, thereby preventing expansion of the structure when
the clay is wetted. As a result, cations and water do not penetrate between the
kaolinite crystal units. This factor explains the relatively low cation exchange
capacity of kaolinite because the ion exchange process is restricted to the ex-
ternal surfaces of the mineral. The fixed structure also accounts for the low-
plasticity (the capacity to be molded) and swelling characteristics of the min-
eral.
Of the minerals possessing a 2: 1 structure, the smectite group and the

mica group have the greatest significance in weathering and soil development.
1
The smectite group is characterized by ion substitutions that occur primarily

in the octahedral layer. This leads to wide variations in chemical compositions
of the group minerals. Montmorillonite, the most common type of the smec-
tite group, has the lattice structure shown in table 3.3. The obvious difference
from kaolinite is that montmorillonite has a three-sheet crystal unit. Like ka-
olinite, the three layers comprising the unit (two tetrahedral sheets and one
octahedral sheet) are bound tightly together by shared atoms. In contrast to
kaolinite, however, thebonding between adjacent crystal units is notably weak,
and the mineral lattice is capable of expansion upon wetting. Therefore, cat-
ions and water molecules easily penetrate the mineral interior where cation
exchange takes place along the surfaces of the crystal units. The result is a
clay with greater plasticity and swelling and a cation exchange capacity that
is much higher than can be generated when only external exchange sites are
available.
Illite is the most common mica-group mineral found in soils. Because its
lattice is 2:1, illite is structurally similar to montmorillonite. It differs from

montmorillonite, however, in that some of the silicon in the tetrahedral sheets
has been replaced by aluminum. This creates an unbalanced charge in the
silica sheets, a charge that is primarily satisfied by potassium ions placed be-
tween the crystal units. As a result, the bonding between the units is stronger
in illite than in montmorillonite. Therefore, the values of c.e.c, swelling, and
plasticity in illite are lower than those found in smectite clays, but they still
exceed those found in kaolinite.
The mineral vermiculite, also part of the mica group, is found in many
+
soils. It differs from illite because it normally has little K
in the interlayer
zone. Instead, that position is more likely to be occupied by Ca ++ or Mg ++ .
These cations do not provide the strong bonding noted in illite and mica; there-
fore, vermiculite is prone to have some expansion.
Other clays found in soils have variations on the structures we have been
discussing. Chlorite, for example, has a crystal unit consisting of alternating
2: 1 layers and octahedral layers. Therefore, it is often designated as a 2: 1 :
clay. Bonding in chlorite is strong because ion substitutions create opposite

charges in making the lattice nonexpanding. In addition, mixed-
the layers,
layer clays can form when different crystal units interstratify with one another
or with hydroxides of magnesium or aluminum. As you might expect, the
physical and chemical properties of mixed-layer clays are quite variable.
Differences in stability are present even within the realm of clay minerals.
Jackson and his colleagues (1952) proposed a distinct sequence of clay min-
eral development in which muscovite or illite is the initial product of weath-
ering and, assuming effective leaching, progressively degrades through
montmorillonite to a final kaolinitic clay (fig. 3.9). The suggestion that ka-
olinite should be the most common end product under normal conditions has
received support from evidence other than mineral indices. Feth and his co-
authors (1964) found that the composition of groundwater draining the gran-
itic rocks of the Sierra Nevada was chemically in equilibrium with kaolinite.
This suggests that other clay minerals in the weathered residuum are unstable
and will ultimately change to kaolinite.
Chapter 3 82
Hot wet climates ( — Si)
Microcline
Orthoclase
c and Others
X
(0
CD
W Muscovite
Oxides -
C/3
Micas
Fe& Al
c CD
e U-
Biotite
°8
<
Z
CO
O Soda-lime
d> Feldspars
^ Augite
Hornblende
and Others
X
Degree of weathering increases
Figure 3.9.
Diagram showing the general It should be recognized, however, that kaolinite itself is a complex mineral
conditions for the formation of the
having a variety of origins that differ from the evolutionary process earlier
various silicate clays and the
oxides of iron and aluminum In suggested. Kaolinite may derive directly from the weathering of feldspar
each case, genesis is without being the end stage in the degradation of other secondary minerals
accompanied by the removal of
soluble elements such as K, Na.
(Keller 1978, 1982). Additionally, when kaolinite forms directly from feld-
Ca, and Mg. spar, it seems certain that the mineral develops in an intervening solution phase
rather than by solid-state transformation or replacement. Nonetheless, as-
suming that a sequential development of clay types is possible, the type and
amount of clay should provide a reasonable basis for estimating the degree of
weathering in a soil; such estimates have in fact been made (Barshad 1964).
It should also be apparent that specific clay minerals will tend to be dominant
in mature soils developed under a particular set of climatic conditions.
The apparent sequential development of clays emphasizes the important

point that clay minerals themselves are highly reactive and readily alter from
one form to another. In most weathering situations, the composition of the
initialclay developed is usually controlled by the ions available in the parent
rock, and many of these ions may be highly mobile. With time, however, cli-
mate and other factors operating within the weathering zone exert a greater
influence, and clays will alter to forms that are more nearly in equilibrium
with the chemical environment.
A note of caution is perhaps needed at this point. In some situations all
clay minerals in a soil are not necessarily derived directly from weathering of
Colman (1982) examined the clay-sized material in the
the parent materials.
weathered outer fringe (called a rind) of basalt and andesite boulders con-
tained in clay-rich soil horizons at least 100,000 years old. The material in
Table 3.4 Gains or losses of chemical constituents in a hypothetical example of

weathering.
Original Rock Saprolite Adjusted % a

Loss or Gain 3
Si0 2 503 41.3 24 78 -25.52

AI 2 O3 18.3 306 183
Fe 2 3 25 11.3 678 + 428
FeO 9.4 1.2 072 - 868
MgO 55 - 55
CaO 126 -126
Ti0 2 1.1 0.1 0.06 - 1.04
H2 0.2 14 6 8.76 + 8.56
Total 99 9 99 1 59 40 -40 5
^To obiam loss or gain multiply each value in the saprolite column by 6 Subtract the adjusted percentage from the
original % in the fresh rock
"Al^ (Uest\)/M203 (saprolite) = 18 3/30 6 = 06
the rinds is predominantly disorganized or amorphous allophane and iron ox-

ides or hydroxides. However, in the surrounding soil itself the clay fraction
consists of highly crystalline, well-developed clay minerals. Because of the rind
composition, it appears likely that the clay minerals in the soil horizon are not
the result of simple transitions from weathering of the boulders. Instead, the
clay minerals were probably introduced to the soil as fully developed crystal-
line forms by some external process.
Hydrous Oxides In addition to clay minerals, a number of oxide and hy-

droxide compounds are formed in the weathering zone as important secondary
minerals. The oxides are normally sesquioxides of iron and aluminum occur-
ring mainly as hematite, Fe 2 3, or its hydrated form limonite, Fe2 3 • H 2 0,
and gibbsite, A1 2 3 • 3H 2 [or preferably Al(OH) 3 ]. In general, these min-
erals are crystalline, but when a poorly ordered ion arrangement exists in the
amorphous. The sesquioxides are usually
structure, they are considered to be
stable end products of weathering unless the chemical environment becomes
strongly acidic.
Estimates Based on Chemical Analyses The degree of chemical weathering

has also been gauged on the basis of total chemical analyses. The most common
approach utilizing such data is a direct comparison of the fresh parent rock
with the saprolite or soil derived in situ from it. To make an analytical com-
parison of the two, A1 2 3 is usually considered to be constant since its mobility
is extremely low under normal conditions. The weight percent values of ox-
ides in the soil are recalculated by using the conversion factor %A1 2 3 (fresh
rock) / %A1 2 3 (weathered material) as a multiplier of each constituent
(Birkeland 1974). Table 3.4 demonstrates this method by showing a hypo-
theticalcomparison of a saprolite developed from a parent metagabbro. The
recalculated values indicate the degree of weathering, and relative gains and
losses from the original composition can be used to compare nearby exposures
or different horizons within the same soil profile. Maximum errors in this
Chapter 3 84
method occur when the assumption of a completely constant aluminum value

is incorrect. As mentioned earlier, aluminum may be mobilized by chelators
or humic acids, and so it is likely that some aluminum will be lost from the
upper weathering zones and some gained in the lower horizons. Nonetheless,
the method provides a good first approximation of the relative degree of
weathering, even though absolute gains and losses are suspect because of the
aluminum problem.
The rate of chemical weathering is also commonly estimated by chemical
analyses of water that has filtered through the rock and soil system. In general,
these studies rely on the assumption that the parent material is the only source
of dissolved constituents in the water. However, because rainwater is not
chemically pure, some knowledge of its composition is required before valid
conclusions about solution within the weathering profile can be made. Cleaves
and his colleagues (1974), for example, estimate that about 37 percent of dis-
solved solids in a small Maryland stream were introduced into the system as
part of the precipitation. Even when the chemistry of rainwater is known, other
factors may lead to faulty estimates of solutes produced by weathering. Cyclic
wetting and drying of soils in arid climates may cause partial re-solution of
salts that earlier were precipitated in the soil (Drever and Smith 1978). Ad-
ditionally, ions are often brought in or removed from a soil by atmospheric
dry fall (deposition not associated with precipitation), vegetation, and land
uses such as irrigation, fertilization, and timbering. Therefore, to be totally
valid the water chemistry approach should be comprehensive in scope, with
all factors causing input or removal of elements being considered (Likens et al.
1977; Reidet al. 1981).

If corrections can be made for precipitation input and effects of the other
factors just mentioned, the dissolved load in rivers should represent the weight
loss of rock material from a drainage basin. Weights can be converted into
volumetric terms, and when considered over a reasonable period of time, rates
of chemical weathering can be expressed in terms of surface lowering or in
soildevelopment (Hembree and Rainwater 1961; Cleaves et al. 1974; Owens
and Watson 1979). In larger areas the rate of chemical denudation also ap-
pears to be a function of rock type and climate (Livingstone 1963; Judson and
Ritter 1964; Strakhov 1967), increasing in magnitude in warm-humid re-
gions. Rapp (1960) demonstrated, however, that solution may be the domi-
nant erosional process even in arctic regions, and therefore, chemical
weathering is a factor that simply cannot be ignored in the overall scheme of
landscape development.
Soils Weathering processes that continue over an extended period of time result in
an unconsolidated mass of soil that is measurably different from the original
rock in its physical and chemical properties. Soils also contain a significant
amount of organic material that is added progressively to the system as de-
caying vegetation and microorganisms. In addition, pronounced layering de-
velops in the weathered mass during the transition from being simply
decomposing rock or alluvium to being a true soil. The vertical arrangement

of the layers constitutes a diagnostic property of soils known as the soil profile.
The profile extends from the surface downward to the fresh parent material.
The time in absolute years needed to form a soil profile, as well as the per-
fection of the profile's development, varies widely with the intensity of the
weathering processes and the character of the original material. Nevertheless,
where the soil profile is well developed, its character reflects the environment
under which it formed and serves as the basis for classification of soils and
interpretation of paleo-conditions.
The Soil Profile

In its simplest form, the soil profile can be visualized as consisting of three
main layers, usually designated as the A, B, and C horizons. The A horizon
is normally considered to be the thin, dark-colored surface layer where organic
matter is concentrated and where clays and mobile components are continu-
ously leached downward or eluviated. The C horizon is usually thought of as
the underlying parent material which is essentially unmodified by the soil-
forming processes. The B horizon, therefore, becomes the transitional zone
between the A and C horizons and historically has been considered as an il-
luviated zone, i.e., a zone of accumulation and concentration of the material

brought down from the A horizon. Each horizon, defined on the basis of its
physical and chemical properties, shows enough internal variation to require

subdivisions indicating some special trait is present.
A variety of properties can be used to distinguish the horizons and zones
of a soil profile. In addition to characteristics already discussed, such as pH,
Eh, and c.e.c, the most important criteria are color, texture and structure,
organic content, and moisture characteristics (Birkeland 1974). Color in soils
is an indicator of high organic content (black, dark brown), ferric iron (yellow-
brown to red), or a concentration of Si0 2 or CaC0 3 (light grey to white).
Small amounts of a pigmentor can cause rather intense discoloration, however,

and so color alone may be a poor index of the total quantity of the pigmenting
substance. Texture is simply the relative proportions of different particle sizes
in a soil horizon, analogous to the property of sorting as used by geologists
(fig. 3.10). Structure in soils, however, is a unique characteristic in that it
designates the shape developed when individual particles cluster together into
aggregates called peds (fig. 3.1 1). In clay-rich soils, the openings between peds
may play an extremely important geomorphic role by providing the only av-
enues for downward percolation through an otherwise impermeable soil. Or-
ganic matter in soils consists mainly of dead leaves, branches, and the like,
called and the amorphous residue, called humus, that develops when
litter,
decomposed. Litter may form at mean annual temperatures as low as

litter is
freezing, but its optimum production occurs at about 25°-30°C and decreases
rapidly above those levels. Microorganisms that convert litter to humus begin
to function at temperatures slightly above freezing (5°C), but the optimum
Chapter 3 86
Figure 3.10.
Percentages of clay (< 002
mm), silt (0.002-0.05 mm), and
sand (0 05-2.0 mm) in basic soil
textural classes as defined by the
U.S. Department of Agriculture.
(From Soil Survey Staff 1951)
100
Figure 3.11.
Major types of soil structure:
(A) prismatic; (B)columnar;
(C) angular blocky; (D) subangular
blocky; (E) platy, (F) granular
(From Soil Survey Staff 1951)
(F)
temperature for their life activities may be as high as 40°C (fig. 3.12). It is
significant that at temperatures between 0° and 25°C, humus is produced in

abundance, but above 25°C little if any humus is accumulated. Humus has
an important effect on soil formation because it includes chelators and in-
creases water absorption. In addition, the development of humus releases CO;
in high concentrations, leading to unusual amounts of carbonic acid within
the humic zone and an associated lowering of the pH.
The total quantity of water that can be held in a soil is the available water
capacity (AWC). This characteristic is very difficult to estimate in the field,
and the lack of a consistent laboratory procedure for its computation has led
to some confusion as to its relevance (Salter and Williams 965). Nonetheless,
1
by combining this parameter with the bulk density (dry weight of soil/unit
Figure 3.12.
Production and destruction of
organic matter in humid climates.
The difference between the

amount produced and the amount
destroyed by microorganisms
controls the accumulation of
humus
0° 5° 10 15 20 J
25 30 35 40° 45 C
30 40° 50 60 70' 80 90° 100 110 120 F
——— A = organic raw material - B == destruction of organic matter
production in humid climate in an aerated environment (aerobic)
volume), an estimate of the depth of wetting can be made (see Birkeland 1974).
Such information is significant in that it relates to many soil properties, es-
pecially those affected by redistribution of material during downward per-
colation of water.
Available water capacity can be calculated if the moisture content at an
upper limit and a lower limit {permanent wilting point) are

(field capacity)
known. The field capacity is determined by allowing a saturated sample to
drain by gravity for at least 48 hours, by which time the remaining water
content is held by adhesion to mineral and organic particles. After field ca-
pacity is reached, water can still be taken from a soil by the normal functions
of plants. If plants continue to extract water to the extent that no more can
be removed, the vegetation wilts. The water remaining in the soil is held with
tensional stresses too high for the plants to break, and this amount of water
represents the permanent wilting point. Both field capacity and permanent
wilting point are expressed as a weight percentage according to the following
equation:
P* = w -w s d
X 100
where P H is moisture percentage, W s is total soil weight, and W d is weight of

soil after drying at 105°C. The available water capacity is simply the differ-
ence between the moisture content at field capacity and that at the permanent
wilting point.
Assuming that some criteria are distinct enough to identify a soil horizon
or zone, the pertinent consideration then becomes what nomenclature should
be used to convey that information. In recent years the simple designation of
A and B horizons as zones of eluviation and illuviation has been seriously ques-
tioned. Soil scientists have found that some soils contain more than one eluvial
or illuvial horizon, obviously necessitating that either an eluvial zone be placed
in the B horizon or an illuvial zone be placed in the A horizon. In addition, it
Chapter 3 88
Table 3.5 Nomenclature of soil horizons
Horizon 3 Characteristics
O Upper layers dominated by organic material above mineral soil horizons

Must have > 30% organic content ifmineral fraction contains >
50%
clay minerals, or > 20% organics if no clay minerals
A Mineral horizons formed at the surface or below an O horizon. Contains

humic organic material mixed with mineral fraction. Properties may
result from cultivation or other similar disturbances.
E Mineral horizons in which mam characteristic is loss of silicate clay. iron.

or aluminum, leaving a concentration of sand and silt particles of
resistant minerals.
B Dominated by obliteration of original rock structure and by illuvial

concentration of various materials including clay minerals, carbonates,
sesquioxides of iron and aluminum Often has distinct color and soil
structure.
C Horizons, excluding hard bedrock, that are less affected by

pedogenesis and lack properties of O, A, E, B horizons Material may be
either like or unlike that from which the solum presumably formed.
R Hard bedrock underlying a soil.
Adapted from the Soil Survey Staff (1960. 1975. 1981)

a Honzons can be divided into subhonzons indicated by Arabic numbers such as B1 A2. B12, etc
,
is now clear that concentrations of clays or sesquioxides are not always caused
by illuviation but may result as a lag when other materials are preferentially
removed. Because of these and other complications, soil processes and char-
acteristics are often incompatible with the general sense of the A, B, C no-
menclature; in response, the Soil Conservation Service (S.C.S.) of the U.S.
Department of Agriculture proposed new terms to indicate the diagnostic soil
horizons (Soil Survey Staff 1960, 1975). The horizons established by the S.C.S.
(table 3.5) may not correlate with the older system; in fact, their definitions
are so precise that laboratory analyses may be required before certain zones
or horizons can be identified.
Surface master horizons are called O or A depending on the amount of
organic matter they contain. The A horizon can be further defined as mollic,
umbric, or ochric on the basis of criteria established in the S.C.S. system.
Mollic A horizons are dark colored, contain more than 1 percent organics,
have a base saturation of more than 50 percent, and usually develop under a
grassland vegetation. Umbric A horizons are similar to mollic types but form
beneath forest zones and have less than 50 percent base saturation. Ochric A
horizons are light colored, contain less than 1 percent organics, and normally
develop under semiarid vegetation.
The E horizon underlies the O or A horizon. It is characterized by intense
leaching that removes Fe +3 or organic coatings from the mineral particles and.
therefore, is usually bleached gray in color. The B horizon has characteristics
that may reflect any or all of the weathering processes discussed earlier. It
commonly has a high clay content due to illuviation or to mineral growth in
situ, reddish hues, iron and aluminum concentrations as sesquioxides, and stable
Table 3.6 Some common descriptive symbols to be used in conjunction with

major soil horizons.
Symbol 3 Meaning
b buried soil horizon
g strong gleying
h illuvial humus
ir illuvial iron
kb accumulation of alkaline earth carbonates, commonly CaC03

m strong cementation
P plowing
t illuvial clay
X fragipan character
From Soil Survey Staff, 1975. 1981

^Symbols used with other profile designations For example, B2t, B1h, Cca
^Formerly designated as ca. a term many publications still use
primary minerals. The properties of the B horizon, however, can change mark-
edly with variations in the fundamental soil-forming controls such as climate
and parent material. The fact that the B horizon has such variable traits
prompted the descriptive subterms argillic (high clay content), natric (high
exchangeable sodium concentration), spodic (translocated organics and ses-
quioxides), oxic (hydrated oxides of Fe and Al and secondary silicates in-
cluding 1 clays), and cambic (well-developed structures with intense
: 1
oxidation and red color common).

The lowermost master horizons, C and R, exist below the B horizon. The
C horizon has no properties typical of the overlying horizons, but it has been
affected by weathering, as features such as oxidation show. It is composed of
unconsolidated material that may or may not be like the material from which
the soil presumably formed (Birkeland 1974). The R horizon is simply con-
solidated bedrock beneath the soil.
In his most recent edition, Birkeland (1984) retains the K horizon to in-
dicate a subsurface horizon characterized by extreme carbonate accumula-

tion. In addition, he suggests a number of formal subhorizons of the A, B, and
C horizons that have designations such as Bt, Bh, Cox, etc. These are not
reproduced here, but those of you interested in soils should refer to that work
for greater detail.
Three kinds of symbols are used to denote horizons and layers in a soil
profile. Capital letters as shown in table 3.5 designate master horizons. Low-
ercase letters are used as suffixes to indicate specific characteristics of layers

in the master horizon (see table 3.6), and numbers are used as suffixes to con-
note vertical subdivision within a horizon or layer. In addition, numbers are

prefixed to the master horizon designations to indicate a significant change in
particle size or mineralogy within the soil. These signify a difference in the
material from which the horizons have formed. In 1975 the S.C.S. used roman
numerals as the prefix but have since changed to arabic numerals (Soil Survey
Staff 1981).The number is never used because it is implied to represent the
1
Chapter 3 90
Section Soil Horizon'" Description™

Figure 3.13.
Hypothetical soil profile
A2 Fine loam, crumb soft, loose.
-
- . - ,—
10 Fine sandy loam, massive slightly
- Bit
hard. Color 7YR 4/2.
20 — — .'—
30 — r*
' :
'. — '
B21t Fine, sandy loam, massive hard.
_. _
</>
Fine sandy loam, massive hard,
2 40 :;7;";^-;~-.: :
^ '-y".
B22tk
<r> calcareous.
a- o_
E '{I
2B1k Sandy gravel, horizontally laminated.
1 50
CaCo 3 Color 10YR 7/1.
o .
.*'.••«' . -
.".
60
Sandy gravel. CaCo 3 coatings on
•'-•-.'• :'-••' •': 2C1k
pebbles.
"' "*,•". —
70 .
-".':-'-.
80 i (
CaCo 3
on I'M
1
'
1 ,
i
3C1k in cracks of granite.
(1) Lowercase k may now be used in place of ca.

(2) Color from Munsell color charts (see Soil Survey Staff. 1975. pp. 463-69 for discus-
sion).
material in the surface mineral zones. Therefore, if no changes occur down-
ward into the profile, prefix numbers are not needed. Transitional zones be-
tween the master horizons are indicated by the use of both capital letters. For
example, a zone transitional between the A and B horizons may have char-
acteristics typical of both. It is designated as AB or BA. the first letter de-
pending on which master horizon most closely resembles. The B2 zone is it
reserved for that portion of the horizon where the diagnostic properties iden-
tifying the B horizon are most prominently displayed. A second number is
sometimes utilized to stress minor variations in the soil. Figure 3.13 represents
a hypothetical soil profile using the current nomenclature. However, because
the S.C.S. designations seem to be in a perpetual state of flux, descriptive
notations shown in figure 3.13 may be changed in the future.
Soil Classification
The classification of soils is, like all classifications, simply an attempt by
someone systematically to group together and name soils that exhibit pro-
nounced similarities. The trait or traits being classified serve as the philo-
sophic basis for the groupings. The choice of diagnostic traits depends on what
the classifier deems important, and since individuals disagree on that point,
no classification will satisfy all members of the pedologic community.
The first real attempt at came from Russian soil sci-

soil classification
from the work of Dokuchaiev

entists in the late nineteenth century, especially
and his students. Although fraught with inconsistencies, Dokuchaiev's scheme
had great influence on American pedologists. His groups were defined in part
by climate and vegetation, and until recently these genetic criteria were ac-
cepted as the basis of all soils classifications used in the United States. In fact,
many Russian terms are still American soils nomenclature.
integrated in
In the United States, C. F. Marbut was the leading force in efforts to
systematize soils. Marbut's classification, developed over a span of years in
the 1920s and 1930s, was based on characteristics that are present only in
"mature" soils. The system, therefore, had no place for soils that were not
fully developed. A more refined classification, designed to rectify many of the
problems inherent in Marbut's system (Baldwin et al. 1983; Thorp and Smith
1949), became the most extensively used soil classification in the United States
until a dramatic shift in fundamental emphasis was introduced. In 1960 the
Soil Conservation Service completely revised the descriptive nomenclature and
the classification of soils (Soil Survey Staff 1960). It is important to note that
this revision is not simply a formulation of new class names (although that
occurred) but represents a fundamental disagreement with the philosophical
basis of earlier classifications. Until the new system was devised, all classifi-
cations were essentially genetic in scope; that is, the major soil classes were
based on climatic and vegetal factors. Although the subdivisions were linked
to observable aspects of the profile, these were not explicitly defined, and soil
scientists inescapably allowed their knowledge of climatic and vegetation dis-

tributions to influence decisions about placing a soil in a particular group. In
many cases, therefore, pedologists were classifying the genetic factors and not
the tangible resulting soil. The S.C.S. system is nongenetic. It requires no cli-
matic interpretation but is based on very specific, often quantitative, criteria.

We will describe it briefly, without becoming mired in its details.
Soils in the S.C.S. system are grouped into 10 orders distinguished by the
major horizons in their profiles (tables 3.7, 3.8; fig. 3.14). The orders are sub-
divided into suborders (table 3.7), defined by a diagnostic physical or chem-
ical soil property that in some cases may require quantitative laboratory data
to be recognized. The terms used to designate a particular suborder represent

the combination of two syllables; the prefix indicates the diagnostic property
of the suborder and the suffix reveals the order, since it is composed of several
key letters of the order name (tables 3.7, 3.9). An Argid, for instance, is an
Aridisol with an argillic horizon, and a Udent is a moist Entisol with a low to
moderate organic content. Further subdivision into great groups and subgroups
is made on the basis of even more detailed properties than those differentiating
the suborders. For example, within the Udent suborder, a great group char-
acterized by a low temperature is a Cryudent, the prefix "cry-" inserted to
indicate coldness. Subgroup nomenclature requires an additional descriptive
word placed before the great group name. A Cryudent with a weakly devel-
oped spodic horizon becomes a Humodic Cryudent.
Chapter 3 92
Table 3.7 Orders and suborders of the Soil Conservation Service classification
Order Suborder Order Suborder
ENTISOL Aquent SPODOSOL Aquod

Psamment Humod
Ustent Orthod
Udent Ferrod
VERTISOL Aquert ALFISOL Aqualf

Ustert Altalf
Udalf
INCEPTISOL Aquept
Ustalf
Andept
Umbrept ULTISOL Aquult
Ochrept Ochrult
Umbrult
ARIDISOL Orthid
Argid OXISOL
MOLLISOL Rendoll
At boll
Aquoll
Alton
Udoll HISTOSOL
Ustoll
From Soil Survey Stall 1960
Table 3.8 Formative elements in names of soil orders in the Soil Conservation
Service classification.
Formative Mnemonicon and

No. of Name Element in Derivation of Pronunciation of
Order* of Order Name of Order Formative Element Formative Elements
1 Entisol ent Nonsense syllable recent

2 Vertisol ert L verto, turn invert
3 Inceptisol ept L inceptum. beginning inception

4 Andisol . ... id L andus, dry. arid
5 Molhsol oil L mollis, soft. mollify
6 Spodosol od Gk spodos. wood ash Podzol odd

7 Alfisol alf Nonsense syllable. Pedalfer
8 Ultisol ult L. ultimus. last Ultimate
9 Oxisol ox F oxide, oxide oxide
10 Histosol ISt G histos, tissue histology
From Soil Survey Staff 1960

a Numbers
of Ihe orders are listed here for the convenience of those \ ho became familiar with them during development of
item of classification
Soils not Figure 3.14.

dominated by 1. Entisols Simple classification of mineral
A horizon swelling clay and organic soils of the world
rests on -^
based on So/7 Taxonomy by the
C Soils dominated USDA (Reprinted by permission
Soils by swelling clay 2. Vertisols from Soil Science Simplified by
with few MiloI. Harpstead and Francis D.
features: Hole © 1980 by The Iowa State
A horizon |
Soils usually moist . . 3. Inceptisols
little
University Press, 2121 South
is over a
weathered State Avenue, Ames. Iowa 50010)
weak B I
Soils usually dry 4. Aridisols
Mineral
soils
(usually
have less i
than 25%
B horizon is
organic
enriched Thick, soft
matter)
in clay
black A1; .
5. Mollisols*
fertile
Soils with J Soils usually
distinct B ] moist Fertile C. . 6. Alfisols
horizons;
Thin A1
moderately over A2
Infertile C 7. Ultisols
weathered
B horizon is enriched
in organic matter 8. Spodosols

and iron
Soils highly
weathered 9. Oxisols
Organic
soils
(usually
have more 10. Histosols
than 25%
organic
matter)
"Some Mollisols have a weak B horizon; not enriched by clay.
It is obvious even from a brief description that the S.C.S. system allows
forextremely detailed classification of soils. However, with 10 orders, 29

common suborders, and countless great groups and subgroups, the number of
possible combinations of terms is enormous; the system requires an intimate
understanding of soil properties for its successful use. Table 3.10 is presented
simply to show the approximate equivalents between the major soil categories
in the S.C.S. classification and the older classification.
Chapter 3 94
Table 3.9 Formative elements in names of suborders in the Soil Conservation

Service classification.
Forma- Derivation of
tive Formative Mnemon- Connotation of Formative
Element Element icon Element
acr Gk. akros, highest acrobat Most strongly weathered

alb L albus, white albino Presence of albic horizon (a
bleached eluvial horizon)
alt L altus, high altitude Cool, high altitudes or latitudes
and Modified from Ando Ando Ando-like
aqu L aqua, water aquarium Characteristics associated with
wetness
arg Modified from argillic horizon; argillite Presence of argillic horizon (a
L. argilla, white clay horizon with illuvial clay).
ferr L ferrum, iron. ferruginous Presence of iron

hum L humus, earth humus Presence of organic matter
ochr Gk. base of ochros, pale ocher Presence of ochric epipedon (a
light-colored surface)
orth Gk. orthos. true orthophonic The common ones
psamm Gk psammos, sand psammite Sand textures
rend Modified from Rendzina. Rendzina Rendzina-like.
ud L. udus, humid udometer Of humid climates
umbr L. umbra, shade umbrella Presence of umbnc epipedon (a
dark-colored surface)
ust L. ustus. burnt. combustion Of dry climates, usually hot in
summer.
Table 3.10 Soilorders of the Soil Conservation Service classification and

approximate equivalents of the Great Soil Groups
Orders in S.C.S.
Classification Approximate Equivalents
1 Entisols Azonal soils, and some Low-Humic Gley soils
2 Vertisols Grumusols
3. Inceptisols Ando, Sol Brun Acide. some Brown Forest. Low-Humic Gley, and
Humic Gley soils
4 Andisols Desert, Reddish Desert, Sierozem, Solonchak, some Brown and
Reddish Brown soils, and associated Solonetz
5 Mollisols Chestnut, Chernozem, Brunizem (Prairie). Rendzinas, some Brown,
Brown Forest, and associated Solonetz and Humic Gley soils
6 Spodosol Podzols. Brown Podzolic soils, and Ground-Water Podzols
7 Alfisols Gray-Brown Podzolic, Gray Wooded soils, Noncalcic Brown soils
Degraded Chernozem, and associated Planosols and some Half-
Bog soils
8 Ultisols Red-Yellow Podzolic soils. Reddish-Brown Latentic soils of the US .
and associated Planosols and Half-Bog soils

9 Oxisols Latente soils. Latosols
10. Histosols Bog soils

Pedogenic Controls and Regimes

Soils can be grouped together systematically on the basis of profile similari-
ties. from a unique combination of certain factors that
In general, these result
control the magnitude and types of soil-forming processes. The external fac-
tors of climate, biota, topography, parent material, and time have long been
recognized as the prime controls, and they are commonly expressed as the soil
equation (Jenny 1941):
S or s —f (cl, o, r, p, t . . .)
where 5 = soil, s = any soil property, cl = climate, o = biota, r = topography,

p = parent material, and t = time. Although the soil equation cannot be solved
in quantitative terms (for discussion see Yaalon 1975), the relative effect of
each factor has been determined, with some difficulty, by analyzing field sites
where four factors are held essentially constant and one is allowed to vary
(Jenny 1941).
The implication here is that soil-forming factors are independent vari-
ables, a perception that is probably invalid except for the factor of time. For
example, numerous cases can be cited to demonstrate the interdependence of
the various factors; climate directly controls vegetation and animal form, rock
type (parent material) influences relief, and relief often affects climate. In
addition, other difficulties arise because many soils are polygenetic; that is,
they developed under more than one set of controlling factors. In such cases,
the soil formed under the prior conditions becomes the parent material that
will respond to the new controlling factors. Therefore, because of the inherent
difficulties in isolating individual factors, it seems likely that solution of the
soil-factor equation in quantitative terms that will satisfy all soil scientists
may not be possible. Nonetheless, a qualitative understanding of which factors
dominate particular environments may allow us to predict how soil properties
will change from one region to another.
Parent material is usually considered to be rocks that are weathering in
situ and unconsolidated sediment that was transported from its place of origin
by various surficial processes and deposited at a different locality. It is gen-
erally conceded that parent material exerts its greatest control in the early
stages of soil development or in very dry regions (Birkeland, 1974). However,
highly impermeable materials or those having relatively stable minerals may
maintain an influence on soil properties for longer time intervals or in other
climatic zones. Perhaps the best demonstration of the effect is where textural
and/or mineralogical variations of parent rock exist in the same climatic and
topographic setting. For example, in the piedmont zone of North Carolina,
soils developed from siliceous mica gneiss and granite gneiss are considerably
different from those developed on diortes and gabbros. The greater content of
ferromagnesian minerals in the mafic rocks led to formation of thick, clay-
loam A horizons and reddish B horizons. These are notably absent in soils
overlying the siliceous gneisses (see Buol et al. 1973 for discussion and ref-
erences).
Chapter 3 96
700
-
600
2 10 km
Soil mapping units
[7j^7n A (oxisols) E (oxisols)
B (ultisols and alfisols) F (ultisols)
G
F^^J C (mollisols) (ultisols)
k<C«| D (ultisols)
Figure 3.15. The relief factor essentially refers to the effect of local topography on soil
Soilcatena showing relationship
development. Therefore, consideration of this factor is really an attempt to
between slope and soil type
(Reproduced from Soil Science analyze soil changes brought about by the influence of slope. In this type of
Society of America Journal. analysis, geomorphologists have found utility in the concept of a soil catena.
Volume 41, 1977, p. 110, by
A soil catena consists of a group of soil profiles whose characteristics change
permission of the Soil Science
Society of America) gradually beneath a sloping surface. The profile changes result because vari-
ations in soil-forming factors are produced by differences in geomorphic pro-
cesses acting on the slope materials, especially drainage of groundwater,
transport of surface sediment, and removal of mobile chemical elements. The
steepness of the slope is important because it tends to control the magnitude
of sediment transport by wash, creep, and mass movements, and it also allows
underground water to move more rapidly through the soil. As a result, soils
in the steep upper-slope location tend to be freely drained, oxidized to red-
brown colors, and coarse-grained. Lower-slope soils, where gradients are less,
commonly have higher moisture contents because the material is poorly

drained. They tend to be clay-rich and blue-gray to gray in color. Midslope
profiles are often transitional in color and texture between the upper- and lower-
slope characteristics (fig. 3.15). Therefore, a catena manifests the interaction
of soil processes and slope processes, and for this reason, the concept is sig-
nificant in geomorphology.
The facility of a catena to reflect process is complicated because slopes
are not always underlain by a single rock type. In addition, variations in cli-
mate, duration of soil development, and landscape evolution will make the
meaning of catena profile changes more difficult to interpret. Nonetheless, the
Alfisols
[ _J
Aridisols
\ j
Entisols
Histosols
Mollisols
Figure 3.16.
catena concept can be successfully applied in any area regardless of geologic The major occurrences of the ten
complexity or climate if sufficient analytical care is exercised (for more com- soil orders in the United States.
plete discussions, see Young 1972; Oilier 1976; Gerrard 1981). The boundaries are approximate
and tentative.
In addition to steepness, slopes also affect soil development because their
orientation often promotes differences in microclimate and vegetation even
where the regional climate is the same. Due primarily to lower direct inso-
lation and temperature, north-facing slopes tend to have greater moisture con-
tents and vegetal cover than south-facing slopes. This leads to thicker A
horizons with a higher organic content.
The effect of climate and its related control on vegetation have been rec-
ognized historically as the most dominant soil forming factors (cl. o in the
Jenny equation). In most of the Great Soil Groups in earlier soil clas-
fact,
sifications (see table 3.10) were based on the genetic relationship between soil
and climate. Similarly, the distribution of S.C.S. soil orders in the United States
is clearly related to climatic zonation (fig. 3.16). Thus, climate and vegetation
can be used in a general way to predict what soil types may be expected in
any region. This follows because certain combinations of climate and vege-
tation (called pedogenic regimes) produce distinct trends in soil-forming pro-
cesses. Several examples of regimes can be used to illustrate this point.
Chapter 3 98
Podzolization The processes that result in the removal of iron and/or alu-
minum from the A horizon and their accumulation in the B horizon are re-
ferred to as podzolization. Podzolization is common in humid-temperature
climates, especially in soils that develop under forest vegetation. For example,
most soils in the eastern half of the United States where precipitation exceeds
70 cm annually are podzolized to some extent. To the west where precipitation
is less than 56 cm, another pedogenic regime (calcification) becomes domi-
nant. Podzolization produces a continuous translocation of mobile substances
as they are released from parent minerals. Because forest vegetation uses less
alkalis and alkaline earths than do grasses or shrubs, the surface litter is nor-
mally quite acidic. As a result, most soluble ions (Ca, Na, Mg) in the upper
+
soil are leached, and colloids of all types adsorb H Therefore, Spodosols,
.
Alfisols, and Ultisols, the soils affected by podzolization, are notably acidic.
Where podzolization is very active, an E horizon may be found overlying
a spodic B horizon. This bleached zone represents the ultimate form of pod-
+3
zolization because almost all the Al and Fe +3 is removed and the residuum
is highly concentrated in silica. Although the soil in a true podzol is acidic,
the E horizon can be developed where the pH of the zone is still within the
insolubility range of Al
+3 and Fe +3 Thus, most of the translocation of these
.
insoluble ions involves chelation rather than dissolution by colloidal acids.

Beneath the zone of maximum eluviation, Spodosols and Alfisols are usu-
ally brown or grey-grown, but as the temperature increases they grade pro-
gressively into Ultisols having yellow and red colors (see table 3.10). The
greater humus content in cooler regions, combined with the translocation and
redeposition of organic matter, accounts for the drabber color in the higher
latitudes. However, as the production of humus declines with increased tem-
peratures, the ferric content rather than the organic matter becomes the pre-
dominant factor in determining the color of the B horizon. The characteristic
color of podzolic profiles, therefore, changes from north to south in the eastern
United States.
The end products of podzolic weathering include resistant primary min-
erals, iron and aluminum sesquioxides or hydrates, and kaolinitic clays. Most
of the mobile ions have been removed from the system in solution, but some
cations may be retained by fixation in the clay mineral structures. Clays ac-
cumulate in the B horizon in any or all of three ways: ( 1 ) the component poly-
mers of clays can be carried in solution from the A horizon and recombined
in the B horizon; (2) clays can be leached as already formed minerals and
physically accumulated; or (3) clays can form in situ by chemical alteration

of parent minerals that were originally in the position of the B horizon.
Laterization The processes resulting in Oxisols constitute a pedogenic re-

gime called laterization. The conditions needed for these processes to function
are high rainfall and temperature, intense leaching, and oxidation. These con-
ditions commonly exist in tropical regions, especially where wet and dry sea-
sons alternate and drainage is good. It is now clear, however, that laterization
does not necessarily require a tropical climate since rapid leaching of the proper
lithologies can form Oxisols in a wide variety of climates, including subarctic.

Thus, to consider Oxisols as the zonal soil of the tropics may not be absolutely
correct.
Laterization is marked by two essential differences from podzolization:
(1) organic accumulation is inhibited; and (2) silica is leached from the system
in addition to the common mobile ions, leaving abnormally high concentra-
tions of hydrated oxides of iron and aluminum in the soil. In some cases silica
combines with available alumina form kaolinitic clays, but where silica is
to
more extensively leached the excess alumina is present as gibbsite Al(OH 3 )
or, more rarely, boehmite (AIO(OH)). The type of parent material and the
organic content exert a direct influence on whether the clays will be kaolinitic
or gibbsitic. Where granites are the parent rocks, less silica will be removed
during laterization because it exists as quartz rather than as amorphous Si0 2 .
Oxisols developed on granites, therefore, tend to have more kaolinite in their

profiles; quartz may be a notable end product. Where basalts are the parent
rocks, gibbsite tends to be more dominant because the more soluble amor-
phous silica is thoroughly leached, leaving little available to form the ka-
olinite. By observation, when limestone is the parent material, the main clay
mineral seems to be boehmite.

Laterization is perhaps the most poorly understood of the pedogenic re-
gimes. We know that iron and aluminum can be dissolved in acidic fluids, yet
the environments of their accumulation are very acidic. Precisely how these
substances are concentrated is not clear, and much controversy has arisen over
interpretations concerning the environment of laterization. In light of this it
may be worthwhile to stress again that although laterization should function
efficiently in warm, humid climates, Oxisols are not necessarily restricted to
the tropics. Such a binding inference can lead to rather hapless paleoclimatic
interpretations, since Oxisols can form in cool climates, and Spodsols, Alfisols,
and Ultisols are not uncommon in the tropics.
Calcification A third pedogenic regime, calcification, functions in sub-

humid to arid climates where precipitation is insufficient to drive the soil water
downward to the water table. Ions mobilized in the A horizon are reprecipi-
tated in the B horizon where zones of CaC0 3 are commonly developed. The
depth of the carbonate zone is a function of the annual precipitation (fig. 3.17)
and in general represents a first approximation of the vertical extent of leaching.
Arkley (1963), however, showed that this relationship holds only in regions
without pronounced seasonality of rainfall or orographic effects. Nonetheless,
for our purposes it seems useful to visualize the depth of the carbonate zone
as a gross index of the magnitude of calcification. Soils derived from those
processes all possess a K horizon or a calcium carbonate zone, but its position
within the soil profile rises proportionately with decreasing rainfall and/or the
efficiency of the calcifying processes.
Chapter 3 100
Rainfall (cm)
Figure 3.17.
50 75
The depth of carbonate zones in
soils as related to mean annual
rainfall
20 30
Rainfall (inches)
Calcification takes place most efficiently under a grass or brush vegetation

because such plants utilize large quantities of alkali and alkaline earths in
their life processes. When the plants die, the mobile ions are returned to the
soil as the litter is decomposed. The amount escaping this recycling process
depends on the precipitation, temperature, and related bacterial activity.
The two S.C.S. soil orders formed primarily by calcification are the Mol-
lisols and Aridisols. Mollisols occur mainly in the temperate zones of the United
States, especially in the northern Great Plains states (the Dakotas and Ne-
braska). In these soils the profile (fig. 3. 18) usually has an undifferentiated,
black A horizon ranging from 30 to 120 cm in thickness, which reflects the
high accumulation of humus in cooler temperature zones. The B horizon is
usually light-yellow to brown, and carbonate may exist there as disseminated
material or in a distinct layer at the base of the horizon. In more arid areas
the Aridisols become dominant. These soils are normally very immaturely de-
veloped; CaCO,, if present as a secondary accumulation, will be near or at
the surface. Generally, then, in cooler areas with less than 64 cm of annual
precipitation, increasing aridity is shown in the profiles by a progressively less
developed A horizon and a carbonate zone rising closer to the surface. Where
the temperature is higher, humus production is less and a red hue becomes
more prominent in the calcified soils. This gives rise to the reddish-colored
Mollisols and Aridisols.
Figure 3.18.
Generalized profile of a Mollisol in
A^ Dark, organic-rich horizon. eastern Great Plains region of the
Maybe acidic. United States Greater aridity
would decrease the size of the A
horizon and raise the carbonate
> 70 cm
zone to a higher level Depth
values will vary with local
conditions Horizons indicated by

ca may now be designated with a
Leached, yellow-brown lowercase k
B < horizon. Usually alkaline
but few carbonates.
Bca y 200 cm
or I
Zone of carbonate
Cca > 220 cm accumulation.
Partly weathered
CM
parent material.
Because of the immobility of the soluble ions, most clay minerals formed
in the calcification regime are montmorillonite or illite types. In some cases
kaolinite is abundant near the top of the profile where slightly acidic soil water
is present, especially where humus is abundant. The retention of the alkaline
earths, however, is not favorable for the development of kaolinite in the deeper
zones.
Finally, as stated earlier, the factor of time in soil development is probably
the only truly independent variable in the soil equation. There is no question
that the development of diagnostic soil properties is a time-dependent phe-
nomenon. Therefore, horizons, clay content, clay mineralogy, organic content,
and other parameters can be expected to change during the tenure of soil for-
mation. Recognize, however, that the value of each property changes at a dif-
ferent rate and enough time, each will eventually reach a condition
that, given
where the property no longer changes or its rate of change becomes negligible.
At that time, the property has attained a steady-state relationship with the
soil-forming environment. When all properties of a soil profile reach this con-
dition, the soil itself is said to be in a steady state. Birkeland (1984) presents
an excellent discussion about the amount of time needed for various diagnostic
properties to culminate in a steady state. As shown in figure 3. 1 9, an A horizon
with pronounced organic accumulation will generally reach a steady-state
condition before aB horizon can be recognized or while it is patently imma-
B horizons (Bt, Bca) require more time. A true oxic B horizon
ture. Diagnostic
presumably attains steady state only when all the weatherable minerals have
been altered to stable forms and, therefore, is the diagnostic horizon requiring
the most time for its complete development.
)
Chapter 3 102
Steady state Steady state

,
i i
o
.^
c Pf
z?
&/
1 o> f
o/ S/
c ID d)
o fc
a
o
<D (D
f
a>
> V /
*° 0)
D
lu a cw
No ne L None 1 1 1
10 10 1
10 2 10 3 10" 10 s 10 6
Time (years) Time (years)
(A) (B)
Figure 3.19.
Because diagnostic components of soil profiles form over variable amounts
Diagram showing the variations in
time to attain the steady state for of time, the major soil orders that are based on those properties should also
(A) various soil properties and be time-dependent (fig. 3.19). Thus, the suggestion by Birkeland (1984) that
(B) various soil orders.
the distribution of major soil orders should correlate with the ages of deposits
or landscapes on which they develop seems reasonable. In fact, that suggestion
is generally supported by the relationships between soil orders and deposit
ages in the United States.
Geomorphic Significance of Soils

Our discussion thus far has been a very brief treatment of the basic processes
involved in weathering and soil development. How is such information im-
portant and relevant to the study of geomorphology? There is no question that
soil characteristics affect geomorphic processes in a number of significant ways;
in fact, a unique soil property may dictate the mechanics of a surficial system.
Once such a property is identified, it provides critical information needed for
environmental control or for regional and local planning (McComas et al.
1969). For example, in certain situations soil properties are known to control
the stability of building foundations (Baker 1975), hydrologic response to pre-
cipitation (Cooley et al. 1973), and the permanence of road construction (Wei-
nert 1961, 1965). Thus, the importance of soils in physical geomorphic systems
should not be underestimated. However, because soil properties are altered by
time and climate change, soil formation has even greater importance in de-
ciphering the sequence of events in Quaternary history. In fact, one of the

greatest uses of weathering and soils is to establish relative ages of glacial
deposits and, by inference, the sequence of glaciations. The same principles
are employed outside of glaciated regions to suggest relative ages of deposits
and often as evidence of climate change. A variety of weathering and soil char-
acteristics have been used in this manner (see Burke and Birkeland 1979 for
a detailed discussion).
The use of soils in Quaternary geomorphology is a dual-edged sword. The
fact that soil-forming processes change with time is a basic ingredient of his-
torical interpretation. However, the fact that soil-forming factors are not con-
stant makes the record preserved in profiles more difficult to ascertain. This
is especially true in areas that have been affected by the pronounced climatic
fluctuations of alternating glacial and interglacial episodes. As the climate
changes, the dominant factors of soil formation in any given area must also
change accordingly. Thus, many soils preserve in their profiles characteristics

that reflect more than one set of soil-forming factors. These polygenetic soils
(sometimes called complex soils) complicate the record because thickness and
maturity of soils are reliable indices of age only when the conditions deve-
loping the soils have been maintained continuously. When conditions change,
the properties of the initial soil are supplanted by new characteristics; because
such alterations vary in the degree of completeness, complex soils are very
difficult to correlate with soils in other areas. A pre-Wisconsinan soil, for ex-
ample, will accrue properties related to the controlling factors at the time of
its formation. A change in those factors at some later time, perhaps post-Wis-
consinan, will obviously cause the original soil properties to be out of phase
with the new soil-forming environment. For all practical purposes, the pre-
Wisconsinan becomes the parent material on which the post-Wisconsinan
soil
climate and biota are working, and a younger soil profile is superimposed on
the older one. Separating the older soil from the younger one is a demanding
field problem that, in this example, would require lateral tracing of the com-
plex soil to a locale where Wisconsinan deposits intervene.

Soils that form on a landscape of the past are paleosols. They can be of
three types: (1) buried soils are developed on a former landscape and subse-
quently covered by younger alluvium or rock; (2) relict soils were not subse-
quently buried but still exist at the surface; and (3) exhumed soils were at one
time buried but have been reexposed when their cover was stripped by erosion
(Ruhe 1965). Buried soils are immediately recognized as paleosols. Relict and
exhumed soils are much more difficult to identify, because it must be proved
that their properties are inconsistent with the modern environment or that
their age is the same as the paleosurface on which they rest. Some additional
geomorphic or stratigraphic data are usually needed to substantiate those re-
quirements.
The most persuasive use of soils in Quaternary geomorphology arises when
soil characteristics are painstakingly integrated with sequences of landform
development and associated sedimentary deposits. An excellent example of
this approach is documented in detailed studies of the Rio Grande valley and
the adjacent slopes and intermontane basins near Las Cruces, New Mexico.
These by the U.S. Soil Conservation Service and known as
studies, sponsored
the Desert Soil-Geomorphology Project, began with the work of R. V. Ruhe
in the 1950s (Ruhe 1964, 1967) and culminated approximately 20 years later
in thorough syntheses of the geomorphology (Gile and Grossman 1979;

soil
Gile et al. 1981). A enormous project cannot be

detailed discussion of this
entertained here. However, a brief abstract of its procedures and results can
demonstrate how soils can be used in the analysis of Quaternary history.
The Desert Project area is physiographically divisible into distinct sub-
areas (fig. 3.20). and northwest margins consist of semiconnected
The east
mountains that rise up to 2750 meters in elevation. The valley border zone is
located in the valley of the Rio Grande river and is characterized by deposits
and surfaces formed by the river and tributaries during and after the Pleis-
tocene epoch. The flood plain of the modern river stands at approximately
'
Chapter 3 104
Jornada del
Figure 3.20.
Muerto Basin
Block diagram showing major Piedmont slope
Valley ^ -^T^^k San Andres
landforms of the Desert Project
border -*§&
\f;ilfJlcf-''iî-pOonaAna^-'\ floor -
—t^T^.^ Mountains
-1— ^?\, 5a" - : ;•
Valley Coalescent_ i^'
^Ssfer" Mountains S^;*-'"—
^,î^5;Organ
rf er races
_ ^>^^$^C~~ ^-i^^-'t ^
Piedmont
slope
Bedrock
1200 meters. Between the mountains and the valley (the piedmont slope) exist
a number of intermontane basins that have been affected by tributary and

slope processes that function on the valley sides. The basin or piedmont areas
are composed mainly of alluvial fan deposits that emerge from the mountains
and coalesce into smooth alluvial plains sloping gently toward the valley bottom.
In some cases the piedmont deposits are graded to the valley-bottom deposits,
but more often they and their related surfaces are not physically connected to
the valley-border sequence.
The regional history was controlled by repeated climatic changes that
produced alternating periods of deposition and erosion. In the valley zone, these
alternations resulted in a series of river terraces, the surfaces of which are
underlain by alluvium of the Rio Grande river (fig. 3.21). Episodes of down-
cutting by the Rio Grande initiated trenching in the tributary arroyos of the
piedmont and valley-border areas. This isolated many of the geomorphic sur-
faces that formed as the upper level of a depositional event. Since the middle
Pleistocene, five depositional episodes, separated by intervening downcutting.
have created a sequence of surfaces that document their relative ages. Soil
formation on the various deposits and surfaces began at different times.
Knowing the geomorphic setting and realizing that climate change was
synchronous and piedmont areas, it might be easy to assume that
in the valley
soil development was similar throughout the region. Actually, the meaning of
the preserved soils is very confusing because their developmental histories var>
on a local basis. In some zones, erosion has truncated diagnostic horizons of
-Torriorthents
Figure 3.21.
Terraces and river alluvium in the
valley border subarea near Las
Cruces, N.M
Camp Rice Formation

(fluvial facies)
a profile. At other localities, soil profiles of one age were buried by material
of a younger depositional event. For example, arroyo trenching initiated in the
Rio Grande valley did not proceed entirely to the mountain front prior to the
incidence of the next depositional phase. In such cases, deposits and soils in
the upper piedmont zone were buried by subsequent fan development (fig.
3.22). Additionally, alluvial plain deposits may not extend all the way to the
,4
axial valley. Furthermore, precise dating by C provided only a broad age
framework for events in the area because the amount of datable material was
limited. What results from these complications is that the history recognized
in deposits and surfaces of the valley border cannot be directly correlated with
those in the valley-side sequence, and soils developed in the two areas are dif-
ferent because of vagaries in local history, parent material, and microclimate.
The crux of this is that Quaternary history in the Desert Project area is
unintelligible without the integration of stratigraphy, soils, and geomor-
phology. The history of the valley-border and piedmont areas was finally linked
by relating the degree of soil development in each subarea to the relative age
of the deposits and geomorphic surfaces within that subarea (Gile et al. 1981).
This could be accomplished because at stable sites (no erosive disruption of
profiles) age becomes the key soil-forming factor. This was especially revealed
in the carbonate horizons, which become thicker and more indurated with age
(fig. 3.23). Thus, even though soils of the two subareas have different char-
acteristics and classifications, the relative degree of their development allowed

them to be correlated and the surfaces associated with the soils could be placed
into a developmental sequence (table 3. 11).
Chapter 3 106
Figure 3.22.
Torriflu vents - -Torriorthents
Cross section showing burial of
deposits and surfaces by younger
fan and terrace alluvium near Las
Cruces, N M
Organ II sl . Pedon 65-2
vq.Xv^ errace anuvium

Terrace alluvium
\^S-._ JPl§anJI)J
i
Camp Rice Fm. \^V Burled O-l surface

(piedmont facies) Arroyo
Channel
Figure 3.23.
Schematic diagram of the
diagnostic morphology for the
stages of carbonate horizon
formation in the two
morphogenetic sequences.
Carbonate accumulations are
indicated in black for clarity
III IV
Stage Gravelly sequence Nongravelly sequence
Thin, partial or Carbonate filaments and/or

complete carbonate coatings on grains.
faint
coatings.
Carbonate coatings Carbonate nodules separated

are thicker and there by low-carbonate material.
are some fillings in
interstices.
III Carbonate occurs essentially throughout the horizon, which is
plugged in the last part of the stage.

IV A laminar horizon has formed on top of the plugged horizon.
Holocene Pleistocene
1 07 Chemical Weathering and Soils
Table 3.11 General relationship between the degree of development in carbonate soils and geomorphic surfaces
in subareas of the Desert Soils Geomorphology Project area near Las Cruces, New Mexico.
Stage of Carbonate
Geomorphic Surface Accumulation (cf. fig. 3.23)
Valley Piedmont Nongravelly Gravelly

Age, yr B.P. or epoch border slope soils soils
Holocene
0-100 (Dune) 3 (Dune) 3
100-7500 Fillmore 6 Organ I
> 7500 (latest Pleistocene) Leasburg* Isaacks' Ranch II, III
Late Pleistocene Picacho Jornada II III, IV
Late mid-Pleistocene Jornada I Jornada I IV (multiple
Early mid-Pleistocene La Mesa laminar

zones)
From Gile 1975 Used with permission of Quaternary Research

3 The dunes are not formally designated by a geomorphic surface name
°Where Fillmore and Leasburg cannot be distinguished they are grouped inlo the Fori Selden surface
The processes of chemical weathering (hydrolysis, oxidation-reduction, solu- Summary

exchange) alter the exposed portion of the geologic framework and
tion, ion
combined with organic processes, produce soils. The degree of chemical change
depends on how mobile the ions of the parent minerals are under the external
and internal controls on the weathering mechanics. In regions with abundant
precipitation, highly mobile ions are usually removed from the weathered zone
unless the original mineral structure is incompletely broken down and ele-
ments such as potassium are fixed in the system. In contrast, where leaching
is minimal, mobile ions (Ca, Na, Mg) are concentrated in the weathered zone.
Immobile ions (Fe +3 Al +3 ) may be transposed in very acidic ground waters

,
or by special organic processes such as chelation. The mobility of most sub-

stances is also dependent on the pH and Eh of the weathering environment.
The type of clay mineral formed in the weathering zone is usually a good
indicator of the intensity of decomposition.

Soils are described and classified according to the soil profile. The char-
acter of the soil profile varies with parent material, climate, biota, topography,
and the length of time involved in its formation. Three major pedogenic re-
gimes — podzolization, laterization, and calcification —
produce the dominant
soil groups; however, changes in the controlling pedogenic factors may result
in complex soils that show evidence of forming under more than one pedogenic
regime. Soils are important elements in reconstructing geomorphic history,
and they directly influence other surficial processes.
Chapter 3 1 08
Birkeland, P. 1984. Soils and geomorphology. New York: Oxford Univ.

Press.
Gile, L. H.; Hawley, J. W.; and Grossman. R. B. 1981. Soils and
geomorphology in the Basin and Range area of Southern New

—
Mexico Guidebook to the Desert Project. New Mex. Bur. Mines Min.
Resources, Mem. 39.
Hunt. C. B. 1972. Geology of soils. San Francisco: W. H. Freeman.
Loughnan, F. 1969. Chemical weathering of the silicate minerals. New
York: American Elsevier.
Soil Survey Staff. I960. Soil classification, a comprehensive system — 7th
approximation. Washington, D.C.: U.S. Dept. of Agriculture. Soil
Conserv. Service.
. 1975. Soil taxonomy. Washington D.C.: U.S. Dept. of Agriculture,
Soil Conserv. Service. Agri. Hndbk. 436.
Physical Weathering, Mass
Movement, and Slopes
Htq ^ e r^
4
I. Introduction IV. Mass Movements of Slope
II. Physical Weathering Material
A. Expansion of Rocks and A. Slope Stability
Minerals 1. Heave, Creep, and Falls
1. Thermal Expansion 2. Slides
2. Unloading 3. Flows
3. Hydration and Swelling B. Morphology of Mass
B. Growth in Voids Movements
C. The Significance of Water Slope Profiles
III. Physical Properties of A. Slope Components and
Unconsolidated Debris Angles
A. Shear Strength 1. The Rock-Climate
1. Internal Friction Influence
2. Effective Normal Stress 2. Slope Evolution
3. Cohesion VI. Summary
B. Measurement of Strength VII. Suggested Readings
109
Chapter 4 110
Introduction The transformation of rocks into unconsolidated debris is the prime geo-
morphic contribution of weathering and soil-forming processes. Whether the
debris produced by weathering will resist erosion depends on the balance be-
tween the internal resistance of the materials and the magnitude of the ex-
ternal forces acting on them. The relative resistance ofanj_naiuralsubstance
is partJyjgflected in the characte rj)TThe_slOj?e that devel ops on it. Extremely
sfeepslopes, for example, can be maintained for long periods only if the un-
derlying rock or soil is so tightly bound together that the forces and agents of
erosion cannot lower the slope angle. On the other hand, gentle slopes in re-
gions of low relief and elevation may be stable for relatively long time spans
even if the underlying material is very friable. Cl early, then, slope character-
ist ics provide us with useful information only whe n we understand t he_erosive
_pjoces ses attacking them.

In a large sense, the evolution of landscapes is the history of regional slope
development. The formation of these slopes encompasses a multitude of geo-
morphic processes, and the properties of slopes reflect in subtle ways the tem-
poral effect of these processes on the resisting framework. Thus, interest in
slopes and slope-forming processes crosses the entire range of geomorphic
thinking, from the analysis of a modern stability problem to the abstraction^
of geologic history.
The mechanics of slope erosion are in many ways closely related to t
processes of physical weathering because the forces that disintegrate ro

and minerals simultaneously lower the internal strength of the unconsolida
cover. We will begin, then, with a brief discussion of physical weathering.
Physical Weathering Physical weathering culminates in the collapse of parent material and its dim-
inution in size. The continuing bjeakdownof rocktajc es place :wiign stress is
ejcert ed along zones of we aknesswithi n the original mate rial. These zones may
_bj_pjajiar_s tructure s^s uch as be ddings or fract ures t hat, upo n rupj ure~_pjodu ce
fragments whose ^ ize_and^ shape are controlled bytTfiTspacing of the planes.
In other cases, failure may occurâl ong min eral bou ndaries, resultingTrTan
accumulation of particles similar in size and shape to the original rock texture.
Although stresses are generated in different ways, the common bond that
unites all processes of disintegration is that in every case a force within the
material itself is responsible for its destruction.
The from either expansion of
stress field involved in disintegration results
rocks or minerals themselves or pressure generated by growth of a foreign
substance in voids within the lithologic fabric. In each method the direction
may change according to the process involved, but the
of the principal stress
most pronounced disintegration invariably occurs where the adjacent rocks
exert the least confining pressure. Intuitively, then, we should expect disin-
tegration to be most pernicious near the surface, where static load from
overburden is minimal and fractures are abundant and closely spaced. With
increasing depth, confining pressure increases, fractures are less common, and
the disintegrating processes become less effective.
111 Physical Weathering, Mass Movement, and Slopes
Expansion of Rocks and Minerals
Thermal Expansion Rocks and minerals expand in response to several phe-

nomena that can rightfully be considered as agents of physical weathering.
There seems to be little doubt that the application of intense heat can cause
physical disruption of rocks. The low thermal conductivity of rocks prevents
\t+4
the inward transfer of heat, allowing the external fringe of a rock mass to
expand significantly while little, if any, change occurs below the outer few
centimeters. Differential stresses are produced by this thermal constraint, and
the rock exterior spalls off in plates or wedges 1-5 cm thick. As figure 4.1
illustrates, this process functions during forest fires; in semiarid forested
mountains of the western United States, it may be the dominant process of
physical weathering (Blackwelder 1927). Whether or not insolation can drive
the process has been debated for many years. Many geologists have gradually,
if not grudgingly, accepted the premise that diurnal temperature fluctuations
are not severe enough to produce thermal spalling (Twidale 1968) because
experimental studies (Griggs 1936a, 1936b) suggested that the process is not
viable. Gray (1965), however, demonstrated that thermal spalling is indeed Figure 4.1.
possible, and geomorphologists have reaffirmed thermal expansion as a method Spalling of granitic boulder
caused by heat expansion during
of rock disintegration (Oilier 1963, 1969; Rice 1976).
forest fire Beartooth Mountains,
Mont
Chapter 4 112
Figure 4.2. Outer glacial

Expansion joints produced by valley
pressure release during valley
entrenchment Vaiont River valley,
Italy (From Kiersch, Civil
Engineer, 34, no 3, p 35, 1964

Used with permission by the
American Society of Civil
Engineers )
Bedding and steep
tectonic fractures
Old set Young set
Unloading Expansion of large segments of rock masses occurs when con-

fining pressure is released by erosion. As denudation removes overburden, the
stress squeezing the underlying framework is lowered, and rocks tend to split
into widely spaced sheets, 1-10 m thick, that are oriented perpendicular to
the direction of pressure release. The sheeting tends to mirror the surface to-
pography, and since outer sheets are relatively easy to erode, the process helps
perpetuate the surficial configuration because subsequent sheets develop with
a similar orientation. Although other processes aid in the removal of the sheets,
it can be readily documented that the original formation of the fractures is a
pressure-release phenomenon. Rock bursts in deep mines, for example, are
explicit proof that something as simple as excavation of tunnels can trigger a
rapid expansion of surrounding rocks. In the natural setting, postglacial en-
trenchment of the Vaiont River in Italy permitted valleyward expansion of
rocks and produced a joint system parallel to the valley sides (fig. 4.2). Hack
(1966) demonstrated that arcuate patterns of streams, ridges, and vegetal types
in the eastern United States are probably controlled by the position of curved
sheets that dilated during erosion of crystalline rocks.
Hydration and Swelling Expansion also occurs when minerals are formed or
when they are altered by the addition of water to their structure. Although
the process begins as a chemical process called hydration, its physical side is
particularly obvious when clay minerals containing layers of OH or H ; are

formed. The creation of the layered structure expands th^miqerals and prop-
agates stress outward from the clay particle.^Clays such as bent onite (Na-
montmorillonite), which do noUiajie-A-frxlÔHoT-rr^ struc-
ture, have the capaci^te-Tfbsorb water into theTTiineral during periods of
wetting. Th e swelling produced by wetting exerts the same outward stress
_
as
during clay formation. Most clays show the trait to some exteîu~BuTTrIe per-
centage of expansion depends on the mineral type plus a myriad of other fac-
tors (table 4.1). Montmorillonite clays, for example, drastically lose their
Table 4.1 Expansion of common clay minerals by hydration.
Clay Mineral % Expansion

Ca-Montmorillonite
Forest. Miss. 145
Wilson Dam, Colo
Cr. 95
Davis Dam, Ariz 45-85
Na-Montmorillonite
Osage, Wyo 1400-1600
lllite
Fithian, III 115-120

Morris, III 60
Tazewell, Va 15
Kaolinite
Macon, Ga 60
Langley, N.C. 20
Mesa Alta. N.M 5
Adapted from Mielenz and King 1955 with permission of the California Division of Mines and Geology
swelling capacity when sodium is replaced by some other cation (Mielenz and
King 1955). Upon drying, the expanded clays lose part or all of the absorbed
water, initiating an alternating swelling and shrinking sequence associated with
episodes of wetting and drying. In contrast, the well-ordered hydrated clays
have a stable structure, and destruction of the OH or water layer occurs only
when the mineral is heated to at least 300°C. The disintegrating effect of
these clays, therefore, occurs during their formation but, in contrast to the
swelling clays, is exerted continuously until relieved. The use of hydrated clays
as agents of disintegration is a one-shot affair, however, because once their
expansive stress is released in a disintegrating event there is no way to rein-
state the internal stress.
The effect of jniner_al_e xj>ansion has been clearly demonstrated in the
physical breakdown of granites jn arid or se miarid regions (Wahrhaftig 1965;

l
Eggler et al. 1969; Isherwood and Street 1976V Tn_these_s ettings the major .
_rjr oduct of granite disintegration js a coa rse angular mass of rock-and mineral
,
fra gments called grus (fig. 4.3), in which feldspars are often unaffected by
decomposition. In the Laramie Range of Colorado and Wyoming, the se-
quence of grus development started during the Precambrian with formation
of hematite by high-temperature oxidation along cleavage planes in the biotite
(Eggler et al. 1969). Although this expanded the biotite in the direction of
the c-axis, the stress was not sufficient to cause disintegration. It did, however,
weaken the geomorphic attack. Subsequent

biotite's ability to resist further
near-surface weathering produced clays from the biotite with as much as 40

percent increase in volume, and the stress generated by this expansion shat-
tered the granite into grus.
Chapter 4 114
In other situations, the granite bedroctinay be weakened in different ways

prior to the grus development hor example, Folk~and Patton (1982) show that
,
the first stage of grus formation in granites of central Texas is the development
o Lmicrosheet joints that parallel the weathering surfac e.. Because these joints
cut indiscriminately across mineral grains, they cannot be the result of biotite
expansion. Instead, they precede and ultimately enhance the grus develop-
ment.
It seems certain that hydration of salts within pores of building stones
also
and concrete develops sufficient stress to cause extensive spalling (Winkler
and Wilhelm 1970). According to E. Winkler (1965), a similar process almost
destroyed Cleopatra's Needle, an obelisk that was brought from Egypt to New
York City in 1880. Salts trapped in spaces within the red-granite monument
did not hydrate until they were placed in the humid climate of the eastern
United States, but since then significant spalling has taken place.
Figure 4.3.
Formation of grus by
S Qlt M/^QthPfinp ig increasingly suggested as a significant component in
disintegration of granitic boulders the physical breakdown of rocks, especially as an explanation of minor weath-
Upper part of photo shows grus ering features such as tafoni. 7a/afliareholes or depressionSj. usually less than
matrix developed when boulders
break apart (Coin is silver dollar)
several meters in width and depth, t hat commonly form on the un derside of
£Ockjaasses_or on steep rock faces. Thev often develop on granitic rocks s et
i n arid climate s. Salt w eathering by crys tallization of salt minerals or hydra-

tion expansion^ of salts has been suggested as the geneti c cause of taTo ni de-
velopment (Evans 1969; Winkler 1975; Bradley et al. 1978). The precise origin
of tafoni, however, remains a mystery (Evans 1969; Selby 1982) because the
features form in a variety of climates other than arid (Calkin and Cailleux
1962; Martini 1978; Watts 1979) and on many rock types other than granite.
In most humid regions, the process of mineral expansion manifests itself
in different end products, jjaefa-afs-peejed off to produce curv e d surfa ces; the
proces s on a larpe sc.a| e is calleHpYfnliatinnanri on a smaller scale spheroitiat—
Figure 4.4.
we aihrnagr-Even though the resulting large domelike masseToi I'ouiidul boul- Northeast side of Half Dome
ders (shown in figs. 4.4 and 4.5) are probably in part a function of pressure taken from the subsidiary dome at
the northeast end of the rock
release, seems certain that water and mineral alteration are intimately in-
it
mass, revealing exfoliation on a
volved (Gentilli 968). Spheroidal boulders are formed because edges and cor-
1
gigantic scale. In the foreground
ners of lithologic blocks are weathered more rapidly than flat surfaces, a is an old shell disintegrating into
phenomenon especially apparent where the parent rock has been fractured undecomposed granite sand
Yosemite National Park, Mariposa
County, Cal
mm
Chapter 4 116
into a blocky framework by perpendicular joint sets. The relatively fresh sphe-
roidal cores are usually surrounded by a zone of disintegrated flakes and spalls
that is enriched in secondary clay minerals. Simpson ( 1 964), for example, found
that the clay matrix in weathered graywacke increased by 5-10 percent in
the spalled zone and also contained abundant vermiculite, an expandable clay
not present in the fresh rock. Evidence such as this seems to indicate that
outward expansion caused by the development of clay minerals peels off the
fresh rock layer by layer, working progressively inward from the surrounding
joint openings.
Growth in Voids
A second group of processes generate stress when some substance grows in
spaces within the rock. The pressure gradient from that in the processes
differs
explained above because it is the openings that are expanded, not the parent
minerals or rocks.
Microcracks in rocks can be produced by processes acting inside the earth
Figure 4.5.
(Simmons and Richter 1976; Whalley et al. 1982) and therefore may already
Gabbro boulder showing
spheroidal weathering Himalaya be present before rocks are exposed at the surface. Because these spaces are
Mine, San Diego County. Cal
•'«M
P "
**>'
r .
not expanded simultaneously or with equal magnitude or direction, the re-

sultant pressures differ locally and the entire system is burdened with a dif-
ferential stress field. Such a pressure distribution is conducive to fracturing

or granulation; the processes responsible for its development are probably the
dominant agents of disintegration.

Plants and organisms aid in th e disintegrati ng process es but th eir greatest ,
effect usually occurs after the parentrock haTalready been converted into soil.
Pla nt roots commonly grow in fragtnrpsô f the parent rock an d p hysically pry
THe solid material apart Nonetheless, co mpared with_other pjgcesse s,_rQojiet
.
growth is of minor consequence.

~~Vjlf mnrTppr vasive processes of physical weathering in volve for ces gen-
erated by crystalliz ation of ice ( frost a ction ) or other minerals in rock spaces .
frofr
In a perfectly closed system , wate r increases 9 percent in volume upon freezing
and__alm ost c ertai nly pro d uces hydrostatic pressures that exceed the tensi le
streng th of all common rocks. Frost action is most e ffective when the rock is
saturated pri or to the freezing event. I n fact, simple alternations of wetting

and drying will sometimes fracture rocks, but the process is accelerated in
combination with freezing (Mugridge and Young 1983).
If more than 20 percent of the available pore space is empty, the expansion
pressure upon freezing may be less than rock strength, and shattering will not
occur (Cooke and Doornkamp 1974). Some evidence exists to suggest that the
intensity of frost action is related to the structure of the pores rather than
simply the percentage of pore space. In a system containing a variety of pore
sizes, ice crystals will preferentially grow in large pores rather than smaller
ones (Everett 1961).
Minerals can also growjn roc k s pa ces, w ith results similar to those of frost
action, as figure 4.6 shows. Most commonly the process functions when per-
colating fluids evaporate within the pores, giving rise to supersaturated con-
ditions and eventual precipitation of minerals. The pressures exerted in
Figure 4.6.
Pebble fractured by growth of
calcite along planes of weakness
Near Roberts, Mont
Chapter 4 118
tionaxe probably greater t hantho se produced_by ice, but their ab-

solute values depend on the concentration of the ionic constituents in the so-
lution. The most common precipitates are sulfates, cajjonajgs^ând chlorides
of^v^rûxLobUe^ catio ns (Ca, N a, Mg, K L_and the process is therefore~more
prone to operate in arid and semiarid regions where the ions are rendered
immobile by insufficient leaching.
The Significance of Water

Even a short review of physical weathering makes clear the importance of
water in the disintegrating processes, jjydrati on^fr ost action, cry stalgrowth.
a nd swelling all require water as a basi c component of the system. The amount
of water need not be great. Many believeTToTexampie, that eveirthin films of
condensing dew in desert regions may be infinitely more destructive than in-
solation (Twidale 1968). Therefore, it seems fair to expect a direct relation-

ship between climate and the prevalence of disintegration.
Peltier (1950) utilized mean annual temperature and precipitation to pre-
and chemical weathering; figure 4.7 shows
dict relative intensities of physical
these data. Physical weathering should be dominant where precipitation is
readily available and the mean annual temperatures are near or below freezing.
Presumably this analysis equates with the importance of frost action as a me-
chanical tool and with the fact that frost action is preeminent in those areas
having the most freeze-thaw cycles during the year. The frequency of freeze-
thaw events has been detailed for the United States by R. Russell (1943) and
L. Williams (1964).
Where unusual local problems exist, the regional climatic characteristics
may have little significance. In those cases it may be extremely important to
understand in detail the climate-lithologic-weathering system, and a more so-
phisticated approach than those just reviewed will be necessary. An excellent
example of this point was provided by Weinert (1961, 1965). In the eastern
part of South Africa, the parent Karoo dolerite has been altered into a mature
soil that is unsatisfactory for maintaining road foundations. In the western
part of the region, this mature soil is not present. Instead, hydration of the
micas has apparently disintegrated the dolerite into a grus that has consid-
erable internal strength and is quite sound from an engineering viewpoint.
Weinert found that the boundary between the sound and unsound surface ma-
terials (as defined by engineering properties discussed in the next section) could
be mapped by the distribution of evaporation and precipitation in the area.
Physical Properties Of The resistance of unconsolidated debris to the forces of erosion is dependent
Unconsolidated Debris on the physical properties of the material. In a real sense these properties de-
termine whether slopes developed on any substance will be stable or, if they
fail, the manner and rate of the resulting sediment movement. In addition, the
physical properties help determine the shape of the slope profile when and if
it attains an equilibrium condition. Clearly, then, the slope material itself di-
rectly influences the resulting process and landform. It is rather disconcerting
to find that most geologists have only a vague knowledge of the basic physical
Chemical weathering Frost action
10 10
Weak Strong .•
r- 20
i
- 20
CD CD
| 30 -1 30
ID
o. 40 Q- 40
E Mod< ;rate
E Weak
CD <D
i 50 CO 50
/
« 60 60
c
Absent orinsi gnificant
70 70
Str ong
80 80
80 70 60 50 40 30 20 10 80 70 60 50 40 30 20 10
Mean annual rainfall (inches) Mean annual rainfall (inches)
Weathering regions
Moderate mechani cal / k
10
weather ng
Strong mechanical
/
rr
20
weathering
.^Slight/
/
Moderate
CD
mechanical — •''mechanical
5 30 eathe 'ing
00
weother ing
ID f Mo< jerate chem icaM
Q. 40
E w< ;athenngwi th
S frost action
w 50 J
C
™ 60
Moderate
// Verys ight
CO
CD
Strc >ng chemical
^ 70
chemical weath enng / /eathe snng
VI
weathering
80
80 70 60 50 40 30 20 10
Mean annual rainfall (inches)
Figure 4.7.
Relative intensities of physical
and chemical weathering under
different temperature-precipitation
conditions.
Chapter 4 120
(A) (B)
Figure 4.8.
Analyses of slope stability properties of soil, which have been identified through years of study by en-
(A) Forces acting on a particle
gineers, even though these characteristics directly control the mechanics of
resting on a slope surface
(B) Stresses acting on a planar many geologic hazards.
surface covered by Before we examine these properties, however, it may be helpful to look
unconsolidated material
again at the concept of driving force, which was briefly discussed in chap-
ter 1. Part (A) of figure 4.8 depicts a boulder resting on a sloping surface.
The force of gravity acts vertically on the particle, and the force magnitude
stems from the weight of the particle, mg (mass times the acceleration of
gravity). Actually the weight may be resolved
two components, one acting
into
perpendicular to the sloping surface and one acting parallel to it. The com-
ponent acting parallel to the surface tends to promote downslope movement

and is measured as W
sin 0, where W
is the weight in pounds or kilograms.
The perpendicular component tends to keep the boulder in place by pushing

it into the surface and thereby resisting the downslope motion: its magnitude
is determined as Wcos 6. Clearly, downslope movement of the particle is en-
hanced on steeper slopes because the sine value increases and the cosine value
decreases as the angle 6 is increased.
In the analysis of slope processes, however, engineers and scientists are
usually concerned with the force acting on some plane existing below the ground
surface along which movement of a block of overlying material takes place.
See (B) of figure 4.8. In this case the exerted force derives primarily from the
weight of the debris overlying the plane. Because the total weight of this ma-
terial cannot be determined like that of a single, discrete boulder, it is cal-
culated indirectly by multiplying the unit specific weight (7) of the material
fl2l7 Physical Weathering, Mass Movement, and Slopes
(lb/ft 3 or kg/m 3 ) times the vertical distance (//) from the plane to the ground
surface. The resolved components are now yh sin 9 cos 9 in the downslope
2
direction and yh cos 9 perpendicular to the plane. The reason for the change
in equations is that the block of soil, represented by the parallelogram ABCD,
is equal to the rectangle BFEC, and the angles FBA and DCE are equal to 9.
Therefore,
cos 9 = -r- and x = cos 9h.

h
Since W= yx, substituting from above gives us W= yh cos 9. Thus, the

pressure acting perpendicular to the plane is
a = W cos 9 = yh cos 9 cos 6 = yh cos 2
and the shear is
r = W sin 9 = yh cos 9 sin 9.
Notice from the above that in the case of the buried plane, we are no
longer talking about force because the value yh is given in units of stress,
which by definition is the force acting on a specific area. This follows because
3
7 is in lb/ft
h is in ft
yh = lb/ft 3 X ft = lb/ft 2 .
When used in this sense stress and pressure are synonymous. The perpendic-
ular component of the total stress is called normal stress and is usually in-
dicated by the symbol a. The downslope component is called shear stress and Interlocking friction
is denoted in mechanical analyses by the symbol r.
Shear Strength
The properties of matter that resist the stresses generated by gravitational
force are collectively known as the shear strength. The detailed analysis of
internal strength is an extremely complex procedure, well beyond the scope
of this book. For our purposes, we can safely say that shear strength of any
material derives from three components: (1) its overall frictional character-
istic, usually expressed as the angle of internal friction; (2) the effective normal
and (3) cohesion. These
stress; factors determine shear strength by the Cou-
lomb equation, <r= normal stress
4> = angle of internal friction
S = c + cr'tan </>,
Figure 4.9.
where S is shear strength (in units of stress), c is cohesion, a' is effective normal Two types of internal friction
stress, and is the angle of internal friction. (A) Plane friction, in which
resistance occurs along a well-
defined plane that may cut
Internal Friction Internal friction is composed of two separate types: plane through individual grains. Angle of
friction, produced when one grain slides past another along a well-defined plane friction is approximated by
angle at which sliding begins
planar surface, and interlocking friction, which originates when particles are
(B) Interlocking friction, in which
required to move upward and over one another. These are shown graphically resistance occurs around particle
in figure 4.9. The angle of plane friction is approximated by the angle at which boundaries
Chapter 4 122
a block will begin to slide over another along the plane separating them. Once
sliding begins, the frictional angle actually decreases slightly, requiring that
a distinction be made between a static and a dynamic angle. In addition, the
plane friction angle varies considerably with smoothness of the plane surface,
moisture at the contact, mineralogy, and other factors.
Interlocking friction is usually greater than plane friction because extra
energy must be used to move interlocked grains in an upward direction. The
angle of interlocking friction also varies with mineralogy and moisture and, it
would seem, must be affected by the density of packing within the mass. In
loose particulate matter of any size, the angle of repose should approximate
the angle of internal friction, but Carson and Kirkby (1972) point out that a
platform holding a cone of debris at its angle of repose can be tilted up to 10°
before the slope fails. This suggests that the angle of repose may be somewhat
less than the maximum angle at which a slope can stand.
Effective Normal Stress The importance of normal stress is its capacity to

hold material together, thereby increasing the internal resistance to shear. In
theory, normal stress acting perpendicular to a shear surface (fig. 4.9) is ab-
sorbed by the underlying slab at the point of contact between grains. In reality,
some of the shear surface is occupied by openings filled with air or water. Since
pore pressure exists in these interstitial spaces, it tends to support part of the
normal stress. The total normal stress (a) therefore includes two elements,
effective normal stress (r/) and pore pressure (n), such that
a = a' + n.
When considering internal friction, effective normal stress (stress exerted at

the solid-to-solid contacts across the shear surface) is the critical parameter
rather than total normal stress.
Surface The value of pore pressure (yu) has a direct bearing on the effective normal
stress because it can add to or detract from the total stress value. For example,
Unsaturated zone figure 4.10 shows a distinct water table in unconsolidated material. At some
level x below the water table, a hydrostatic pressure is exerted that is equal
Water Table
to the specific weight of the water (y H ) times the vertical distance (h) between
the water table and the level of x. Although the pore pressure beneath the
water table is acting in all directions, Terzaghi ( 1 936) points out that a portion
Buoyant of the pressure will be exerted in a direction opposite to the normal stress and
force = yw •
therefore will provide some relief from the overburden weight. In contrast, in
the unsaturated zone above the water table, some of the water will be pre-
vented from moving downward because it is attached to particles by capil-
laritySimply stated, this attached moisture increases the weight of the soil.
Relating water content to the effective normal stress, three possible situations
Figure 4.10. can be envisioned.

Buoyant stress produced by
hydrostatic pressure beneath a
1. In a completely dry soil, the pore pressure is atmospheric and n is
water table zero. Therefore, the effective normal stress and the total normal
stress are the same since
a = a 0.
Below the water table, pore pressure is positive (> atmospheric

pressure), causing the effective normal stress to be lower because
<j = a
3. Above the water table, n is negative and the effective normal stress
is higher:
a' = o- - (— n).
Because the effective normal stress directly influences internal friction, it is
clear that dry or partially saturated soils, especially those with a high clay
content, should have greater shear strength and stand at higher slopes than
equivalent materials that are thoroughly saturated.
Cohesion Figure 4.1 1 demonstrates the relationship between shear strength

and effective normal stress. The graph indicates that as the effective normal
stress increases, the values of shear strength also rise. The relationship be-
tween the two variables defines a straight line that passes through the origin
of both axes. The angle between the line and the abscissa represents the angle
of internal friction. The material represented in figure 4.1 1 A has no discern-
ible strength when the effective normal stress decreases to zero, a condition
that is common in coarse, unconsolidated detritus. Solid rocks, however, pos-
sess shear strength even when a' is removed (fig 4.1 IB) because the constit-
uent particles are bonded or cemented together. The strength revealed here is Stress (a)
cohesion, a factor that is unaffected by normal stress. (A) Unconsolidated material
Clay-rich soils also have some cohesion, presumably because adsorption
of ions and water by clay minerals creates a binding structure among the par-
ticles. The cohesive strength depends on the attractive force between the par-
ticles and the lubricating action of the interstitial liquid. As Grim (1962) points
out, the molecules of the inner layers of water adsorbed to the surfaces of clays
are oriented by the electrical charge of the minerals; because of this, fluidity
and lubrication are not possible when moisture content is low, even though
water is present. For clays to become plastic and exert a lubricating action,
the adsorption of water layers must continue until the outermost ones can be
held but are no longer fixed in a rigid, oriented position. Fortunately, simple
tests can indicate how much water a soil can absorb before it begins to behave (B) Rock material
more water is added, when the substance will
like a plastic substance; or, if
lose all its cohesion and become a muddy fluid.

Figure 4.11.
If water is gradually added to a dry, pulverized soil, the voids fill and the
Relationship between shear
mixture becomes increasingly more plastic in its behavior. As more water is strength and effective normal
added, however, the cohesion decreases, and when all the pores are filled, any stress. (A) Unconsolidated
material has no shear strength
further input of water results in complete destruction of the internal fabric
when effective normal stress is
and the production of a fluid. Atterberg (1911) suggested two simple tests to zero. (B) Rock material has shear
indicate the transition from the solid to the plastic state and from the plastic strength (c) from cohesion even
where no effective normal stress
and the liquid limit.
to the liquid state; they are, respectively, the plastic limit
is present.
The two limits are expressed as moisture contents determined by the weight
of contained water divided by the weight of the dry soil. The range of water
content between the two limits is the plasticity index.
— — 1
Chapter 4 124
Table 4.2 Average Atterberg limits in clays saturated with common cations
3
Montmorillonite lllite Kaolmite Halloysite"
Plastic Liquid Plastic Liquid Plastic Liquid Plastic Liquid

Limit Limit Limit Limit Limit Limit Limit Limit
Na* 91 442 36 65 27 41 42 46
te 63 173 41 75 33 52 45 48
+
Mg^ 59 164 39 84 29 50 54 60
++
Ca 68 155 39 86 31 54 48 60
Adapted from Gnm 1962 Used with permission of McGraw Hilt Book Compart/
Dala from White 1955
a4 samples
"3 samples
c 2 samples
1 sample 2H2O sample 4H2O

tf
1
Profile of test results — The values of Atterberg limits are affected by a number of auxiliary fac-
Green Creek tors. First, they are related to the types of clay minerals included in the soil,
280
although the precise limits for any particular clay species vary appreciably
270 "S
(see table 4.2). In general, plastic and liquid limits are higher for montmo-
^s^
rillonite clays than for illites or kaolinites, mainly because the montmorillon-
260 I «== — ites are able to disperse into very small particles with an enormous total water-
250
— V
'N absorbing area. The type of exchangeable cation in the montmorillonite, how-
"5 x
ever, may engender wide variations in the limit values, especially in the case
y of Na + and Li
+
Plasticity indices vary in a similar manner. Second, limit values
^240 .
c tend to increase when the particle size is smaller. The activity, defined as the
o
'« 230
,
M ratio of the plasticity index to the abundance of clays (Skempton 1953). shows
> ,
< i
1
\ that plasticity increases proportionately with the percentage increase of clay-
DJ 1
y
220 1
/ sized particles. As the activity increases, soils tend to have a lower resistance
1
1
'
/ to shear, and so the parameter has some engineering significance. Finally,
210
drying of a soil reduces its plasticity because shrinkage during the dehydration
- < process brings particles closer together. The attraction between particles is so
200
W W
P L
wN strengthened that penetration by water is difficult. Conversely, repeated wet-
190 I t I 1 1 1 1
ting combined with only partial drying may increase plasticity (Grim 1962).
20 40 60 80
Water content (%) Although the tests for Atterberg limits are rather unsophisticated, the
results are consistent when determined by more than one analyst, and they
Figure 4.12. are geologically significant (Casagrande 1948; Seed et al. 1964). More im-
Variation with depth of soil test portantly, the moisture content in fine-grained material is a guide to its in-
results in the Leda clay,
St Lawrence River valley Natural
ternal strength. Some soils, called sensitive soils, exist with natural water
water content commonly is contents above their liquid limits (fig. 4. 12), an apparent inconsistency with
greater than the liquid limit in the limit concept. The fact is, however, that some soils develop an open "ho-
these sensitive soils p
is the W neycomb" structure (fig. 4.13) that is capable of holding water in excess of
plastic limit; WL is the liquid limit
WN is measured water content its Although the structure is unstable, in an undisturbed soil the
liquid limit.
material will be solid and possess some strength. The disruption of the internal
structure by erosion, earthquake shocks, or other phenomena will cause the
excess water to be released, and the solid material will become a fluid. A simple
test for the liquid limit can reveal the presence of sensitivity and the potential
for dangerous mass wasting.

1 25 Physical Weathering, Mass Movement, and Slopes
Figure 4.13.
Open soil structures in sensitive
clays Structures in (A) and (B)
allow water content to exceed
liquid limit. Disturbance of this
structure causes temporary
Undisturbed Undisturbed fluidity until substance is
salt water deposit fresh water deposit
remolded (C). (From Lambe, 1953,
(A) (B)
pp 315-18. Used with permission
by the American Society of Civil
Engineers )
Remolded
(C)
Measurement of Strength
Although some field tests have been developed, most detailed analyses of shear
strength in soil are done in a laboratory by means of the direct shear test, the
triaxial compression test, or the uniaxial (unconfined) compression test. (The
details of these tests are provided in most textbooks of soil mechanics.)
Measurement of shear strength is a complicated task, made additionally
difficult because some techniques do not yield certain critical data. In the di-
rect shear test, for example, pore pressure cannot be determined. It should
also be kept in mind measure stress at
that laboratory tests are designed to
Strain
the moment of failure, which is analogous to shear strength. Natural mate-
rials, however, especially soils, begin to deform well below the stress level that
Figure 4.14.
causes rupture (fig. 4.14) and may continue to deform at a rate that increases Relationship between stress and
steadily with additional stress. Ultimately the material ruptures at a critical strain Permanent deformation
stress called the breaking strength. begins at Y (yield stress) but
rupture of the substance does not
Determination of strength in rocks presents an added complication be- occur until the stress value
cause tests on small samples may not be reliable indicators of the overall reaches B (breaking strength)
strength of a large rock mass. This occurs because structures within the mass Materialbehaves like a plastic
substance between Y and B.
such as joints and other fractures may make the total mass relatively nonresis-
However, analyses of rock samples taken in zones between the
tant to erosion.
fractures may indicate that the rock is quite strong. Obviously methods to
estimate rock strength in the field are needed. One method devised by Selby
(1980, 1982) is presented here as table 4.3. The Selby classification allows
any rock mass to be placed into one of five categories representing overall rock
mass strength. Placement in a specific group is based on numerical ratings (r)
given for various strength parameters. All parameters are not considered to
be of equal significance in the determination of strength, so each is assigned
a percentage value representing its importance. The sum of the weighted values
is an estimate of the rock mass strength.
The appeal of an approach such as Selby's is that a rapid estimate of rock
mass strength can be made with normal geological field equipment except for
the Schmidt hammer, which is a portable engineering tool designed to mea-
sure hardness (intact strength) of materials. In addition to stability recom-
mendations, information gained from strength analyses may be important clues
to understanding slope profiles and evolution.
Chapter 4 1 26
Table 4.3 Field classification of rock strength
1 2 3 4 5
Parameter Very Strong Strong Moderate Weak Very Weak

Intact rock strength 100-60 60-50 50-40 40-35 35-10
(N-type Schmidt
hammer "R") r : 20 r : 18 r : 14 r : 10 r : 5
Weathering unweathered slightly moderately highly completely

weathered weathered weathered weathered
r : 10 r 9: r : 7 r : 5 r : 3
Spacing of >3 m 3-1 m 1-0 3 m 300-50 mm <50 mm

joints r : 30 r : 28 r : 21 r : 15 r : 8
Joint Very favorable Favorable Fair Unfavorable Very unfavorable

orientations Steep dips into Moderate Horizontal dips. Moderate dips Steep dips out
slope, cross dips into or nearly out of slope of slope
joints interlock slope vertical (hard
rocks only)
r : 20 r : 18 r 14
: r : 9 r : 5
Width of <0 1 mm 1-1 mm 1-5 mm 5-20 mm >20 mm

joints r : 7 r : 6 r : 5 r : 4 r : 2
Continuity of none few continuous continuous, continuous.

joints continuous continuous no infill thin infill thick infill
r : 7 r : 6 r 5: r : 4 r : 1
Outflow of none trace slight moderate great

groundwater <25C/mm/10 m2 25-1258/mir /10 m2 >125C/mm/10m 2
r : 6 r : 5 r : 4 r : 3 r : 1
Total rating 100-91 90-71 70-51 50-26 <26
Used with permission ot Zeitschnft fur Geomorphologie Published by Gebruder Borntraeger Vertagsbuchhandlung Berlin Stuttgart Johannes strafe 3A 7000 S -
Mass Movements of Mass movements are of three primary types: slides, flows, and heaves, each
Slope Material having distinctive characteristics (Carson and Kirkby 1972). In slides, co-
hesive blocks of material move on a well-defined surface of sliding, and no
internal shearing takes place concurrently within the sliding block. In con-
trast, flows move entirely by differential shearing within the transported mass,
and no clear plane can be defined at the base of the moving debris. The velocity
in flows tends to decrease from the surface downward. In heave, the disrupting
forces act perpendicular to the ground surface by expansion of the material.

This movement does not in itself provide a lateral component of transport, but
it facilitates slow, downslope movement by gravity and serves as an important
forerunner of more rapid mass movements such as rockfalls.

Although each of these movements could theoretically function alone, it
seems certain that all are involved to some extent in most natural slope fail-
ures. Many processes can in fact only be explained by some combination of
the primary types of movement. As figure 4.15 shows, mass movements can
be thought of as multifarious events. The location of a particular process near
Flow
Figure 4.15.
Classification of mass movement
processes.
Slide Heave
a corner of the triangle in the figure indicates the dominance of the primary
movement occupying is to the corner, the more
that corner; the closer a term
dominant that type of movement. Superimposed on the diagram are lines that
represent the relative transport velocities of each process and the usual water
content within the debris being moved.
Slope Stability
The various mass movements are alike in that all begin when the shear stress
tending to displace material exceeds the resisting strength. As long as slope
material maintains its internal resistance at a level greater than the driving
impetus, no slope failure will occur. Stability, therefore, represents some bal-
ance between driving forces (shear stress) and resisting forces (shear strength),
and can be expressed as a safety ratio:
resisting force (shear strength)

F=
driving force (shear stress)
F values greater than 1 connote slope stability, but as the ratio approaches
unity a critical condition evolves and failure is imminent. Clearly, any factor
that lowers the safety ratio (see table 4.4) can trigger mass movement, and
this tendency can be produced by increasing the driving force, lowering the
resistance, or both. Theoretically, failure occurs when F= 1; this value is an
excellent example of a geomorphic threshold. In reality, however, because of
Chapter 4 128
Table 4.4 Factors that influence stress and resistance in slope materials.
Factors That Increase Shear Stress

Removal of lateral support
Erosion (rivers, ice, wave)
Human road cuts, etc)
activity (quarries,
Addition of mass
Natural (ram, talus, etc.)
Human (fills, ore stockpiles, buildings, etc )
Earthquakes
Regional tilting
Removal of underlying support
Natural (undercutting, solution, weathering, etc.)
Human activity (mining)
Lateral pressure
Natural (swelling, expansion by freezing, water addition)
Factors That Decrease Shear Strength

Weathering and other physicochemical reactions
Disintegration (lowers cohesion)
Hydration (lowers cohesion)
Base exchange
Solution
Drying
Pore water
Buoyancy
Capillary tension
Structural changes
Remolding
Fracturing
After Varnes 1958, with permission of the Transportation Research Board
imprecise measuring techniques, some failures have occurred when the F value
was and some slopes have been maintained temporarily with
slightly positive,
small negative values. Once failure does occur, the type of movement depends
on precisely how the forces interact with one another.
The stability of slope material above a suspected plane of failure can be
estimated if the components of stress and strength are known.
On shallow planar surfaces like that shown in figure 4.8(B), the driv-
ing stresses are numerically equal to yh sin 6 cos 6. Resistance is determined
as shear strength, which is derived numerically as ^ = c + (yh cos 2 — n)
tan </>. Therefore,
_ _ c + (yh cos 2 fl — ju) tan 4>
yh sin 6 cos 6
In most analyses the vertical height of the water table above the slide
plane is expressed as a fraction of the soil thickness above the plane (m), where
m= 1 .0 if the water table is at the surface, and m= if it is at or below the
sliding plane. Thus, the pore pressure can be noted as
M = y K mh cos
2
and
_ c + (7 — my H .) h cos 2 8 tan
yh sin cos
The following hypothetical example will show how to determine whether

the slope is stable or close to failure. If laboratory tests tell us that = 10°,
c = 45 lb/ft 2 t ,
= 165 lb/ft3 6 ,
= 8°, h = 12 ft, m = 0.6, and yw = 62 .4
lb/ft 3 then
= 45 + (165 - 0.6 X 62.4) 12 X .98 X .18

F
165 X 12 X .14 X .99
_ 45 + (165 - 37.44) 2.12

274.43
_ 45 + 127.56 X 2.12
274.43
„ 45 + 270.42 315.42
1.15
274.43 274.43
The slope in our hypothetical case is stable, but changes in the controlling
factors could easily lead to failure. For example, a rise in the water table to
m= 0.9 would decrease the F value to 1.0 and presumably cause slope failure.
The analysis of slope stability for curved sliding surfaces such as those
found in rotational slips is more complicated because it involves balancing
moments (force X distance) rather than simply the forces acting on the sur-
face. Rotational failures are usually analyzed by dividing the material above
the slip plane into vertical slices and treating each slice by a modified form of
the planar analysis discussed above. The values of all slices are then summed
to determine the total resistance and shear stress and the estimate of the slope
stability. (For an excellent discussion of slope stability, see Selby 1982.)
Heave, Creep, and Falls Heave is instrumental in the process of creep, the
almost imperceptibly slow movement of material in response to gravity. Sea-
sonal creep or soil creep the downslope movement of regolith that is aided
is
periodically by the heave mechanism. No continuous external stress is placed

on the mass; it moves under gravity when its cohesion and frictional resistance
are spasmodically lowered. The process functions in the upper several feet of
the soil, and its effect decreases rapidly with depth. The phenomenon of soil
creep was first recognized in the latter part of the nineteenth century, and its
ubiquity was gradually accepted as its effects were observed in the field. His-
torically, evidence that suggests a soil creep influence on slope materials has
included downslope curvature of bedding (fig. 4.16), stone lines, downslope
growth of trees or and accumulations of soil upslope from

tilting of structures,
a fixed obstruction (see Young 1972). Recently, however, observations and

measurements of seasonal creep have become more sophisticated. Precise sur-
veying methods and trenches such as Young pits (Young 1960) are fairly stan-
dard techniques in current studies of the process. A complete review of the
common methods employed to measure soil creep can be found in Selby (1966)
and Revue de Geomorphologie Dynamique (1967), and new techniques are
continually being developed (Finlayson 1981).
Brf
Figure 4.16.
Creep in vertical Romney shale Although burrowing animals and vegetation may cause random distur-
Western Maryland Railroad cut
bances in soils, their effects are minor compared with the heave produced by
one mile west of Great Cacapon
Washington County, Md swelling or freezing and thawing. In the heave mechanism, expansion disturbs
soil particles perpendicular to the ground surfaces; when the soil contracts,
the vertical attraction of gravity acts on the particles. The expansion-con-

traction cycle, therefore, adds a lateral component to particle motion in any
soil having an inclined surface. Because gravity is reasonably uniform over
the Earth's surface, the distance of transport in each heave event, and pre-
sumably the rate of creep in any climate, should vary with the slope angle and
should decrease with depth beneath the surface. Schumm (1967a) has dem-
onstrated a significant correlation between the rate of surficial rock creep and
the sine of the slope angle, but documentation of this relationship for fine soils
or below the surface is lacking. Actually, as figure 4. 17 shows, the contraction
event is never perfectly downward but usually moves in a direction about
midway between the normal and the vertical. As the lateral distance traveled
in each heave is less than would be theoretically predicted, a clear relationship
between slope angle and creep rate may be difficult to demonstrate. In fact,
considerable evidence suggests that even the direction of creep movement is

Figure 4.17.
Movement of near-surface
often random on a short-term basis, complicating the presumed relationship
material by heaving During even more (Fleming and Johnson 1975; Finlayson 1981). In addition, detailed
expansion (E), particle is measurements (Kirkby 1967) have shown the creep rate to decrease with depth
displaced perpendicular to the
surface During contraction (C).
(fig. 4. 1 8). Presumably this relates to the lower frequency of heaving at depth
particle settles in a vertically and to the greater difficulty of expansion with increasing overburden. In any
downward direction under case, below a depth of 20 cm movement ceases or becomes drastically smaller
influence of gravity Actual
movement is shown by line D (Young 1960; Kirkby 1967).
.
Young pit results

Figure 4.18.
Rate of creep as it relates to type
Surface
and depth of material. Rate in all
• soils decreases with depth.
.*
*
•
•
•
*
•
•
•
•
•
•
•
•
•
•
Q.
•
• » y
Q •
Median sine i
slope = 0.22 0.42 d 0.37 0.13

•
•
•
•
Bluff Peat
1 1
Rate of movement (mm/yr)
Based on 6-10 measurements

Based on 1 1 -20 measurements
Based on more than 20 measurements
Downslope creep rates are extremely variable because of differences in

slope angle, moisture content, and measuring technique (Caine 1981). Par-
ticle size also introduces variability in the rate because fine-grained clays tend
to swell more upon wetting, and frost heaving is greatest in silty material.
Nevertheless, available measurements indicate that downslope rates range from
0.1 to 15 mm/yr where the soil is vegetated. Rates may increase to 0.5m/yr
or more on uncovered talus slopes where frost action is prevalent (Gardner
1979; Selby 1982). Volumetric rates can also be calculated as the volume of
soil moved annually across a plane set perpendicular to the surface and par-
allel to the contour of the slope, for a unit horizontal distance along the plane.
In semiarid regions the rate seems to be somewhat higher. In arctic climates
a special kind of creep process called solifluction is extremely important in the
geomorphic scheme; it is discussed in chapter 1 1
A second type of creep, continuous creep (Terzaghi 1950), is fundamen-

tally different from seasonal creep in that (1) it is driven by gravity alone,
(2) it may affect consolidated rock, and (3) it can function at levels well below
the surface. Continuous creep is the strain response to stress that is generated
by the weight of overburden. It begins at the yield stress and continues even
though no additional stress is placed on the material. Continuous creep is es-
pecially pronounced where rocks or semiconsolidated materials with low yield
by stronger substances. For example, a weak clay
stress (fig. 4.14) are overlain
unit sandwiched between resistant strata is prone to deform by continuous
creep. Excavations of any kind through that rock sequence will reduce the
lateral confining pressure on the clay unit, and it will begin to flow. Creep of
this type is not important in terms of the volume of material it moves or the
Chapter 4 132
distance of transport, but it is very significant as a precursor of rapid, some-

times catastrophic, mass movements. In many cases landslides are immedi-
ately preceded by accelerated creep; persons who intervene in the natural
setting can trigger these events by not recognizing the potential for continuous
creep (Kiersch 1964).
The heave mechanism is also an essential element in some rapid mass
movements, especially falls. Falls in both rock and soils involve a single mass
that travels as a freely falling body with little or no interaction with other
solids (fig. 4. 19). Movement usually is through the air, although occasional
bouncing or rolling may be considered as part of the motion. Rock falls are
most common where the parent material is well jointed and a steep slope is
developed on the rock face. The fractures are enlarged progressively by heaving,
mainly in the form of freezing and thawing, until the gravitational force ex-
ceeds the internal resistance. Undercutting of the rock or soil face by erosive
agents acting at the base of the material accelerates the process. The removal
of the subjacent support tends to increase tension in the overhang and so helps
to create and expand incipient cracks.
Topples (fig. 4.19) are similar to falls except that forward movement of
a material block is produced by rotation around a fixed hinge located at the
base of the block. Although the process is not commonly cited, in some cases
toppling may be the most important factor in cliff retreat (Caine 1982).
Slides As rapid mass movements are grouped according

in all classification,
to the classifier's opinion as towhat aspects of the phenomena are most im-
portant. Because most rapid movements are not observed as they occur, their
fundamental properties of motion must be interpreted after the event. This,
combined with the fact that no clear-cut distinction can be made between the
primary modes of transport, has resulted in a number of viable classifications
of rapid mass movements (Sharpe 1938; Ward 1945; Varnes 1958; Hutch-
inson 1968). The classification prepared by Varnes (1958) is adopted here as
figure 4.20 because it relates well with our process orientation. The classifi-
cation is based primarily on the type of material being moved and the primary
type of movement. Additions to the classification (Varnes 1978) are shown in
table 4.5.
As defined earlier, slides are slope failures that are initiated by slippage
along a well-defined planar surface. The sliding mass is essentially unde-
formed; however, it may partially disintegrate during the sliding motion, giving
rise to flow movement in the latter phase of the event. The plane of sliding
may be shallow and approximately parallel to the ground surface as in the
Figure 4.19.
Types of mass movements
(A) Rock fall on Interstate 70 west
of Denver, May 5-6. 1973
(B) Rock topple leading to a rock
fall in Morgan County, Ky.
—
Chapter 4 134
BEDROCK SOILS
(including rock fragments, sheared bedrock, weathered zone, organic soils)
Rockfall Joint opened, e.g. by Soil fall

hydrostatic
pressure or Original support removed
CO frost wedging
_J e.g. sea or river erosion,
_l quarrying
<
Mixed sediments
EXTREMELY RAPID
Rotational slump Planar — block slide Slides
Rupture
surface
EXTREMELY
SLOW TO MODERATE
Rockslide
Scarp face Dip slope — water-bearing silt
control by joints control by bedding planes Firm clayey gravel
General lateral
movement of soft clays
VERY SLOW TO VERY RAPID
EXTREMELY RAPID
UNCONSOLIDATED MATERIALS
MAINLY LARGE NON-PLASTIC MIXED ROCKS AND SOIL MOSTLY PLASTIC
ROCK FRAGMENTS SORTED SAND OR SILT
Rock fragment flow Sand run Loess flow Slow earthflow
Loess
I
Shale-
Debris avalanche
Bedrock
Sand Weathered shale
EXTREMELY RAPID RAPID TO VERY RAPID EXTREMELY RAPID
APPROXIMATE RATE OF MOVEMENT
Rapid earthflow Regolith

Extremely
VERY RAPID 3m/second ra P ld
VERY RAPID TO
EXTREMELY RAPID Very rapid
3 minute
Rapid
1 5m day
Moderate
RAPID TO VERY RAPID r— 5m month
Sand— 1
Slow
1 5m/year
Very slow
3m 5 years
Extremely
VERY RAPID slow
Figure 4.20.
Classification of landslides
Table 4.5 Classification of mass movement types in different parent materials.
Type of Material
Type of Movement Engineering Soils

Bedrock
Predominantly coarse Predominantly fine
Falls Rock fall Debris fall Earth fall
Topples Rock topple Debris topple Earth topple
Rotational Few Rock slump Debris slump Earth slump

Units
Slides Rock block slide Debris block slide Earth block slide
Translational
Many Earth slide
Units Rock slide Debris slide
Lateral Spreads Rock spread Debris spread Earth spread
Flows Rock flow Debris flow Earth flow

(deep creep) (soil creep)
Complex Combination of two or more principal types of movement
From Varnes 1978. T R B Special Report 176 Landslides Used with permission of the Transportation Research Board National Academy ot Sciences
case of rockslides and debris slides, or it may penetrate to some depth as a

concave surface along which rotational slip (fig. 4.21) or slumping may occur.
Slides on shallow planar surfaces, called translational slides, are the most
common of sliding phenomena. Like all forms of mass movements, the initi-
ation of translational slides occurs when the material's resisting strength is
exceeded by the shear stress. As long as F is greater than 1 at the potential

surface of sliding, the slope will be stable. If an increase in the driving force
or a decrease in resistance brings the ratio to unity, sliding will ensue. Com-
monly the driving force is increased by an addition of mass to the sliding block,
but other factors can produce the same effect. Earthquakes, for example, gen-
erate a horizontal mass force that passes through the center of gravity of the
slope material and adds an extra driving element. Steepening of the slope also
adds driving force because shear stress increases as the slope angle (0) in-
creases. Thus, human activity or erosion that increases slope may trigger mass
movement. Actually this is a rather simplistic view since steepening also re-
duces the shear strength by complicated changes in cohesion, pore pressure,
and effective normal stress (Terzaghi 1950; Carson and Kirkby 1972).
The sliding phenomenon can also be produced by a variety of events that
reduce the internal resistance of the debris. From observation, sliding usually
occurs after prolonged or exceptionally heavy rainfall, indicating that the low-
ering of resistance is predominantly a function of water. In the past, the water
Chapter 4 136
Figure 4.21.
Crown
(A)Features of a rotational slide. Slope reversed
(B) Example of rotational slump In
which toe area has disintegrated
into an earthflow.
Right flank
Originalground
surface
Transverse
cracks
Transverse
ridges
Radial
Transverse
cracks
crack
Surface of rupture
(shear plane)
Longitudinal
fault zone
effect was interpreted to be lubrication along the sliding surface. Terzaghi

(1950), however, refutes this notion by pointing out that water applied to many
common minerals such as quartz is actually an antilubricant. Furthermore,
most soils in humid regions contain more than enough water to cause lubri-
cation at all times, yet they also fail after rainstorms. Clearly water affects
strength in other ways. You will recall that shear strength is a function of PSS
1
cohesion (c), effective normal stress (a ), and friction (0) such that
S = c + (a) tan </>.
The response of these factors (c, a' , <j>) to wetting is significantly more im-
portant in the initiation of slippage than is lubrication (Sidle and Swanston Figure 4.22.
Differences in rockslides caused
1982). For example, the rise of the water table or the piezometric surface, by slab failure and by rock
which accompanies all prolonged rainfalls, may be the most common culprit avalanche Slab failure (SF)
in sliding. As the water table rises, the pore pressure (fi) at any point within involves only the outer sheet of
rock, which splits along one
the saturated mass increases, ultimately resulting in a decrease of effective prominent fracture. A rock
normal stress (a) and the concomitant reduction of shear strength. avalanche (RA) occurs when
Rockslides are usually associated with major structural features within fracture system penetrates to the
potential surface of sliding (PSS).
the rock such as the stratigraphy of the rock sequence or joint patterns. Mas-
sive rock units normally have several prominent joint sets as well as a super-
imposed network of randomly spaced and oriented fractures. Prior to the
creation of the joints these rocks have great shear strength associated with the
high cohesion in lithified materials. Once joints begin to form, however, the
process becomes self-generating because shear stresses are concentrated pref-
erentially on the unfractured zones of solid rock. With time, more and more
of the original rock is consumed by jointing, and eventually the near surface
becomes a cohesionless mass of densely packed angular blocks (Terzaghi 1962).
At this stage, except for particle size, the rock mass resembles an aggregate
of dry sand in which the cohesion within the individual particles has little
bearing on slope stability. The shear strength of the total mass stems entirely
from internal friction because the cohesion across the joint openings is zero.
Rockslides can be divided into two types: rock avalanches, and slab failure
(Carson and Kirkby 1972). Both obey the same mechanics, and they differ
only in the amount of fracturing within the rock and the angle on the potential
surface of sliding (fig. 4.22). In slab failure, cracks develop where a rock mass
expands because the horizontal confining pressure is removed, allowing strain
to proceed outward in the direction of the pressure release. In that sense, the
process in rocks is similar to pressure-release sheeting. As lateral stress is re-
moved, a tensional zone develops in the upper part of the mass, and cracks
form within the zone. The tensional fractures penetrate to a depth that is con-
trolled by the strength of the material. Because rocks usually have high tensile
strength, fracturing does not penetrate the total depth of the tensional zone.
This is not always the case in unconsolidated substances.
Chapter 4 138
The stability of the outer slab depends on the depth of the fracture relative
to the height of the unconfined surface that is undergoing expansion. Equa-
tions have been derived to predict the maximum height attainable before slab
failure occurs (Terzaghi 1943), but predictive models are not always com-
pletely successful. For example, in one study (Lohnes and Handy 1968), ten-
sion cracks probably did not appear until slab failure was imminent. As a
result, the unsupported face was considerably higher than would have been
possible had the cracks developed at an earlier stage. In addition, Terzaghi
(1962) considered the worst possible case — the weakest rocks—and found that
most rock types could probably stand vertically up to heights of 1300 m. The
observation that few vertical cliffs stand at this height even in stronger rocks
indicates that fracturing drastically reduces the height that can be maintained
on an unsupported rock face.
Rock avalanches occur when the joint network becomes essentially con-
tinuous down to the potential surface of sliding. An avalanche differs from
slab failure in that it involves the entire mass above the sliding surface whereas
slab failure includes only the material outside the outermost continuous joint;
figure 4.22 shows both events. Both types of rockslide differ from rockfalls in
that the rocks fail only when fractures intersect the potential surface of sliding.
Falls, on the other hand, can occur above the potential slide plane. Both heaving
and sliding processes may be involved in some natural mass movements
(Schumm and Chorley 1964). In fact, the incidence of sliding is so compli-
cated that Terzaghi (1950) was able to list 19 possible causes.
Flows In true flows, the movement within the displaced mass resembles
closely that of a viscous fluid, in which the velocity is greatest at the surface
and decreases downward in the flowing mass. In many cases, flows are the
final event in a movement begun as a slide, and the distinction between the
two is usually unclear. For most types of flow, abundant water is a necessary
component, but dry flows called rock fragment flows by Varnes (1958) do
occur when rockslides or falls increase drastically in velocity and lose their
identity as a unitized mass. When rocks slide or fall down a steep slope, the
material disintegrates as it crashes into the relatively flat surface at the base
of the slope. From there the rocks travel as masses of broken debris (called
sturzstroms German), moving with enormous velocities over the gentler
in
slopes in the piedmont or valley bottom. For example, the wet. mud-soaked
sturzstrom at Mount Huascaran, Peru, sped at approximately 400 km an hour
over a distance of 14.5 km, wreaking destruction along its path and killing
some 80,000 persons (Eriksen et al. 1970; Browning 1973).
The exact mechanism required to move these large bouldery masses (usu-
ally greater than million m 3 ) for such long distances is the subject of some
1
controversy. Shreve (1966b, 1968) proposes that the material slides continu-
ously on a layer of compressed air that was trapped beneath the debris as it
came to the bottom of the steep slope. The air-lubrication hypothesis has been
challenged by Hsu (1975) who suggests that sturzstrom movement is pri-

marily a flow phenomenon, reviving a conclusion made earlierby Heim (1932).
In most cases the highest strata involved in the fall are contained in the rear
portion of the deposited debris. Shreve correctly indicates that this distribu-
tion negates viscous flow as the transportingmechanism, because in that pro-
cess the uppermost layers are transported at higher velocities and so would be
farther downstream in the blocky debris. The deposits, however, also have geo-
metric features similar to those formed by lava flows and glaciers, and care
must be used to distinguish between the different types of deposits (Porter and
Orombelli 1980). Hsu therefore feels that the debris moves by flow, but that
the mechanism from viscous transport in that individual particles are

differs
dispersed in a dust-laden cloud, and the kinetic energy driving the flow is
transferred from grain to grain as they collide and push one another forward.
The entire mass moves simultaneously until all the original energy is dissi-
pated by friction from the particle collisions.

Flow by this process would explain the distribution of the source rocks
within the deposits. It would also permit great distances of transport because
the frictional resistance decreases when grains are immersed in a buoyant in-
terstitial fluid that reduces the effective normal stress. Variations in flow dis-
tance and velocity in different events probably depend on the properties of the
interstitial substance.
In soils, the transition from debris slides to debris flows also requires an
increase in water content or another buoyant substance. In the intial phase,
the mass breaks into progressively smaller parts as it moves downslope (fig.
4.20) even though the velocity of the advance may be slow. If the mass is wet,
a debris avalanche may be generated as a long, narrow flow that extends well
beyond the foot of the slope. The erosive nature of the movement commonly
begins gulley formation on relatively undissected slopes, examples and de-
scriptions of which can be found in Hack and Goodlett (1960) and Rapp (1960).
Debris flows and slides usually result from heavy rains or the sudden melting
of frozensoils. Torrential rain is especially effective in producing debris flows
where vegetation has been stripped from a deep soil on moderate or steep slopes
(Heller 1981; Renwick et al. 1982; Wasson and Hall 1982). For example, the
annual southern California pattern of summer brush fires followed by torren-
tial winter rains produces excessive runoff that converts the unbound soil into
a high-density flow (Sharp and Noble 1953). These flows, following preex-
isting drainage, have tremendous erosive and transporting power, and may
cause damage many kilometers downvalley while simultaneously eroding the
sediment-yielding slopes in the headward reaches.
Earthflows and mudflows involve movement of fine-grained slope material
and range from slow to rapid. In slow earthflows, the original failure of the
slope is usually in the form of slump, often when the mass becomes saturated
with groundwater. As explained earlier, a rising water table and pore pressure
tend to lower shear resistance, and slippage results (Wells et al. 1980). If the
Chapter 4 140
slumped mass is relatively wet, it may slowly bulge forward at its front by
viscous flow and take the form of tongues, superimposed piles of rolled mud.
or bulbous toes (fig. 4.23). This movement may continue at a slow pace for
many years until some stability is finally reached. Rapid earthflows or mud-
flows also begin as slides or debris avalanches (Pomeroy 1 980; Cummans 1 98 1 );
in fact, most can be traced back to bowl-shaped scars indicative of sliding.
The primary slip, even if the motion is small, may remold the soil and cause
it to lose much of its original undisturbed strength as it releases the water in
excess of the liquid limit. The material changes instantaneously from a plastic
solid to a viscous liquid and flows downslope away from the slipped zone.
To summarize, the distinction between slides and flows is often rather neb-
ulous. Several generalizations can be proposed, however, to help put mass
movements in some reasonable perspective:
1. Most flows rise naturally as the final stage of a movement that

begins as a slide. For slope stability analyses, therefore, it is
probably more important to understand the mechanics of sliding

and the factors that might produce it.
2. The mobility or rate of mass movements depends to a large degree

on the amount of water or other buoyant substances in the
displaced material. This is particularly true in flow movements.
Morphology of Mass Movements

It is appropriate to ask how we can reconstruct the mode of mass transfer,
especially since subtle transitions from one mechanism to another are common,
and most interpretations of movement characteristics are made after the event
is over. Unless some concrete relationship exists between the surface config-
uration of the displaced material and the genetic process, we are facing an
insoluble dilemma. Fortunately some evidence has been presented to suggest
that the desired morphologic relationship is real. In a study of 66 landslips in
New Zealand, Crozier (1973) arranged the common types of mass movements
into five primary process groups: fluid flow (mudflows, debris flows, debris
avalanches), viscous flow (earthflow, bouldery earthflow), slide-flow (slump/
flow), planar slides (turf glide, debris slides, rockslides), and rotational slides
(earth and rock slumps). Each of the 66 landslips studied was described quan-
titatively by seven morphometric indices, listed in table 4.6 and illustrated in
figure 4.24. The relationship between each process group and the index values
was tested statistically to ascertain whether the correlation was significant
enough to warrant use of morphometry as a genetic determinant.
Crozier found that the classification index (D/L) was the best indicator
of the process group, reaffirming Skempton's (1953) assertion about the im-
portance of this parameter. As one would expect, the D/L value decreases
?Wfc
^s^.-zsmt <\*r* i
Figure 4.23.
Types of flows and slides
(A) Lobate earthflow (center of
photo) Near Roberts, Mont
(Photo by R R. Dutcher)
(B) Hebgen Lake earthquake.
Upstream view of Madison slide,
showing the west edge of the
slide debris, blocked highway,
and dry river bed below the slide
Madison County, Mont., August
1959- (Photo by J. B Hadley, U.S.
Forest Service photo, fig 54 in
U S Geol Survey Prof. Paper
435-K 1964) (C) A slump-earthflow
caused by reactivation of an old
stabilized slump block The
slump-earthflow follows the outline
of an old slump block apparently
subsidiary to the stabilized
landslide on the left side of the
active area Charles Mix County,
S.D.
*.
Chapter 4 142
Table 4.6 Morphometric indices used to determine process of mass movement
Index Description
Classification D/L — maximum depth of displaced mass prior to its displacement over
maximum length
Dilation Wx/Wc — width of convex part of displaced mass to concave indicates

lateral spreading
Flowage (Wx/Wc — 1) X Lm/Lc X 100 — Lm is length of displaced mass. Lc is

length of concave segment
Displacement Lr/Lc — Lr is length of the surface of rupture exposed in concave

segment Low value indicates instability
Viscous Flow Lf/Dc — Lf is length of bare surface on displaced material. Dc is the

depth of the concave segment
Tenuity Lm/Lc — Indicates how dispersed or cohesive the material is during

displacement
Fluidity Amount of flowage expected from particular type of material on distinct

slope. Varies with water content
After Oozier 1973 With permission of Zeitschnft fur Geomorphologie

a Compare figure 4 23
markedly with greater flow (table 4.7) because the displaced material will
extend farther downvalley than it would if moving as a sliding block. Some
uncertainty will remain, however, unless the classification index is used in con-
junction with other indices. Importantly, a definite inverse relationship was
found between D/L and four other morphometric indices (flowage, tenuity,
dilation, fluidity), each of which
is presumably controlled by the water content
of the material during movement.

its
The relative mobility of sturzstroms may also be estimated from the ge-
ometry of the deposits. Heim (1932) observed that the distance traveled by a
sturzstrom is a function of the height of the rockfall, the size of the mass, the
characteristics of the mass, and the characteristics of the route followed. Later
Shreve (1968) considered the H/L ratio to be an equivalent coefficient offric-
tion, since in sliding H
and L are related by
H= tan <f> L,
where tan is the coefficient of friction and H and L are respectively the height
<t>
of fall and the horizontal distance of the movement. The H/L ratio should,
therefore, indicate resistance and be inversely proportional to the mobility of
the movement. Hsu (1975), however, concluded that the equivalent coefficient
of friction is somewhat misleading, and he introduced another parameter called
the excessive travel distance (Le) to reveal mobility characteristics. This factor
is defined as the "horizontal displacement of the tip of a sturzstrom beyond
the distance one expects from a frictional slide down an incline with a normal
coefficient of friction of tan 32° (0.62)." It is expressed as
Le = L- ///tan 32°.
In general, Le seems to correlate positively to the volume of the fallen mass.

2° scarp
Figure 4.24.
Head Main (1°) scarp Different morphometric indices
shown in diagram can be used to
Crown describe and distinguish
processes of movement:
(A) landslip terminology;
(B) longitudinal section; (C) plan
view.
Lx convex
Transverse crevice
Toe Terracette
Surface of
rupture
Scar
ass
Tip
Radial
crevice
Hollow
(A)
Parent slope
Original surface
D&Dc
New surface
Left flank
Buried original surface Surface of rupture
(B) (C)
Table 4.7 Average values of the depth/length ratio in different types of landslips
as calculated for different areas.
Type of Landslip Average D/L Ratio
Flows 1.58
Planar slides 6.33
Rotational slides 20.84
After Crozier 1973 With permission of Zeiischrill ft/r Geomorphologie

Chapter 4 1 44
Although more work is needed before a clear relationship between mor-

phology and process can be defined, the morphometric approach exemplified
by these studies holds real promise. Predictions of slope stability, possible modes
of failure, and areas that might be affected are potential benefits if we can
understand how previous movements occurred in any given region.
Slope Profiles Profiles developed on the surface of natural slopes are regarded as reflections
of the major geomorphic factors — climate, rock type and structure, time, and
process. This rather innocuous statement leaves much to be desired because
it may suggest to you that the relationship between slope form and geomorphic
factors is straightforward. On the contrary, there is no simple method to de-
cipher which factors will determine the precise characteristics of a slope pro-
file. In fact, all of the factors are involved in some way to produce slopes.
Therefore, many debates about the development of slopes revolve around how
much a factor is involved in slope formation rather than whether a factor is
involved. Invariably this leads to individual value judgments since very few,
if any, studies can absolutely isolate the effect of one factor by keeping all
others constant.
Geomorphologists have paid considerable attention to the geometry of
slopes and the angles developed on different parts of the profile. Ideally, slope
profiles can be divided into four, general components (fig. 4.25): a convex upper
segment, a segment having a constant slope
cliffface (or free face), a straight
angle, and a concave segment at the hillslope base (Wood 1942; King 1953;
Carson and Kirkby 1972). More detailed distinctions of components have been
suggested. For example, the classification shown in figure 4.26 recognizes nine
slope components and additionally suggests that each is associated with pro-
cesses that probably dominate the zone. In any actual situation, all of the com-
ponents shown may not be present in the profile, or they have negligible
Figure 4.25. significance. The upper convexity, for instance, is usually more prevalent in
Major components of slope humid-temperate regions than in semiarid or arid climatic zones because the
profiles CC =concave segment;
soil creep process is known to be more important in the humid environment
S = straight segment, CF = cliff
face; CV = convex segment (fig. 4.26).
In addition to the types of components found in profiles, measurements
in a variety of climatic zones have revealed the interesting fact that slope an-
gles appear
to be concentrated in groups with rather small ranges of values
(see Carson and Kirkby 1972; Young 1972). Most pronounced are those that
cluster at 43°-45°, 30°-38°, 25°-29°, 19°-21°, 5°-ll°, and l°-4°. Al-
though any slope angle is possible, the frequency of these recurring groups is
tantalizing to geomorphologists because it probably reflects the underlying
control of the great geomorphic variables.
Each of these groups has well-defined maximum and minimum values,
which have been termed limiting angles by Young (1972) and threshold an-
gles by Carson and Kirkby (1972). The general interpretation of the angular
distribution is that angles within any group represent a stability regime for
ndicates movement in
a downvalley direction
Figure 4.26.
slopes formed in a particular climatic and lithologic setting. Under those con- Diagrammatic representation of
ditions, threshold values can be exceeded if the intrinsic properties of the parent the hypothetical nine unit
landsurface model.
material are altered or if the climate changes. When threshold values are
reached, any further change requires a fundamental response in the system
that adjusts the slope angle and places it within a different stability group.
Exactly why and how slopes adjust, however, is a debatable question. Ac-
cording to one hypothesis, groups and their limiting angles represent char-
acteristic angles for the processes that are working on the slopes. Thus, it is
clear that geomorphologists recognize process as an important ingredient in

the development of slope components and slope angles.
Chapter 4 146
Figure 4.27.
The relationship between mass
strength and profile angle for all
rock units studied in Antarctica

and New Zealand.
I I I I I I I I
10 20 30 40 50 60 70 80 90 100
Angle of slope (S°)
The was understood years ago by Gilbert (1877),

significance of process
who felt was controlled by either weathering
that the development of any slope
or sediment transport. Since then, the terms weathering-limited and trans-
port-limited have been generally accepted to describe slopes formed under
each process control. Weathering-limited slopes are created where the rate of
soil or regolith production is lower than the rate of its removal by erosion. As
a result, most of these profiles are determined by the character of the parent
rock.Such profiles seem to prevail in dry climates or in mountainous terrain
where erosion is rapid. In contrast, transport-limited slopes are formed where
the rate of weathering is more rapid than erosion. Slopes produced under this
regime normally develop on any unconsolidated parent material regardless of
environment, but they are typically dominant in humid-temperate zones where
is continuous. These
vegetation cover profiles are less affected by parent rock
and more dependent on the type and rate of slope processes.
Selby (1982) has made a cogent argument that weathering-limited slopes
are directly dependent on the relative resistance of the underlying parent rocks.
As evidence, he has demonstrated a high correlation between rock mass
strength (see table 4.3) and the angle developed on various slope segments
(fig. 4.27). A line drawn around the data points shown in figure 4.27 creates
what Selby calls the strength equilibrium envelope, and the slopes represented
by points within that envelope are referred to as strength equilibrium slopes.
Presumably, as long as the mass strength is maintained, strength equilibrium
slopes will keep a constant angle and the slope surface will retreat parallel to
itself.
It is easy to visualize a cliff face controlled by a constant mass strength
and retreating in a parallel manner. Whether we can expect rock strength to

remain constant is debatable, however, because as discussed earlier once joints
form in bedrock, the process becomes self-generating and more of the rock
will become fractured. This, of course, decreases rock mass strength and re-
quires an adjustment in the slope angle. In addition, during the process of
slope retreat, material eroded from the cliff face by rockfalls accumulates lower
in the profile as a talus. Technically, talus refers to the slope formed from the
accumulation of debris eroded from a cliff face, although many geomorphol-

ogists have used the term in reference to the debris itself. In either sense, talus
extends upslope with time and eventually may bury the original rock face (fig.
4.28). If and when this occurs, recognize that a transition has been made from
a weathering-limited slope to a transport-limited slope and the strength of the
parent rock is no longer significant. Instead, it is the relationship between re-
sistance of the talus debris and the erosional forces that should determine the
slope angles.
Theoretically it can be shown that in cohesionless material the angle of
internal friction 4> is equal to the slope angle 6 (Carson and Kirkby 1972).
Therefore, a strong correlation should exist between slope angles and the phys-
ical strength properties of unconsolidated material. In talus accumulates, values
are uncommonly high (43°-45°) because the mass is densely packed and the
rock fragments are interlocked (Carson and Kirkby 1972). If the void per-
centage is and the stable angle of slope
large, little pore pressure will develop
Figure 4.28.
will correspond to the angle. Continuous breakdown of the talus deposits
Upslope extension of talus slopes
without much clay formation would produce a sandy matrix that should stand
near the angle of repose for cohesionless sands, approximately 35°.
There is ample evidence that in a wide variety of climates a weathered
mixture of rock rubble and soil underlies slopes between 25° and 28°. (Young
1961; Melton 1965b; Robinson 1966). As the original talus deposits are pro-
gressively broken down by weathering, the mass gradually loses its open-pore
framework. During times of abundant water and high water tables, the ma-
terial attains positive pore pressures that reduce the effective normal stress by
buoyancy. This obviously lowers shear strength and changes the relationship
between internal friction and the potential failure surface. Thus, the recur-
rence of slope angles at 25°-29° may lie in the mechanics associated with
saturated soils. As summarized by Skempton (1964), cohesionless materials
subjected to pore pressures are likely to experience shallow landsliding along
failure planes that approximate
tan 6 = '/2 tan </>.
Assuming an original angle of 45° for coarse talus deposits and 35° for a
sandy mantle, the stable slope developed on these materials when they become
saturated would be about 26° and 19°, respectively. In clay-rich soils the 4>
angle is much lower, and stable slope angles are considerably less.
Carson (1969) proposed that instability in slopes requires the progressive

replacement of steep slopes by gentler ones. In this model, many landscapes
should go through more than one phase of instability, but the exact number
depends on the characteristics of the rocks and how they ultimately break
down. In the initial stage, a steep rock cliff is replaced by talus or slopes de-
veloped on thoroughly fractured rocks. This phase might be followed by a
Chapter 4 148
change to lower slopes, and eventually to the gentle slopes formed on clay-
rich soils. Each slope is only temporarily stable, for as weathering changes the
mantle's properties and pore pressures vary, the mass reaches its slope threshold
value. Further change causes the slope to adjust rapidly into a new stability
range consistent with the revised properties of the mantle. Because of the vari-
ability of soil properties and pore pressures, any limiting angle values are pos-
sible, even though they apparently cluster in recurring groups. The types of
material, the number of instability phases, and the threshold values combine
in any area to control the progression of slope development. The net effect of
the variables is eventually to form slopes that have long-term stability with
respect to rapid mass movements; at that point, creep and surface water ero-
sion become much more significant as slope processes.
The salient point of this discussion is that recurring angles measured on
slopes may be easily explained by the relationship between erosive process and
the different strengths of unconsolidated materials caused by textural varia-
tions. However, whether all slope materials experience an evolution in texture
as envisioned by Carson is debatable, and perhaps unnecessary to explain slope
angle and form.
Our discussion of slope profiles thus far has attempted to demonstrate
that processes of weathering and erosion are intimately involved in slope de-
velopment. Process, however, is not an independent variable because it is di-
rectly controlled by climate and geology. In fact, of the many variables cited
as being responsible for hillslope form, only geology and climate can be con-
sidered as independent variables.
The Rock-Climate Influence

It was shown earlier that slopes in weathering-limited situations are controlled
by the mass strength of the parent rock. This is especially significant in the
maintenance of a cliff face. The lithologic influence on slopes is shown in both
declivity and profile shape. Coherent rocks tend to support steeper slope an-
gles, and with equal cohesion, the more massive the bedding, the steeper the
slopes. Where strata contain alternately weak and resistant rocks, an irregular
profile may develop, and resistant units will assume higher than normal angles
where they overlie weaker rocks.
In regions where a cliff face is not present, lithology may still exert a
control on slopes. It is an accepted fact that topography generally reflects lith-
ology and that "resistant" rocks underlie hills and nonresistant rocks become
the valleys. In this sense, however, resistance is not defined by intrinsic prop-
erties of a particular rock type but is a relative feature determined by how
rapidly slopes developed on the rock retreat and whether the rock stands rel-
atively high in the local topography (Young 1972). Therefore, it is not so much
the rock itself that determines resistance, but whether the slopes formed over
the rock are controlled by processes of weathering or processes of removal. If
a slope is weathering-controlled, resistance is related to how rapidly the rock
is weathered; it is a direct function of the rock properties. In transport-limited
slopes, the resistance is attributable to the rate at which regolith can be eroded:
the properties of the weathered mass and the type and magnitude of the ero-
sional processes become important in slope development. For these reasons,
the resistance of a particular rock type and its influence on slopes can be re-
versed if the rock is located in different climates. For example, the charac-
teristics of slopes formed on limestones in humid climates contrast markedly
with those developed in arid climates.
With regard to climatic influence, geomorphologists have long recognized
that the most common slope profile in humid-temperate regions is a distinct,
convex upper slope and a concave lower slope. Contrary some beliefs, straight
to
slope segments do occur in regions with a humid-temperate climate, and some
profiles do contain steep cliff faces. Most cliff faces, however, are ephemeral
in the sense that as soon as undercutting ceases, a talus slope forms and will
extend upslope until it covers the original cliff wall (fig. 4.28). If the lithology
of the rock sequence underlying the slope is not uniform, cliff faces may persist
because resistant units are maintained as caprocks when the weaker under-
lying strata retreat faster, essentially undercutting the stronger rocks.
Convex upper slopes are usually interpreted as a function of soil creep;
the lower concavity probably results from soil wash, although not all slopes
have this segment, particularly when there is active erosion at the slope base
(Strahler 1950). The convexo-concave profile is most likely to be attained after
mass movements have produced a long-term angular stability. At this stage,
creep and wash become the dominant slope processes; the straight segment,
representing stability of slope material, is gradually diminished in size. The
processes of water erosion on slopes will be discussed in the next chapter. Rec-
ognize here, however, that water flowing over and through slope material com- -Cliff
bines with mass movement to mold slope profiles, and in some cases water Debris slope
erosion may be the dominant process involved. I
Desert plain
Semiarid and arid climates tend to engender slope profiles that are more
angular than those found in humid-temperate regions, even though the same
convex, straight, and concave segments may be present (fig. 4.29). Steep cliffs (A) Typical arid slopes
usually are present above a straight, debris-covered segment that normally

stands at angles between 25° and 35°. At the base of the straight segment a Convex
pronounced change in slope occurs, and angles decrease over a short distance
to less than 5°, a normal slope for most desert plains. The limited vegetal cover
-Straight
and low precipitation associated with arid zones assure that mass movements
Concave
occur at higher angles and that creep is subordinated to wash. As a result the
upper slope convexity, so prominent in humid regions, is much less pro-
(B) Typical
nounced. humid-temperate slopes
Straight segments are maintained by the wash process, which is accel-
erated on the sparsely vegetated surfaces. Unlike similar segments in humid Figure 4.29.
climates, these usually have only a thin veneer of rock debris. They are not, Typical slope profiles in (A) and
regions and (B) humid-temperate
then, slopes of accumulation such as talus slopes but instead probably rep-
regions
resent true slopes of transportation, on which the amount of debris supplied
segment from the cliff face or by weathering of the underlying
to the straight
rocks removed in equal quantities to the desert plain. The angle of slope
is
represents some balance between the processes that break debris down and
Chapter 4 150
the actual transporting mechanism (Schumm and Chorley 1966). Most stu-
dents feel that a general relationship between particle size and slope angle can
be demonstrated.
Although other climatic regimes have characteristic slope forms, in most
cases they are produced by the same mechanics that operates in the humid-
temperate or arid zones. In the periglacial environment a special influence is
exerted by magnified frost activity; a more extensive treatment of that envi-
ronment is presented in chapter 1 1.
Very little research has attempted to determine what aspects of hillslope
profiles are most closely related to climate. An example of this approach is a
study by Toy (1977). Toy utilized a rigorous statistical analysis to compare
slope properties within two extended traverses in the United States (Kentucky
to Nevada and Montana to New Mexico) along which considerable climatic
Slope replacement variation is experienced. The selection of sampling localities was stringent.
Parent rock at each measuring site was restricted to shales dipping at less than
12 3 4 5°. Each slope analyzed was south-facing, within 5 miles of a weather station
having records for the same 21 -year period used as the climatic base, and had
no effects of human activity. Toy found that climate could account for 59 per-
cent of the variability in the upper convex segments and 43 percent of the
variability in the slope of the straight segments. Arid slopes in this study were
shorter, had steeper straight segments, and had shorter radii of curvatures
Parallel retreat
developed at the convex crests than slopes in humid regions. In addition, of
the climatic variables used in the study, those most closely associated with
Figure 4.30.
Three hypotheses ot slope slope variations were (1) spring and summer precipitation, (2) potential
evolution. Higher numbers evapotranspiration, and (3) water availability (total precipitation minus total
indicate increasing age of the
potential evapotranspiration during the 21-year period).
slope. (Adapted from Young,
1972, fig. 14, in Slopes With Toy's findings cannot be used to make sweeping generalizations about the
permission of A Young ) climatic effect on slope profiles because they apply only to one type of parent
rock. However, the study is a good demonstration of the research design needed
to estimate the influence of one geomorphic factor by reducing or eliminating
the effects of others.
Slope Evolution
In addition to geology and climate, the factor of time can also be considered
as an independent variable. Its effect, however, is difficult to determine, es-
pecially when the time interval involved is very long. As we saw in chapter 1.
some of the great debates in geomorphology revolve around the question of

how slopes respond to continued erosion. Do slopes progressively flatten through
time, providing landscapes with evolutionary steps or stages? Or do slopes
reach an equilibrium between form and geomorphic factors that is maintained
through time because slopes retreat in a parallel manner? Unfortunately, these
questions are more easily asked than answered.
Three main types of slope evolution have been suggested: slope decline,
slope replacement, and parallel retreat (fig. 4.30) In slope decline, the steep
upper slope erodes more rapidly than the basal zone, causing a flattening of
the overall angle. It is usually accompanied by a developing convexity on the
upper slope and concavity near the base. Slope decline alone cannot in fact
explain a concave profile on the lower slope unless some deposition occurs at
Figure 4.31.
Abandoned wave-cut bluffs, Michigan Profiles of bluffs of three different
ages developed in the same
material.
<**&*^.-•'
.00 5 .00 10.00 meters
(no vertical exaggeration)
the base. In slope replacement, the steepest angle is progressively replaced by

the upward expansion of a gentler slope developed near the base. This process
tends to enlarge the overall concavity of the profile, which may be in either a
segmented or a smoothly curved form. Slopes evolving by parallel retreat are
characterized by the maintenance of constant angles on the steepest part of
the slope. Absolute lengths of slope parts do not change except in the concave
zone, which gets longer with time.
There is little doubt that all three types of slope evolution can be dem-
onstrated in actual situations (Savigear 1952; Brunsden and Kesel 1973; Cun-
ningham and Griba 1973; Haig 1979; Nash 1980; Selby 1980, 1982; Colman
1983). In general, slope decline is most notable in humid regions, and parallel
retreat seems to be more prevalent in arid regions.
In most of the foregoing cases, demonstration of slope evolution is based
on the ergodic hypothesis (Chorley and Kennedy 1971), which contends that,
in the proper setting, spatial elements can be considered as equivalent to time
elements, and space-time transformations are therefore acceptable. As an ex-
ample, Nash (1980) suggests that profile variations in a series of lake bluffs
inMichigan represent slope changes that have occurred during intervals be-
tween the development of the different bluffs. Because each bluff was formed
in the same morainic material, it isassumed that the oldest bluff originally
had a profile similar to that of the modern bluff. Therefore, the observed dif-
ferences between the two profiles represent changes that have occurred on the
older slope since the time of its formation (fig. 4.31).
The types of studies just mentioned have inherent value because they are
based on real observations. However, they are probably valid only if significant
changes in other geomorphic variables, such as climate, do not occur during
the tenure of slope development. Thus, they are restricted to geologically short
time spans. Because of this, the possibility remains that the observed changes
are merely transitions toward an equilibrium form that, when attained after
a greater period of time, would experience no further profile alteration.
Finally, it should be pointed out that numerous attempts have been made
to characterize slope evolution by employing theoretical techniques such as
numerical and simulation models (for references see Carson and Kirkby 1972;
Young 1972; Selby 1982). Although modeling suggests routes that slope evo-
lution may follow, the techniques suffer because the assumed character of the
Chapter 4 152
original slope profile is pure conjecture; i.e., there is no sure way to know the
form of the initial profile. As a result, some geomorphologists believe they add
little to our comprehension of slopes unless they are based on long-term and
detailed field measurements (Dunkerley 1980; Selby 1982).
Summary The processes of physical weathering tend to break rocks and unconsolidated
debris into smaller particle sizes. The force needed to accomplish this disin-
tegration is provided by expansion resulting from unloading, hydration of min-
erals, or growth of foreign substances in spaces within the parent material.
Many important processes of disintegration require the presence of water.
Physical weathering, combined with gravity, is instrumental in determining
the type and rate of mass movements; ultimately it has a direct bearing on
the slopes developed in any region.
Mass movements occur as slides, flows, and heaves, or by water-induced
transport of surface debris. The magnitude and type of mass movement are
partly dependent on the physical properties of the parent material. Sheai
strength (a function of internal friction, effective normal stress, and cohesion)
determines how vigorously any substance will resist the force attempting to
produce mass movement. Thus, slope failure or other mass movements can
result from an increase in shear stress (driving force), a lowering of shear
strength (resistance), or both. Physical weathering tends to decrease the shear
strength of materials and thereby helps to initiate mass movements and con-
trol the form of the resulting slopes. Climate and lithology interact to influence
slope profiles. The effect of time is shown by the manner in which slopes evolve.
Carson, M., and Kirkby, M. 1972. Hillslope form and process. London:
Cambridge Univ. Press.
Crozier, M. J. 1973. Techniques for the morphometric analysis of landslips.
Zeit.f. Geomorph. 17: 78-101.
Grim, R. 1962. Applied clay mineralogy. New York: McGraw-Hill.
Oilier, C. D. 1969. Weathering. Edinburgh: Oliver and Boyd.
Selby, M. J. 1982. Hillslope materials and processes. Oxford: Oxford Univ.
Press.
Terzaghi, K. 1950. Mechanism of landslides. In Application of geology to
engineering practice, edited by S. Paige, pp. 83-123. Geol. Soc.
America Berkey Vol.
Varnes, D. J. 1978. Slope movement types and processes. In Landslides —
Analysis and control, edited by R. Schuster and R. Krizek, pp. 12-33.
Washington, D.C.: Transportation Research Board Spec. Rpt. 176.
Young, A. 1972. Slopes. Edinburgh: Oliver and Boyd.
The Drainage Basin Development, —
Morphometry, and Hydrology
5
I. Introduction B. Surface Water
II. Slope Hydrology and Runoff 1. Paleofloods
Generation 2. Effect of Physical Basin
A. Infiltration Characteristics
B. Subsurface Stormflow and 3. Morphometric
Saturated Overland Flow Relationships
III. Initiation of Channels and the V. Ba: >in Denudation
Drainage Network A. Slope Erosion and Sediment
A. Basin Morphometry Yield
1. Linear Morphometric 1. Wash
Relationships 2. Sediment Yield (Soil
2. Areal Morphometric Loss)

Relationships 3. Factors Affecting
3. Relief Morphometric Sediment Yield
Relationships a. Precipitation and
B. Basin Evolution Vegetation
IV. Basin Hydrology b. Basin Size
A. Underground Water c. Elevation and Relief
1 . The Groundwater d. Rock Type
Profile e. The Human Factor
2. Movement of B. Sediment Budgets
Groundwater C. Rates of Denudation
3. Aquifers, Wells, and VI. Summary
Utilization Problems VII. Suj *gested Readings
153
Chapters 154
Introduction Having examined the processes that break rocks apart and transport debris
down slopes by mass movements, we are now ready to consider rivers how —
they form and how sediment and water are delivered to the river system. Two
basic truths about rivers were realized long before geomorphology emerged
as an organized science: (1) streams form the valleys in which they flow, and
(2) every river consists of a major trunk segment fed by a number of mutually
adjusted branches that diminish in size away from the main stem. The many
tributaries define a network of channels that drain a discernible, finite area
recognized as the drainage basin or watershed of the trunk river.
Each basin is separated from its neighbor by a "divide," and so basins

serve as excellent fundamental units of geomorphic systems. Any feature, flu-
vial or otherwise, within a basin can be reasonably considered as an individual
subsystem of the basin, having its own and energy
set of processes, geology,
gains and losses. Furthermore, because measure the amount
it is possible to
of water entering a basin as precipitation and the volume leaving the basin as
stream discharge, hydrologic events can be readily analyzed on a basinal scale.
Similarly, most of the sediment produced within the basin limits is ultimately
transported from the basin via the trunk river. Thus, considered on a long
temporal scale, the rate of lowering of the basin surface can be estimated. The
mechanics of fluvial processes usually reflect some balance between the sed-
iment to be transported and the water available to accomplish the task.
Most earth scientists are introduced to watersheds by learning that
drainage patterns or individual stream patterns often mirror certain traits of
the underlying geology, as figure 5.1 and table 5.1 show. The gross character
of patterns is a useful tool in structural interpretation (Howard 1967) and as
a first approximation of lithology; a search of topographic maps and aerial
photographs for a distinctive network arrangement is a logical first step in the
study of regional geology.
In a hydrologic sense, however, prior to World War II, most basins were
described in qualitative terms such as well-drained or poorly drained, or they
were connoted descriptively in the Davisian scheme as being youthful, mature,
or old. The mechanics of how river channels or networks actually form or how
water gets into a channel was understood (or misunderstood) only in vague
terms by both geologists and hydrologists. Realizing this early twentieth-cen-
tury view of streams and drainages, it is startling to examine the avant-garde

approach presented by R. E. Horton during this period. His attempt to explain
stream origins in mathematical terms and to describe basin hydrology as a
function of statistical laws can be cited as the birth of quantitative geomor-
phology. We now know that many of Horton's original ideas are only partially
correct. However, there can be little doubt that modern geomorphic analysis
of drainage basins has its roots in Horton's original work, and his thinking
was instrumental in the rise of a new breed of geomorphologist (Horton 1933,

1945).
155 The Drainage Basin —Development, Morphometry, and Hydrology
Figure 5.1.
Basic drainage patterns.
Descriptions are given in
table 5.1.
Dry-<?
(E) Radial (F) Annular (G)Multi-basinal (H) Contorted
Table 5.1 Descriptions and characteristics of basic drainage patterns illustrated

in figure 5.1.
Basic Significance
Dendritic Horizontal sediments or beveled, uniformly resistant, crystalline rocks.

Gentle regional slope at present or at time of drainage inception. Type
pattern resembles spreading oak or chestnut tree
Parallel Generally indicates moderate to steep slopes but also found in areas of
parallel, elongate landforms. All transitions possible between this
pattern and type dendritic and trellis.
Trellis Dipping or folded sedimentary, volcanic, or low-grade metasedimentary

rocks: areas of parallel fractures; exposed lake or sea floors ribbed by
beach ridges. All transitions to parallel pattern Type pattern is regarded
here as one in which small tributaries are essentially same size on
opposite sides of long parallel subsequent streams
Rectangular Joints and/or faults at right angles Lacks orderly repetitive quality of trellis
pattern; streams and divides lack regional continuity
Radial Volcanoes, domes, and erosion residuals A complex of radial patterns in a

volcanic field might be called multiradial
Annular Structural domes and basins, diatremes, and possibly stocks.
Multibasinal Hummocky surficial deposits, differentially scoured or deflated bedrock;

areas of recent volcanism, limestone solution, and permafrost. This
descriptive term is suggested for all multiple-depression patterns whose
exact origins are unknown.
Contorted Contorted, coarsely layered metamorphic rocks. Dikes, veins, and

migmatized bands provide the resistant layers in some areas Pattern
differs from recurved trellis in lack of regional orderliness, discontinuity
of ridges and valleys, and generally smaller scale
After Howard 1967 Used with permission of American Association of Petroleum Geologists
Chapter 5 156
Figure 5.2.
The slope hydrologic cycle Some
of the precipitation (P) is
intercepted by vegetation (I) or
lost by evapotranspiration (ET)
Upon reaching the ground
surface, becomes part of stream
it
discharge (Q) by direct runoff (R),
interflow (IF), or groundwater flow

(GW) after it reaches the water
table (V)
Slope Hydrology and The ultimate source of river flow is, of course, precipitation, which represents
Runoff Generation the major influx of water to any drainage basin. Precisely how much of that
precipitation actually becomes part of stream flow and what route a particular
drop of water follows to reach a channel are topics of great concern to hy-
drologists. In fact, they constitute the components of what is commonly re-
ferred to as the slope hydrological cycle (fig. 5.2).
Rainfall does not usually strike the ground surface directly because its
fall is impeded by leaves or trunks of the vegetal cover. This process, called
interception, may significantly reduce the amount of rainfall striking the sur-
face. Interception losses are quite variable because they depend on numerous
meteorological factors, vegetation types, land use, and seasonality. In addition,
the loss is dependent on storm duration, being high initially and decreasing
as the storm continues. During long storms, the vegetation may eventually
become saturated and all additional rainfall is passed to the surface. None-
theless, interception may remove 10 to 20 percent of precipitation where grasses
and crops are the dominant vegetation and up to 50 percent under a forest
canopy (Selby 1982). Rainwater that does reach the ground surface may still
be prevented from becoming part of stream flow because it is susceptible to
further loss by evapotranspiration.
Infiltration
Once on the ground surface, precipitated water may follow different routes to
a stream channel. The water flowing into a channel in direct response to a
precipitation event is called storm runoff or direct runoff. Hydrologists have
always been concerned about how much runoff will issue from any precipi-
tation event and how quickly the runoff will enter the stream channels. These
Figure 5.3.
Flood hydrograph
factors determine if and when a flood will occur and how high the river level
will be at peak flow. The response of a stream to a storm is determined by the
use of a flood (or storm) hydrograph, which depicts river flow as a function
of time (fig. 5.3). The volume of water (discharge) appears on the hydrograph
in theform of direct runoff, or as base flow (water infiltrated to the water table
and released more slowly from the underground system). Although discharge
measured at a station may contain both of these components, it should be clear
that direct runoff occurs only in response to a storm, while in dry periods all
the discharge is base flow. The maximum or peak flow usually develops soon
after precipitation ends, separating the hydrograph into two distinct segments,
the rising limb and the recession limb. The rising limb in general reflects the
input of direct runoff, and the increase in discharge with time is relatively
rapid. The recessional phase, however, is more by the gradual de-
controlled
pletion of water temporarily stored in the system, and so the decrease in dis-
charge with time is less pronounced. Also the shape of the recession limb is
not influenced by the properties of the storm causing the increased discharge
but is more closely related to the physical character of the basin. Since a hy-
drograph represents the sum of hydrographs from all subareas of a basin, if
geology and topography are fairly constant throughout, then rainfalls having
similar properties should generate hydrographs with the same shape. On this
premise, a type hydrograph for a basin, called a unit hydrograph, has been
developed, in which the runoff volume is adjusted to the same unit value (one
for the entire area). The unit hydrograph has been used as a connecting link
in many morphometry to hydrology.
studies attempting to relate basin
Horton suggested that the primary control of storm runoff is how
In 1933
easily rainwater can sink into the material on a sloping surface (a process
known as infiltration). In general terms the infiltration theory of runoff can
be described as follows. Imagine a hypothetical slope that is part of an un-
dulating surface topography developed by weathering and mass movements.
As precipitation falls on the slope surface, the water is absorbed into the ground
at a rate called the infiltration capacity, which is a function of a number of
factors such as soil texture and structure, vegetation, and the condition of the
Chapter 5 1 58
surface prior to precipitation. Thus, infiltration capacity varies on a large scale
-Infiltration capacity
depending on regional geology, but it can also change on a local basis (even
on a given slope) if the controlling factors vary. In addition, infiltration ca-
Rainfall intensity
pacity usually changes markedly during any precipitation event. As figure 5.4
shows, it usually starts with a high value that decreases rapidly in the first
several hours of the storm and then more slowly as rainfall continues, until it
finally attains a reasonably constant minimum value. The infiltration capacity

changes because surface conditions are being changed, especially as aggre-
Time- gated soil clumps are broken apart and surface entry by the water becomes
more difficult as pores become clogged with the clay particles. Also, the in-
Figure 5.4. filtration capacity can be only as great as the lowest transmission rate in sub-
Infiltration capacity and rainfall
surface horizons. In prolonged storms, for example, a clay pan or a caliche
intensity plotted against time
Infiltration capacity decreases zone in the B horizon may determine the ultimate infiltration rate. In the in-
with duration of storm Runoff terval between rains, the infiltration capacity rises again as the surface dries
occurs only when rainfall intensity
is greater than infiltration capacity
and regains its aggregated structure, and the storage space within the soil in-
creases as water gradually drains downward. The frequency of rainfall there-

fore becomes a significant factor, because a rapid succession of rains without
enough time intervening will prevent the infiltration capacity from returning
to its original high value. Rains of relatively small intensity can then trigger
disastrous floods.
In Horton's model, as long as the infiltration capacity exceeds the rate at
which rainfall strikes the surface (known as rainfall intensity), all incoming
water will be infiltrated and none will run off. However, in those periods when
rainfall intensity exceeds the infiltration capacity, runoff will occur as water
moving down the slope surface, a process now known as Hortonian overland
flow. The velocity of this flow usually falls in a range between 10 and
500m/hr, meaning that flow initiated at the top of a 100-meter slope would
enter the channel within minutes or hours of its development (Dunne 1978).
Thus, Hortonian overland flow was considered to be the primary determinant
of peak flow and total direct runoff (fig. 5.3) that appears in stream channels
in response to a storm. The basic assumption here is that all infiltrated water
is delayed in its procession to stream channels because it percolates downward
to the groundwater table, and from there moves riverward at the slow velocity
of groundwater flow Such base flow would eventually be part of river
(fig. 5.2).
flow but would enter the channel after the direct runoff accruing from the
initiating storm was moved downstream. This interpretation seemed to be sup-
ported by other hydrologic analyses, especially the unit hydrograph concept
being developed at the same time (see discussion by Chorley 1978), and was
readily accepted as the basic model of storm-generated runoff.
Overland flow traverses a surface as a broad shallow sheet or in small,
linear depressions called rills. The flow is most common on slopes having lim-
ited soil development and a sparse vegetal cover such as those in arid or semi-
arid regions or in mountainous terrains. We now know, however, that Hortonian
runoff is minor or nonexistent in humid-temperate regions that are densely
vegetated and have well-developed soils (Kirkby and Chorley 1967; Rawitz et
1 59 The Drainage Basin —Development, Morphometry, and Hydrology
al 1970; Dunne and Black 1970a, 1970b). In those environments, even steady-
state infiltration capacities normally exceed rainfall intensities, and only rare
precipitation events can produce Hortonian overland flow.
In light of the above, we must ask the logical question as to where the
water comes from to produce the rapid peak flows observed in small basins of
humid regions if all rainwater reaching the surface is being infiltrated.
Subsurface Stormflow and Saturated Overland Flow

A major flaw in Horton's original model was the belief that infiltrated water
moves directly downward under the influence of gravity. Research subsequent
to Horton's analyses has demonstrated clearly that a lateral component of
subsurface flow also exists in most slope materials. As water moves downward
from the surface as a wetting front to the water table, it often encounters a
soil horizon of low permeability that not only influences the equilibrium in-
filtration capacity but also tends to divert water downslope along its surface.
This happens because a saturated condition builds up above the low perme-
ability zone, and lateral flow is initiated parallel to the barrier. This type of
water movement, known as interflow or throughflow (Kirkby and Chorley
1967), can occur above the water table and allow water to follow a more direct
route to the channel than normal groundwater. Throughflow is probably most
common in the more permeable A horizon.
Where no barrier exists, infiltrated water will reach the water table and
elevate its position. The water table rises rapidly adjacent to the channel, where
antecedent soil moisture is greatest, and slowly in the upper-slope zone. As a
result, the water table steepens immediately next to the channel and generates
accelerated groundwater flow in that area. The combination of throughflow
and accelerated groundwater movement is called subsurface stormflow. It was
first recognized and defined at the same time as Horton's infiltration theory
(Hursh 1936; Hursh and Brater 1941), but the significance of the concept was
not fully realized until much later.
Subsurface stormflow from the lower parts of slopes may actually produce
runoff in the form of bank seepage in the early part of a given storm. There-
fore, it is possible for some of the subsurface stormflow to contribute to the
peak discharge. It is some situations much of
also possible, however, that in
the storm runoff is generated in the subsurface through macropores such as
root channels or "pipes," which are linear openings having a much greater
permeability than the normal soil matrix (Mosley 1979). These linear path-
ways provide a rapid delivery of subsurface flow to the channel and under the
may be significant contributors to the rising limb of the storm
proper conditions
hydrograph. Nonetheless, most studies consider subsurface flow to be of minor
consequence except during the recessional phase of runoff following peak flow.
This occurs because measured rates of subsurface stormflow range between
Chapters 160
Rain Rain
Figure 5.5.
Runoff from a steep, well-drained
slope in northern Vermont
(A) Early in storm saturated zone
I I I J J I I lllllll
(shaded) yields a small amount of
subsurface stormflow (SSSF).
(B) Late in storm water table rises
to surface Subsurface water
returns to the surface as return
flow (RF). Precipitation on the
saturated area (DPS) adds to the 10 20 30 40 "0 10 20 30 40
return flow (From Water in
Distance from base of slope (meters) Distance from base of slope (meters)
Environmental Planning by
T Dunne and L. B Leopold (A) (B)
Copyright© 1978 by
W H Freeman and Company. All
rights reserved)
0.003 and 1.0 cm/hr (Dunne 1978), which is much too slow to become part
of the peak discharge in most small basin flows. This slow response requires
that another source of runoff be present to produce the rapid peak flow con-
ditions.
The results of detailed studies conducted in Vermont (Ragan 1968: Dunne

and Black 1970a, 1970b) showed that in many storms the level of the water
table rises until it actually intersects the ground surface near the stream channel
(fig. 5.5). In addition, zones of throughflow partially fed by laterally moving,
infiltrated water may become saturated to the surface (Kirkby and Chorley
1967). This creates a situation where some of the infiltrated water is returned
once again to the surface and begins to flow downslope toward the channel.
The flow is aptly called return flow. Return flow is obviously a form of overland
flow, but its origin and location is distinctly different from the Hortonian over-
land flow discussed earlier. The velocity of return flow is much greater than
subsurface stormflow, attaining speeds of from 3 to 15 cm/sec (Dunne 1978).
In addition, direct precipitation on the saturated area marked by return flow
or zones of saturated throughflow adds significant amounts of water to the
volume of direct runoff.
The combination of return flow and direct precipitation on saturated areas
is called saturated overland flow. It is now documented that in many humid
areas saturated overland flow is the major contributor of direct runoff to stream
channels. For example, Ragan (1968) estimated that on average it supplied
55 to 62 percent of the total storm runoff in a small watershed near Bur-
lington, Vermont, and was predominant in determining peak discharge. Sub-
surface storm flow provided 36 to 43 percent of the total flow and exerted its
greatest influence during the recessional phase of the runoff.

It now seems clear that there are many sources of direct runoff other than
Hortonian overland flow. It also is becoming apparent that the area of a wa-
tershed that actually provides runoff during any storm is not constant. This
perception, known as the variable source concept (also called the partial area
..:
North
(A) (B)
Summer Immediately after

100 200 300 400 feet
snowmelt period 10 feet

Contour interval
Autumn
Figure 5.6.
Variations in saturated areas on
concept), essentially indicates that the area over which quick runoff occurs well-drained hillslopes near
Danville, Vt. (A) Seasonal changes
varies seasonally and during a given storm (fig. 5.6). This change is funda-
of pre-storm saturated area
mentally controlled by topography, soil characteristics, antecedent moisture, (B) Expansion of saturated area
and rainfall properties. Areas with moderate to poorly drained soils, gentle during a single, 46 mm
rainstorm
Solid black is at beginning of
slopes, and concave recessions along the valley walls are prone to have the
storm Light shade represents
greatest expansion of contributing areas during storms and on a seasonal basis. saturated area at end of storm
Although the generalizations discussed above seem reasonable, they are where water table has risen to the
surface (From Water in
based on a limited number of studies and much more work is needed to refine
Environmental Planning by
the model (Selby 1982). Nonetheless, the observations about runoff genera- T. Dunne and L B Leopold
tion have an important bearing on the geomorphic processes operating within Copyright© 1978 by
W. H. Freeman and Company, All
a drainage basin, and further research in this field should be vigorously pur-
rights reserved)
sued (Freeze 1980).
In the Horton model when rainfall intensity exceeds the infiltration capacity, Initiation of Channels
overland flow occurs and only then does erosion become possible (fig. 5.4). For and the Drainage
a stream channel to develop, the erosive force (F) of the overland flow must Network
surpass the resistance (R) of the surface to being eroded. According to Horton
Chapter 5 162
(1945), as overland flow begins to traverse the slope, the force it exerts on a
depends on the slope angle, the depth of the water, and the
soil particle specific
weight of the water such that
F = y— sin 0,
where 7 is the specific weight, d the depth, and sin 6 the slope angle. Actually
the force (F) represents a shear stress exerted parallel to the surface by the
water.The stress progressively increases downslope because the depth rises as
more and more water is added to the volume of the overland flow. Depending
on the size of the slope material, a threshold stress is eventually reached at
some point on the slope where F>
R, and particles are dislodged or entrained.
The processes and magnitude of slope erosion by water will be discussed later
in the chapter when we examine sediment yield from drainage basins. Our
concern here is the initiation of channels on sloping surfaces.

Surface resistance (R) is affected by the type and density of the vegetal
cover. As we saw earlier, vegetation intercepts raindrops before they can strike
the surface, thereby preserving cohesion in the aggregated-clay soil structures.
In addition, rootlets tend to bind soil particles, and litter often serves as a
protective mat above the surface material. Vegetation also inhibits the free
flow of water and retards its velocity. In areas devoid of vegetal cover, the
surface commonly develops a hard crust as it dries in the direct sunlight, which
provides a high initial resistance to erosion. This may be destroyed during a
storm, indicating that resistance, like the force of overland flow, may vary as
a rainfall event progresses. It should also be noted that the factors that de-
termine resistance are essentially the same ones that control infiltration ca-
pacity.
Erosion by overland flow begins when F exceeds R, taking the form of a
series of shallow, parallel rills that are oriented perpendicular to the slope con-
tours. Slight variations of the surface topography produce a greater depth of
flow and more erosive force in the low spots. Erosion is accelerated at those
points, and a rilled surface results rather than a slope that discharges water
as an unconfined sheet. The actual point where rill formation begins depends
on how efficiently the force of overland flow increases as the water moves down
the surface. This ultimately relates to the infiltration capacity, the rainfall
intensity, and the resulting rate of runoff (the runoff intensity).
Assuming a constant slope and runoff intensity, the distance between the
watershed divide and the upper position of rills is a measurable segment called
the critical length (X,.); the surface between Xc and the divide is recognized
as a "belt of no erosion" (figs. 5.7 and 5.8). Horton considered the critical
length to be the most important single factor in the development of stream
networks, but its significance can perhaps be appreciated more fully on a local
scale because it is highly sensitive to changes in the factors that control ero-
sion. Assume, for example, that a farmer wishes to plant additional row crops
in an area within the belt of no erosion. As the land is cleared of its original
'
Figure 5.7.
Hypothetical slope showing
overland flow. No erosion occurs
until the force of overland flow (F)
Level of overland flow
exceeds the resistance of the
surface material (R). Upslope from
that point no erosion occurs. Xc is
the distance from the divide to
point where erosion begins (After
Horton 1945 Courtesy of The
F>R
Geological Society of America)
vegetal cover and replaced by the crops, the resistance and infiltration ca-
pacity are drastically lowered. Under prevailing precipitation more runoff oc- Divide
curs, the force of overland flow increases, and X c is shortened. The narrowing
Belt of no erosion
of the belt of no erosion allows rills and gulleys to form in the newly cropped
region. Depending on their depth, these may make the area impassable for 1
Rill
the heavy equipment needed for plowing and harvesting. erosion
The
HHI
analysis just presented is probably applicable in areas that are prone
to the development of Hortonian overland flow. In humid-temperate regions, ' '
'
however, overland flow emanating from the upper slope areas is rare, and
therefore the concept of a critical length has no particular significance. In
cases where saturated overland flow dominates, the erosive action probably
Figure 5.8.
begins closer to the slope base (Kirkby and Chorley 1967; Kirkby 1969), and Map view of critical length (Xc )
rills are gradually extended upslope. The headward growth most likely occurs and belt of no erosion Rill erosion
starts downslope from belt of no
because the headcut in any rill exposes a substratum that has a lower resis-
erosion when F > R. (Compare
tance to erosion than the slope surface. The formation of rills does not explain fig 5 7)
stream channels because a rilled surface is still part of a slope system. Their
development, however, is the first necessary step toward a true river. It is now
generally recognized that rills initiated on a slope cannot long remain as par-
allel unconnected channels. Deeper and wider rills develop where the length
over which erosion can occur is the greatest. These master rills carry more
water, and because of the greater depth, they undergo downcutting until all
the flow is contained within the channel and the rill becomes a tiny stream.
Because they become slightly entrenched, master rills capture adjacent rills
when bank caving or overtopping during high flow destroys the narrow divides
between them. The repeated diversion of rills, a process called micropiracy,
tends to obliterate the original rill distribution, and gradually the initial slope
parallel to the master channel is replaced by slopes on each side that slant
toward the main drainage line.
The development of new slope direction in accordance with the master

channel was called cross-grading by Horton (1945). In the final stage, only
one stream, confined in the master rill channel, crosses the slope. The side
slopes presumably develop a new rill system sloping to the position of the ini-
tial stream, and the process repeats itself, culminating in a secondary master
rill serving as an incipient tributary. Each smaller tributary evolves in a sim-
ilar way until the network of streams takes form.
Chapters 164
Figure 5.9. The network pattern develops by repeated division of single channel seg-
Development of bifurcation angles
ments into two branches, a process known as bifurcation. Schumm (1956)
(A) The original angle is
One branch
preserved. (B) suggests that the angle between the limbs of a bifurcated channel probably
becomes dominant (C) Angle evolves in one of three possible ways (fig. 5.9): (1) both limbs grow headward
decreases and branches merge
while preserving the original angle at their juncture; (2) one branch straightens
intoone channel; occurs on steep
slopes (From Schumm 1956 its course and becomes dominant; or (3) the angle on steep slopes progressively
Courtesy of The Geological decreases until the branches reunite into a single channel. Any or all of these
Society of America)
procedures might be found in the evolution of a network, constrained only by
the fundamental erosive controls and the geologic framework. Divides be-
tween adjacent basins are predetermined by the extent to which streams can
expand headward. Because critical length varies with resistance, the areal ex-
tent of the uneroded uplands partially reflects the geology and its history.
Within the basin itself, smaller interfluves may be present where cross-grading
has not operated. These small areas parallel the main stream and preserve the
slope of the original surface.
Basin Morphometry
One of Horton's greatest contributions was to demonstrate that stream net-
works have a distinct fabric, called the drainage composition, in which the
relationship between streams of different magnitude can be expressed in
mathematical terms. Each stream within a basin is assigned to a particular
order indicating its relative importance in the network, the lowest order streams
being the most minor tributaries and the highest order, the main trunk river.
Figure 5.10 shows several methods of ordering streams. Horton's cum-

bersome method was refined by Strahler (1952a) so that stream segments
rather than entire streams become the ordered units. As the figure shows, a
segment with no tributaries is designated as a first-order stream. Where two
first-order segments join they form a second-order segment; two second-order
segments a third-order segment, and so forth. Any segment may be joined by
a channel of a lower order without an increase in its order; i.e., third-order
segments may have an infinite number of second- or first-order tributaries.
Only where two segments of equal magnitude join is an increase in order re-
quired. The apparent inconsistency in Strahler's method of not accounting for
all tributaries is removed in the network analysis conceived by Shreve (1966a.
1967). He considers streams as links within the network, with the magnitude
Shreve(1967) of each link representing the sum of the link numbers of all tributaries that
feed That is, networks in which the downstream segments are of the same
it.
Figure 5.10.
magnitude have equal numbers of links within their basins. Shreve's link system
Methods of ordering streams
within a drainage basin gives a number that at any point within the basin is equal to the number of
first-order streams upstream from that point.
-
Table 5.2 Common morphometric relationships
Linear Morphometry
Stream number in each order (N ) N = Rb s °

R ~ s
1
Total stream numbers in basin (N) N= b

Rb - 1
°~'
Average stream length L = L,RL 1
- I lP ~ 1 1 where
wher
Total stream length L = L,R b s 1
I I
V u—1 Ju = RL
Bifurcation ratio R = N /No+^

Length ratio RL = I o/I +,
Length
y of overland flow C = — 1
2D
Areal Morphometry
Stream areas in each order ~A = A,Ra ~^

Length-area L = AA 06
1
Basin shape RF = —
Lb
Drainage density D= —
1L
F = —
N
Stream frequency s
C = —
1
Constant of channel maintenance
D
Relief Morphometry
Relief ratio Rn = H/L

Relative relief Rhp = H/p
Relative basin height / = h/H
Relative basin area x = a/A
Ruggedness number (Melton 1957) R = DH
Adapted from Strahler 1958 Courtesy of the Geological Sooely of America

s = order of master stream
o = any given stream order
H = basin relief
P = basin perimeter
Although the Strahler system of numbering was most commonly em-

ployed in earlier studies of basin morphometry, many researchers now find
Shreve's link ordering conducive to highly sophisticated analyses that are well
beyond the purpose of our treatment here (for example, see Smart and Wallis
1971; Abrahams 1980; Abrahams and Miller 1982). It is now generally rec-
ognized that in every basin a group of measurable properties exist that define
the linear, areal, and relief characteristics of the watershed (table 5.2). These
variables seem to correlate with stream orders, and various combinations of
the parameters obey statistical relationships that hold for a large number of
basins. Two general types of numbers have been used to describe basin mor-
phometry or network characteristics (Strahler 1957, 1964, 1968). Linear scale
measurements allow size comparison of topographic units.
Chapter 5 166
The parameters may include the length of streams of any order, the relief,
the length of basin perimeter, and many other measures. The second type of
measurement consists of dimensionless numbers, often derived as ratios of
length parameters, that permit comparisons of basins or networks. Length ra-
tios, bifurcation ratios, and relief-length ratios are common examples. Table
and most commonly used linear, areal, and relief equations,
5.2 gives the basic
but many more relationships not shown have been derived from these.
Linear Morphometric Relationships The establishment of stream ordering

led Horton to realize that certain linear parameters of the basin are propor-
tionately related to the stream order and that these could be expressed as basic
relationships of the drainage composition. Much of linear morphometry is a
function of the bifurcation ratio (/?*), which is defined as the ratio of the
number of streams of a given order to the number of the next higher order.
The primary use of the bifurcation ratio is to allow rapid estimates of the
number of streams of any given order and the total number of streams within
the basin. Although the ratio value will not be constant between each set of
adjacent orders, its variation from order to order will be small, and a mean
value can be used. Also, as Horton pointed out, the number of streams in the
second highest order is a good approximation of R b When
. geology is reason-
ably homogeneous throughout a basin, Rb values usually range from 3.0 to
5.0.
The length ratio (R L ) is similar in context to the bifurcation ratio; it is
the ratio of the average length of streams of a given order to those of the next
higher order. The length ratio can be used to determine the average length of
streams in an unmeasured given order (L ) and their total length (L ). The
combined length of all the streams in a given basin is simply the sum of the
lengths in each order. For most basin networks, stream lengths of different
orders plot as a straight line on semilogarithmic paper (fig. 5.1 1), as stream
numbers also do.
The relationships between stream order and the number and length of
segments in that order have been repeatedly verified and are now firmly es-
tablished (Schumm 1956; Chorley 1957; Morisawa 1962; Chose et al. 1967;
Selby 1967; and many others).
Areal Morphometric Relationships The equity among linear elements within

a drainage system suggests that areal components should also possess a con-
sistent morphometry, since dimensionally area is simply the product of linear
factors. The fundamental unit of areal elements is the area contained within
the basin of any given order (A ). It encompasses all the area that provides
runoff to streams of the given order, including all the areas of tributary basins
of a lower order as well as interfluve regions. Schumm (1956) demonstrated
(fig. 5.12) that basin areas, like stream numbers and lengths, are related to
stream order in a geometric series.
Although area by itself is an important independent variable (Murphey

et al. 1977), it has also been employed to manifest a variety of other param-
eters(shown in table 5.2), each of which has a particular significance in basin
geomorphology. One of the more important factors involving area is the
1,000,000
100,000
100,000
10,000 1000
o
o
o
1000 100
a w 10,000 _
CO
E
o Z
CO
CD
k.
CO
c
100 10 W CO
(D
o
O)
c 1000
CD
100
4 6 8 10 12 2 3
Order of streams Stream order no.
Figure 5.11. Figure 5.12.

Relation of stream order to the Relationship between stream
number and mean lengths of order and mean basin area in two
streams in the Susquehanna River drainage basins. (After Schumm
basin. (After Brush 1961) 1956. Courtesy of The Geological
Society of America)
drainage density (D), which is essentially the average length of streams per
unit area and as such reflects the spacing of the drainageways. The drainage
density is of interest because it is by the interaction be-
directly controlled
tween geology and climate. As these two factors differ from region to region,
wide variations in D can be expected (table 5.3). In general, resistant surface
materials or those with high infiltration capacities have widely spaced streams
and, consequently, low drainage densities. As resistance or surface perme-
ability decreases, runoff is usually removed in a greater number of closely
spaced channels, and D tends to be much higher. As a rule of thumb, where
geology and slope angles are the same, humid regions develop a thick vegetal
—
Chapter 5 168
Table 5.3 Drainage density in regions with different geology and climate.
Drainage Density 3 Climate Geology Area
3-4 Humid-temperate Resistant sandstone, Appalachian Plateau

flat-lying
8-16 Humid-temperate Nonresistant, flat-lying Central-Eastern U.S.

rocks
20-30 Dry summers Fractured and Southern California
subtropical. weathered igneous
seasonal and metamorphics
50-100 Semiarid Variable lithology and Rocky Mountains
structure
200-400 And-semiand Flat-lying, non-resistant Badlands. S. Dakota
sedimentary rocks
1100-1300 Humid-temperate Weak clays Northern
New Jersey
Data from Strahler 1968 and Schumm 1956

a ln miles/sq mi
mat that increases resistance and infiltration and thereby perpetuates a lower
drainage density than would be expected in more arid basins. Thus, drainage
density not only reflects the geologic framework, it may serve as a useful pa-
rameter in climatic geomorphology (Daniel 1981). Methods for rapid esti-
mation of drainage density have been devised (McCoy 1971; Mark 1974;
Richards 1979; Bauer 1980).
The density factor is also related to a parameter known as the texture
ratio (T):
T= N/P,
where TV is the number of crenulations on the most irregular contour and P is
the length of the basin perimeter. Ratios determined for many areas show the
mean texture ratio to be a useful descriptive number. A region with a coarse
texture has a mean value < 4, medium texture 4-10, and fine texture > 10;
figure 5.13shows a fine-textured region where drainage density is high.
Drainage density is related to the texture ratio as a simple power function
D = aTh ,
where a and b are constants. The texture ratio can therefore be used as a
substitute to indicate relative density values.
The drainage density not only reflects geology and climate, but it also has
been used as an independent variable in the framing of other morphometric
parameters. For example, the constant of channel maintenance and the length
of overland flow (table
5.2) both utilize a reciprocal relationship with density
to demonstrate the link between factors that control surface erosion and those
that describe the drainage net (Schumm 1956). The constant of channel main-
tenance indicates the minimum area required for the development and main-
tenance of a channel; that is, the ratio represents the amount of basin area
needed to maintain one linear unit of channel length. As Schumm points out
(1956, p. 607) this relationship requires that drainage networks develop in an
Figure 5.13.
Fine-textured topography with
high drainage density formed on
sedimentary rocks. Santa Fe and
Los Alamos counties, N.M
£#p£
^4mm
*4**W
'&*
..'
**»«
Chapter 5 170
orderly way because the meter by meter growth of a drainage system is pos-
sible only if sufficient area is available to maintain the expanding channels.
The ruggedness number (drainage density X basin relief) is another param-
eter that employs drainage density; it is useful in relating morphometry to
flood peak discharge (Patton and Baker 1976).
Relief Morphometric Relationships A third group of parameters shown in

table 5.2 is used to indicate the vertical dimensions of a drainage basin; it
includes factors of gradient and elevation. Like stream numbers, length, and
area, the average slope of stream segments in any order approximates a geo-
metric series in which the first term is the mean slope of first-order streams.
This relationship is reasonably valid as long as the geologic framework is ho-
mogeneous. Channel slopes and surface slopes are closely akin to the param-
eters for length. Horton suggested, for example, that the length of overland
flow as a function of only the drainage density is at best an approximation
because overland flow also depends on slope parameters.
As relief refers to elevation differences between two points, slopes that
connect the points are integral factors. The choice of reference points differs,
but the most useful relief parameters are the maximum relief (highest ele-
vation in the basin-lowest elevation in the basin) or the maximum basin relief
(highest elevation on the basin perimeter-the elevation at the mouth of the
trunk river). The relief ratio (Schumm 1956), the maximum basin relief di-
vided by the longest horizontal distance of the basin measured parallel to the
major stream, indicates the overall steepness of the basin.
A different relief study is the hypsometric analysis (Strahler 1952b), which
relates elevation and basin area. As figure 5.14 shows, the basin is assumed
to have vertical sides rising from a horizontal plane passing through the basin
mouth and under the entire basin. Essentially, a hypsometric analysis reveals
how much of the basin is found within cross-sectional segments bounded by
specified elevations. The relative height (y) is the ratio of the height (/») of a
given contour above the horizontal datum plane to the total relief (//). The
relative area (jc) equals the ratio a/A, where a is the area of the basin above
the given contour and A is The hypsometric curve
the total basin area.
(fig. 5. 1 4B) represents the between y and x and simply
plot of the relationship
indicates the distribution of mass above the datum. The form of the curve is
produced by the hypsometric integral (HI), which expresses, as a percentage,
the volume of the original basin that remains. In natural basins most HI values
range from 20 to 80 percent, the higher value indicating that large areas of
the original basin have not been altered into slopes. Low integral values simply
mean that much of the basin stands at low elevation relative to the area of the
original upland surface. One objection to the hypsometric analysis is the tedium
involved in determining the integral, but several alternative methods of cal-
culation (Chorley and Morley 1959; Haan and Johnson 1966; Pike and Wilson
1971) have removed much of this difficulty.
Summit plane Figure 5.14.

Ingredients of a hypsometric
analysis (A) Diagram showing
how dimensionless parameters
Land surface used in analysis are derived
parameters to
(B) Plot of the
produce the hypsometric curve.
(From Strahler 1952. Courtesy of
The Geological Society of
America)
Contour
Divide or perimeter
Summit
Mouth
Area a Area A
(entire basin) Proportion of total basin area
(A) (B)
Basin Evolution
Although morphometric values differ from basin to basin, each network still
obeys the same statistical relationships discussed above. Many authors have
suggested that morphometry reflects an adjustment of geomorphic variables
that is established under the constraints of the prevailing climate and geology
(Chorley 1962; Leopold and Langbein 1962; Strahler 1964; Doornkamp and
King 1971; Woldenberg 1969; and many others). Essentially this means that
once a network is established, the basinal characteristics can be defined by the
same quantitative terms at any time during the drainage growth. As the basins
and networks evolve, an equilibrium is eventually produced by the interplay
of climate and geology and maintained as a time-independent phenomenon.
Once the components within a basin become balanced, any changes of climate
or geology will be compensated for by adjustments of the basin parameters in
such a way that the relationships of drainage composition will be preserved.
As originally conceived, however, these relationships issued from well-devel-
oped stream systems, and the measurements needed to derive the equations
were made on topographic maps of these basins. Such an approach provides

no insight as to how quickly morphometric balance is attained or what changes
in its character occur as the basin ages. It seems appropriate, therefore, to
consider the influence of time on the morphometry of a basin.
Some studies have provided a glimpse into the question of how rapidly
morphometry is established and what changes occur in its nature as the basin
evolves. There seems to be little doubt that a quantitatively balanced drainage
net forms rapidly in erodible material. This was clearly demonstrated by
Schumm (1956) in the Perth Amboy (N.J.) badlands, by Morisawa (1964)
Chapter 5 172
on the uplifted floor of Hebgen Lake (Montana), and by Kirkby and Kirkby
(1969) on the raised beach and harbor floor around Montague Island (Alaska).
These studies showed that a balanced drainage composition had evolved in a
period no longer than a few years (in the Alaska case, only a few days). Cer-
tainly, the time needed for drainage development to occur in erodible material
is However, the amount of time needed
insignificant in terms of geological time.
in absolute terms probably varies according to the resistance of the material,
the climate, and the initial slope angles. Unfortunately, examples of drainage
establishment are documented only in areas where the least resistant mate-
rials underlie the system. Precisely how long it takes to form a balanced net-
work in regions underlain by resistant crystalline rocks is rather conjectural.

Drainage evolution has also been examined experimentally. Parker (1976,
discussed in Schumm 1977), using the Rainfall-Erosion Facility at Colorado
State University, found that patterns grow by headward extension until dis-
section reaches the watershed margin. Drainage density increases during this
phase, especially along the outside of the expanding drainage. Further evo-
lution, however, revealed that density achieves an equilibrium value main-
tained because tributaries near the center of the basin are lost while new
tributariesform and extend near the watershed margin. The total drainage
density and the mode of growth depended on the slope of the basin. The basin
with a lower slope had a higher density because tributaries formed inside the
basin during the early extension growth, while in the steeper basin, streams
extended to the margins before interior links were added. In general, Parker's
work showed that the network will develop as many streams and as much length
as is needed to efficiently drain the basin. However, during the period of growth,
drainage density may change, and in different parts of the basin, the change
may be in opposite directions.
Once basin elements attain a statistical balance, further changes in mor-
phometry are usually revealed drainage texture, or hypsometry
in the shape,
of the basin. The any basin presumably

area limits of are determined by the
hydrophysical controls denoted by Horton and by the competition for space
between adjacent basins. During the period of expansion to its peripheral limits,
however, each parameter should change in such a manner as to maintain the
original quantitative relationships among the factors. Hack (1957) showed
that for a large number of basins, the stream length and basin area are related
by the simple power function
L= 1.4 A 0.6
where L is the distance from any locality on a stream to the divide at the head
of the longest segment above the given locality, and A is the basin area above
the given locality. Hack noted that to preserve the original geometric balance,
each variable must change at the same rate and the exponent in the equation
should be 0.5. The larger exponent he observed requires that basins become
more narrow and elongate as they grow. Elongation is very striking in the
growth of parallel drainage patterns, as exemplified in the postglacial history

of theOntonagon region of northern Michigan (Hack 1965). There the net-
work formed as the Ontonagon Plain was exposed during the destruction of
Figure 5.15.
Growth of two parallel drainage
patterns in Ontonagon region,
the
northern Michigan. (After Hack
1965)
Streams between Mineral River Little Cranberry River

and Cranberry River
Figure 5.16.
Relation of stream order to stream
length and number in drainage
basins of the Ontonagon Plain,
Mich. A, Streams between Mineral
and Cranberry rivers; data from
maps; stream orders not known.
Cranberry River.
B. Little
C, Weigel Creek. D, Mill Creek.
(From Hack 1965)
Stream order (indicated by number)
glacial Lake Duluth. Initial stream development followed parallel grooves that
extended down the plain to the lake margin (fig. 5.15), and although this ac-
centuated the basin elongation, figure 5.16 demonstrates that the lengths and
areas of every order follow the 0.6 power function recognized elsewhere.
The same tendency has been noted in arid basins (Miller 1958). Thus,
the trend toward elongation has probably been noted in enough widely diver-
gent settings to suggest that it is an inherent property in the growth of drainage
basins (Jarvis and Sham 1981). Some studies, however, suggest that as small
basins evolve into large ones, they reach a stage where they widen as fast as
they elongate. The exponent then will reach a value that approximates 0.5
(Mueller 1972; Shreve 1974; Moon 1980).
Chapters 174
Part of the problem of shape variation with time is similar to the exper-
imental observations of density discussed above. Both of these factors may be
associated with the suggestion by Schumm (1956) that the region of most
intense erosion migrates with time, being near the mouth during the early
phase of basin development and near the headward margin in later stages.
These observations demonstrate that various parts of a large basin may be

evolving at different rates. The drainage texture also seems to vary with time.
As demonstrated by Ruhe (1952), both drainage density and stream fre-
quency increase systematically in areas underlain by glacial deposits of pro-
gressively increasing age. The rate of the textural change is not constant; it
probably was greatest during the first 20,000 years and then decelerated. The
marked transition in the rate of textural evolution perhaps represents the time
at which complete equilibrium was established and the basins were filled with
as many streams as possible (Leopold et al. 1964).
It is interesting to note that throughout the period of growth in Ruhe's
study, the channel lengths and numbers seem to obey Horton's geometric laws,
suggesting that texture may be a surrogate for time in the analysis of basin
evolution. This propositionwas given added credence when Melton (1958)
found stream frequency (F) and drainage density (D) to be related by a simple
equation
F = 0.694 D 1
.
In deriving this equation, Melton used as his sample 156 basins with widely
divergent geology, erosional histories, and presumably, ages. Since all stages
of drainage development were thrown into the statistical pot, the significant
empirical relationship between the variables most likely reflects a general trend
followed by basins as they grow. In fact, it was suggested that the dimension-
less ratio F/D 2 , called the relative density, should indicate how completely the
stream network fills the basin (Melton 1958). High relative density values
suggest that stream lengths are short and the basin outline is not yet com-
pletely filled with the stream network. As the drainage evolves and expands
into each basin niche, the ratio decreases until equilibrium is established.
It has been commonly accepted that Melton's growth law is an example
in which spatial parameters can be substituted for time. However, the uni-
versal applicability of the growth equation has been questioned. Abrahams
(1972) suggested that space and time are interchangeable in morphometry
only if the basins analyzed are environmentally similar and of the same order.
Further studies, however, did not completely substantiate this premise (Wil-
cock 1975), and a reasonable argument could be made that all factors — linear,
areal, and relief — are probably involved in the statistical relationships af-
fected by time.
Even though it may be necessary to include all types of factors in the
derivation of growth laws, it is clear that alterations of basin morphometry do
occur with time. These changes do not violate the steady-state concept because
parameters vary with one another in a systematic way. It is tempting to as-
sume, therefore, that a well-balanced network with discernible morphometry
Figure 5.17.
Gravels of Luther Time
\y/\ Map of terrace deposits in a small
H Gravels of Burnett Ranch Time piedmont area of the Beartooth
Mountains, Mont. Major stream
I Gravels ol Spring Creek Time piracy occurred near A
between deposition of Burnett
m Pre-Wisconsin Gravels of
the East Rosebud River Ranch gravels and deposition of
Mesozoic and Tertiary Luther gravels. (From Ritter 1972)
ot the Bighorn Basin
Paleozoic Rocks at the

Beartooth Mtn. Front
IPrecambnan Rocks ol the

Beartooth Mtns.
Glacial Deposits
evolved in an orderly manner from infancy to its present state. Realistically,

such an assumption may be totally erroneous because basin morphometry may
tell us little, if anything, about basin history.
A cogent example of this is the drainage system of Volney Creek, a small
basin in southern Montana within the watershed of the Yellowstone River.
Volney Creek heads in the piedmont region of the Beartooth Mountains, rising
approximately 3 km
from the mountain front. Its basin, covering an area of
60 km 2 is underlain entirely by Mesozoic and Tertiary clastic sedimentary
,
rocks. The basin exhibits all the normal morphometric characteristics of nearby
piedmont basins, and it is presumed to be in equilibrium because its mor-
phometry is balanced statistically as discussed above. The geomorphic history
of the basin, however, shows that its evolution was anything but orderly (Ritter
1972). Two terraces standing well above the present level of Volney Creek are
capped by gravel containing crystalline clasts that had to be derived from the
Beartooth Mountains; in fact, upstream tracing of the terraces demonstrates
that they cross through the present basin divide and continue mountainward
in the neighboring valley of West Red Lodge Creek (see map, fig. 5.17). The
significance is that prior to early Wisconsinan glaciation (Burnett Ranch time
in fig. 5.17), the valley now occupied by Volney Creek was the drainage av-
enue master stream of the area. The crystalline-rich gravel capping
for the
the terraces indicates that the paleoriver drained from the mountains, and its
basinal area must have been vastly greater than that of the modern Volney
Chapters 176
Creek. Immediately preceding the glaciation. the river was diverted into its
present position in the valley of West Red Lodge Creek, leaving the Volney
segment abandoned was occupied by the very small modern river.
until it
common in piedmont regions and

Diversions of the Volney Creek type are
may be important phenomena in the expansion of any drainage system, es-
pecially in the early stages (Howard 1971). Such changes, nonetheless, are
catastrophic events in basin development because they drastically alter basin
properties such as area, relief, and stream length. The internal adjustment to
changes spurred by piracy must occur rapidly, for most basins possess a bal-
anced morphometry even though such spasmodic events must be common-
place.
Basin Hydrology It is important for planning purposes to estimate how much water exists in a
drainage basin and whether it is available for use. Hydrologists usually em-
ploy a concept known as the water balance or hydrologic budget to make such
estimates. The water balance simply refers to the balance that must exist be-
tween all water entering the basin (input), all water leaving the basin (output),
and changes in the amount of water being stored. The major types of input
are rainfall and snow, and the major outputs are in the form of evapotrans-
piration and streamflow. Water is stored as soil moisture and groundwater, and
changes in these values actually represent losses or gains of available water.
For example, a positive change in groundwater storage indicates that the un-
derground reservoir is being recharged, but in the budget it represents a loss
because the availability of the water is being lowered. On the other hand,
groundwater runoff (base flow) represents an input because water availability
is increased as it is released from storage.
Underground Water
The increased demand for water that accompanied population growth, indus-
trial expansion, and extended irrigation in the United States has brought with
it a marked increase in the utilization of groundwater. The volume of water
contained in cracks and pores in the Earth's underground reservoir is enor-
mous, being estimated at almost 8 million km-
1
in the outer 5 km of the crust
(Todd 1970). Unfortunately, this vast resource is not evenly distributed, and
some regions are blessed with abundant groundwater while others are seri-
ously deficient. Ironically, many of the most rapidly expanding areas of the
United States are in regions with low reserves of surface and groundwater.
The lack of available water, coupled with a growing demand, poses a real chal-
lenge to geologists and requires that we continue to expand our knowledge
concerning the distribution, movement, and utilization of groundwater.
Groundwater geology is a separate discipline in itself: however, because it is
an important part of basin hydrology, we must briefly examine how ground-

water systems work.
Zone name . j^ûi,.^^^^ ,, Water name

Figure 5.18.
Soil water Soil water
J tj J b Schematic diagram of zones and
water types in the groundwater
profile (Not to scale)
Aeration Vadose water
Capillary fringe
ft'' Water table Capillary water
x
Phreatic Phreatic water
The Groundwater Profile Groundwater in porous and permeable rocks or

unconsolidated debris usually has a rather distinct distribution that can be
visualized as a vertical zonation known as the groundwater profile, shown in
figure 5.18. Below the zone of soil moisture, water moves downward under
the influence of gravity into and through the zone of aeration. In this zone,
pore spaces are partly occupied by water, called vadose water, and partly by
air that is physically connected to the atmosphere. At some lower depth all
the pore spaces are occupied by water, marking the beginning of a thin, sat-
urated zone, the capillary fringe, where water, held in tension, will not drain
freely into a well. Beneath the capillary fringe the spaces are also filled with
water in a zone known as the phreatic zone. The distinction between these two
saturated zones is that phreatic water will drain freely into a well because the
hydrostatic pressure in the phreatic zone is greater than atmospheric pressure.
The water table marks the top of the phreatic zone and represents the level
at which the hydrostatic pressure is equal to the atmospheric pressure.
Movement of Groundwater After descending to the phreatic zone, ground-

water does not stagnate but is capable of further movement. Unlike flow in
the zone of aeration, however, movement in the phreatic zone is not entirely
controlled by gravity. Each unit volume of water contained in the phreatic
zone possesses a certain amount of potential energy, called its potential or

head. The amount of potential from droplet to droplet, being dependent
varies
on each drop's pressure and elevation above some datum. When pressure and
elevation are known, the potential for each unit volume of water can be cal-
culated, and water particles having the same potential can be contoured along
surfaces known as equipotential surfaces (fig. 5.19). Although some diffusion
occurs, groundwater particles move along paths that are perpendicular to the
equipotential surfaces.
Chapters 178
Figure 5.19.
Movement of groundwater Ground surface
— ==
90 Equipotential surfaces
Flow lines
Ground surface
according to distribution of
potential in the underground Water table Water table
system Water moves from high to
low potential and perpendicular to
the equipotential surfaces.
Well =1
Figure 5.20.
Water table and loss in head as Well =2
water moves from well 1 to well 2
in unconfined aquifer.
^ Water Table
The movement of groundwater is a mechanical process whereby some of

the initial potential energy of the water is lost to friction generated as the
water moves. It follows that water moving from one point to another must
have more potential energy at the beginning of the transport route than at the
end. Thus, combining the above, groundwater always moves according to the
following rules: (1) it moves from zones of higher potential towards zones of
lower potential, and (2) it flows perpendicular to the equipotential surfaces
(Hubbert 1940).
The groundwater flow is directly proportional
velocity and discharge of
water moves from one point to
to the loss of potential (head) that occurs as
another (fig. 5.20). This was demonstrated by Darcy in his experiments on
flow through permeable material and led to his law of groundwater flow ex-
pressed as
V= K ^' ~ "2
'
where V is velocity, h is the head, L is the distance between points 1 and 2.
and A" is a constant of proportionality representing the permeability (or the

hydraulic conductivity) of the medium. In figure 5.20. the velocity of flow can
be calculated from the difference in hydrostatic level between two wells
1 79 The Drainage Basin — Development, Morphometry, and Hydrology
Recharge
Figure 5.21.
Confined aquifer and piezometric
surface
Artesian Well
Piezometric
surface
Impermeable rocks | Confined aquifer

(Aquitards)
(hi — h 2) if the permeability of the material is known. Discharge can also be

calculated by adding the cross-sectional area to the equation so that
Q = KIA
where
:re Q is discharge,
dis A is the cross-section area of an aquifer (w X d), and
/, called the hydraulic gradient, equals

hi-h 2
L
As mentioned above, the top of the phreatic zone is an important hy-
drologic feature known as the water table (fig. 5.18). Because hydrostatic
pressure everywhere along the surface of the water table is equal to atmo-
spheric pressure, the potential of water there is completely a function of ele-
vation, and water at that surface will always move from higher to lower
elevations. This partially explains why the water table is generally a mirror
image of surface topography. Where surface water and groundwater systems
are physically connected, the levels of rivers, lakes, swamps, etc., are merely
surface extensions of the underground water table (fig. 5.19).
Aquifers, Wells, and Utilization Problems Aquifers are lithologic bodies that
store and transmit water in economic amounts. The most common aquifer is
called an unconfined aquifer because it is open to the atmosphere and its hy-
drostatic level (the level at which water stands in an open hole) is within the
water-bearing unit itself. The hydrostatic level in an unconfined aquifer is the
water table. In some other aquifers the water is held in a porous and perme-
able unit that is not connected vertically to the atmosphere but is overlain and
underlain by impermeable layers called aquitards. Water contained in these
confined aquifers (fig. 5.21) will rise above the top of the aquifer when it is
penetrated by an open hole. The level to which water will rise is called the
piezometric surface, and its height above the aquifer itself depends on the
difference in potential at the point where precipitation enters the aquifer (re-
charge zone) and the position of the hole (fig. 5.21). Sometimes the piezo-
metric level is above the elevation of the ground surface, and water will flow
freely out of the hole without pumping as artesian flow.
Chapter 5 180
Figure 5.22. Radius of influence 72 hrs.

Successive cones of depression
caused by drawdown of water
table during 72-hour pumping of
unconfined aquifer
Water table
The development of an aquifer for water supply requires wells; the larger
the demand, the larger and more numerous the wells. As a well is pumped,
the hydrostatic level (water table or piezometric surface) surrounding the well
is molded into an inverted cone known as a cone of depression (fig. 5.22). The
cone develops because the release of water (or of pressure in a confined aquifer)
is greatest near the well, causing pronounced lowering of the hydrostatic level,
called drawdown, adjacent to the well. The effect of pumping decreases away
from the well, and so the drawdown is less at the perimeter of the cone. In the
initial phase of pumping, the rate of drawdown is high, but as pumping con-
tinues the rate gradually decreases until the cone attains a nearly constant
form. The dimensions of the quasi-equilibrium cone depend on the rate at
which the well is pumped and the hydrologic properties of the aquifer.
Utilizing groundwater can result in a number of environmental problems

if the system is not carefully studied before development. For example, ex-
cessive drawdown can occur locally if wells are placed too close to one another
and the radius of influence (maximum diameter of the cone of depression) of
adjacent wells overlaps. This produces abnormally high drawdown in the zone
of overlapping because the actual lowering of the water level is the sum total
of drawdown produced by all the interfering wells. On a regional scale, most
problems arise when more water is pumped from the aquifer over a period of
years than is returned to the aquifer by natural or artificial recharge, a prac-
tice known as overdraft. In such a case the aquifer is actually being mined of
its water, and on a long-term basis the normal hydrostatic level may be dras-
tically lowered.
The effects of continued overdraft differ, but two types of responses will
demonstrate the problems that can result from misuse of the groundwater
system. First, in confined aquifers, the drawdown of the piezometric surface
reflects the decrease of pressure within the aquifer. Because the pressure is
lowered, water from the overlying aquitard seeps downward into the aquifer.
As the aquitard drains, the normal load exerted by the weight of the overlying
rocks compacts the aquitard and decreases its thickness. Ultimately, the pro-
cess culminates in measurable subsidence of the ground surface. Some areas
(Mexico City, Las Vegas, Central Valley, California, Houston-Galveston area)
have experienced 1 to 5 meters of overdraft subsidence, creating a variet> of
annoying and hazardous conditions such as cracking of buildings, strain on
buried pipelines, and destruction of well casings.
Ground surface
Figure 5.23.
Relationship of salty and fresh
Water table groundwater in an aquifer of a
coastal region Lowering of water
table by overdraft requires a 40-
Sea level
boundary between
fold rise in the
the fresh and salt water and
increases the possibility of
pollution of the aquifer by salt
water.
A second major effect of overdraft usually occurs in coastal regions where

the aquifer is physically connected to the ocean. There two fluids (ocean water
and fresh water) having different densities are separated by a sharp boundary,
as diagrammed in figure 5.23. The location of the interface depends on the
hydrodynamic balance between fresh water (density = 1 .000 g/cm 3 ) and salt
water (density = 1.025 g/cm 3 ). In general, the depth below sea level to the
saltwater boundary is about 40 times the height of the water table above sea
level. In such a situation, continued overdraft can cause pollution of the aquifer
because minor lowering of the water table necessitates a much greater rise of
the saltwater-freshwater interface, a phenomenon known as saltwater intru-
sion. For example, a 2 m drawdown of the water table requires a concomitant
80 m rise of the salt water, and any well extending to a depth greater than the
new interface level will be polluted with nonpotable water.
Many cities along the coasts of California, Texas, Florida, New York, and
New Jersey have been affected by saltwater intrusion. The phenomenon, how-
ever, can function wherever two same
fluids of different density exist within the
aquifer. For example, the water supply of Las Vegas, Nevada, is in jeopardy
from pollution by high-magnesium groundwater located about 24 km south
of the city. The impending intrusion is due to the large overdraft from the
aquifer beneath Las Vegas.
Surface Water
The major loss of water from most drainage basins occurs as stream discharge,
that is, the volume of water passing a given channel cross section during a
specified time interval, or
Q= wdv,
where Q is discharge in ft 3 /sec(cfs) or m 3 /sec(cms), w is width, d is depth,

and v is velocity. Measurement of discharge is a relatively simple procedure
whereby the total width of the channel is divided into evenly spaced segments.
The depth and velocity are measured in each compartment, and total dis-
charge is determined by summing the discharges of all the subsections. The
difficult measurement to obtain is velocity. Surface velocity is not a good es-
timate of the mean velocity, and so measurements must be made within the
current. Many devices have been developed to measure velocity, but probably
Chapters 182
Figure 5.24.
Photo of Price Current Meter
the most widely used is the Price current meter. This instrument, designed by
W. G. Price in 1882, has a group of conical cups mounted on a vertical rod
that is rotated by the force of the water striking the cups (fig. 5.24). As the
shaft spins, it periodically closes an electrical circuit, the moment of closure
being recognized by the meter operator as a clicking sound in earphones con-
nected to the circuitry. The number of axial revolutions per unit time is easily
converted into velocity values.
To be of any scientific value, discharge must be measured repeatedly at
the same locality, giving hydrologists a better understanding of how flow varies
with time. In the United States these sampling localities, called gaging sta-
tions, have been in operation since the late 1800s; therefore, a wealth of flow
data for a large number of streams
is available from the U.S. Geological Surve\
and many state water surveys. At most stations, discharge is measured every
1 5 minutes and the average of these values published as a mean daily dis-
charge. The mean annual discharge is the average of the daily values over the
entire period of record, presuming that the station has been maintained for
-
Q (cms)
Figure 5.25.
10 12
Rating curve for low flow, Rock
Creek near Red Lodge, Mont
300 400 700

Q (cfs)
longer than one year. Actually, discharge measured as described above

is not
because the procedure is it is estimated from a
too time-consuming. Instead
rating curve (fig. 5.25), which relates a wide spectrum of discharge values to
the elevation of the river above some datum. Once the rating curve is con-
structed, the height of the river above the datum (called the stage or gage
height) is the only variable observed directly, and its value is used to predict
the discharge.
Geomorphologists are interested in the frequency and magnitude of flow
events because each has a decided bearing on how watershed systems work.
The frequency of a given discharge is compiled into a flow duration curve
(fig. 5.26), which relates any discharge value to the percentage of time that
it is equaled or exceeded. At any station, then, the lowest daily discharge in
the period of record will be equaled or exceeded 100 percent of the time. The
largest flow will be equaled only once out of the entire number of days in the
sample, giving it a percentage value slightly greater than zero.

Another common approach to finding the frequency-magnitude relation-
ship is to consider only the peak discharge during each year (annual series)
or only the discharges above some predetermined value {partial duration se-
ries). The statistical samples in the series approaches are much smaller than
in analyses utilizing daily records; however, they are useful in studies of major
Chapter 5 184
100,000 _
Figure 5.26.
Flow duration curve for the
Powder River near Arvada, Wyo.
1917-1950. (From Leopold and
Maddock 1953)
10,000
1,000
Mean annual discharge =
430 cfs (12.2 cms)
Median discharge =
,160 cfs (4.5 cms)
O)
«
o
20 40 60 80 100
Percent of time flow equalled or exceeded discharge indicated
flow events. The annual discharges shown in table 5.4 have been ranked ac-
cording to their magnitude during the years of record, and a recurrence in-
terval for each flow is shown. The recurrence interval, simply the average time
between two flow events of equal or larger magnitude, is calculated as
n + 1
R
m
where R is the recurrence interval in years, n is the total number of discharge
values in the sample, and m is the rank of a given flow.
A plot of discharge and recurrence intervals on probability graph paper
(fig. 5.27) allows hydrologists to estimate the magnitude of a flood to be ex-
pected within a specified interval of time. Various types of probability graphs
have been used in these analyses (lognormal. Gumbel Type I. Gumbel Type
III, Pearson Type III), but no one type is inherently better than the others
(see discussion in Dunne and Leopold 1978 and Benson 1971).
On Rock Creek (Mont.), as figure 5.27 shows, the flow should equal or
exceed 81 cms (2860 cfs) once during each 25-year interval. As rivers unfor-
tunately do not understand statistical theory, there is no reason to believe that
25-year floods will be evenly distributed over time. In the next 50 years there
may be two 25-year floods in successive years, or floods may occur in the first
and last years of the period, or a flow equaling or exceeding the discharge of
185 The Drainage Basin — Development, Morphometry, and Hydrology
Table 5.4 Annual flood series data, 1932-1963, for Rock Creek near Red Lodge,
Mont. Each maximum annual flood is ranked, the highest being
ranked 1. Recurrence intervals have been calculated.
Maximum Flood Rank Recurrence

Year cfs cms Magnitude Interval
1932 935 265 22 1 45

1934 533 15.1 31 1.03
1935 1490 42 2 9 356
1936 1240 35.1 15 213
1937 1930 54.6 5 64
1938 991 28.0 21 1 52
1939 661 187 29 1 10
1940 673 19.0 28 1 14
1941 780 22 1 24 1 33
1942 1840 52.1 6 5.33
1943 2010 569 3 106
1944 1630 45 1 7 457
1945 1990 56.3 4 8
1946 774 21 9 25 1 28
1947 846 23.9 23 1.39
1948 1070 30.3 19 1.68
1949 1190 33.7 17 1 88
1950 1100 31.1 18 1 78
1951 1460 41 3 10 32
1952 2590 733 2 16
1953 1300 36.8 13 246
1954 1570 444 8 40
1955 578 164 30 1.06
1956 1430 40.5 11 290
1957 3110 88.0 1 32
1958 1250 35.4 14 2.29
1959 1200 340 16 20
1960 680 19.2 27 1 19
1961 751 21 3 26 1 23
1962 1030 29 1 20 1.60
1963 1350 382 12 2 67
a 25-year flood may not happen at all. However, the probability that a flow
of a particular magnitude will occur in any year is the reciprocal of the re-
currence interval
where is the probability of flow being equaled or exceeded in any one year
P
and R the recurrence interval. Thus, a 25-year flood has a 4 percent chance
is
of occurring in any given year; a 10-year flood has a 10 percent chance, and
so on. Exactly which year it will happen is simply unpredictable.
Even though it is impossible to predict when a given flow will occur, mag-
nitude-frequency analyses have practical value in river management, espe-
cially for lower frequency floods such as the 20-year (q w ), 10-year (q 10 ), or
5-year (<? 5 ) floods. Using an annual series, the mean annual flood, which is
1
Chapter 5 1 86
Figure 5.27.
Flood frequency curve for Rock
Creek near Red Lodge, Mont
-
3000
• /
-
as
E 2000
•J
c
c
<
-
/
1000 •
•
• >•
mi i i I mi i i 1 i 1 l llllillll i l 1 nil llll
1.01 1.05 1.2 1.5 2 I
1.1
Recurrence interval (years)
the arithmetic mean of all the maximum yearly discharges in the sample, is
the flow that should recur once every 2.33 years ((72.33) • In °ur Rock Creek
example the mean annual flood has a discharge of 1290 cfs and plots at a
recurrence interval of approximately 2.33 on our probability curve (fig. 5.27).
When longer time intervals are involved in watershed management, it may
be important to know the probability of a particular flood's occurring during
the projected life of a design structure. This is normally estimated by the prob-
ability equation
("-#
where q is the probability of a flood with a recurrence interval (7) occurring
in the specified number of years («). Thus, the chance of a 50-year flood's
occurring in the next 50 years is 63 percent (Costa and Baker 1981).
The U.S. Geological Survey publishes extensive flood- frequency infor-
mation on a state and regional basis. Techniques for deriving flood frequenc\
curves from gage data or for estimating maximum probable floods at any site
have been described in detail (Dalrymple 1960).

Backwater flow Pecos River

Figure 5.28.
Schematic of on- and off-lap
sequences and peak flood stage
in a tributary valley for the 1954
and 1974 floods on the Pecos

River, Texas. Sections in the
proximal region (area 2) contain
both floods, while distal regions
(area 1) farther up the tributary
record only the larger 1954 flood,
Paleostage reconstructions are
based on the elevation of the
most distal sediments of each
flood unit
Paleofloods Perhaps the single most

difficult problem faced by hydrologic
planners is magnitude of the largest flood that can be expected
to estimate the
to occur in any given basin. Such estimates require extension of the flood-
frequency curve beyond the limit provided by measured flow events. This is
very risky business because critical decisions about the type and cost of flood
protection may be based on these estimates. Simple extension of the frequency
curve along the trend of the line requires questionable assumptions, and most
hydrologists agree that line extensions past twice the period of record are in-
valid (see Costa and Baker 1981). In some cases historical observations about
stages attained in floods that occurred prior to installation of gaging stations
may be used to extend the curve (Benson 1950). In the United States, how-
ever, such observations would only extend the curve several hundred years at
best. In addition, the accuracy of observations found in historical records are
suspect, and the method may not be any more valid than simple line extension.
One of the more promising techniques to extend flood-frequency curves
involves the use of Holocene stratigraphy. Essentially, the method utilizes the
positions of overbank deposits in the sequence of floodplain sediments as in-

dicators of rare flow events. Techniques used to make the analyses are thor-
oughly presented in Costa (1978), Patton et al. (1979), and Baker et al. (1983).
The most common stratigraphic approach is to utilize slack water sedi-
ment deposited by narrow bedrock valleys. In these
rivers confined in relatively
settings, large stage increases result from small increases in discharge, and
the valley cross sections are not subject to major change during floods. In high-
flow conditions, water is backflooded into tributary mouths, shallow caves, or
protected areas downstream from bedrock spurs jutting into the valley. The
low-velocity slack water entering these areas deposits fine-grained silts and
sands which contain organic matter that can be dated by radiocarbon anal-
yses, thereby providing the recurrence interval for the flood. Projection of the
highest level of the deposit into the main valley gives an estimate of the peak
flow depth (fig. 5.28); the discharge can then be estimated by using the slope-
area hydrologic method (for details, see Dalrymple and Benson 1967).
Chapter 5 188
Slack water deposits have been used to extend frequency curves over a
period of 2000 to 10,000 years; in addition, the technique has been applied in
widely diverse climatic zones (Moss and Kochel 1978; Patton et al. 1979; Patton
and Dibble 1 982; Kochel et al. 1 982; Baker et al. 1 983). The method, however,
is subject to certain assumptions that may lead to error. Nonetheless, com-
parison of results obtained by this method with those from conventional tech-
niques of curve extension shows the promise of the slack water analysis. For
example, the 1954 flood of the Pecos River in Texas was estimated by con-
ventional methods to have a recurrence interval ranging from 81 years to 10
million years. The slack water method estimate is 2000 years, a much more
reasonable value (Kochel et al. 1982).
Other stream deposits and approaches may be used to determine the age
and/or magnitude of past floods (Costa 1978, 1983). For example, overbank
gravel deposits associated with soils or organic-rich layers can also be used in
a manner similar to that employed in the slack water method (Costa 1974a;
Patton and Baker 1977).
Effect of Physical Basin Characteristics Physical characteristics of a basin

play a significant role in determining the magnitude of the flood peak. The
basin lag, for example, is the time needed for a unit mass of rain falling on
the basin to be discharged from the basin as streamflow. It is usually estimated
as the time interval between the centroid of rainfall and the peak of the hy-
drograph (fig. 5.3) and probably consists of two separate parts: (1) the time
involved in overland flow and (2) the channel-transit time. Lag time for any
basin is fairly consistent, with only minor variations in the parameter caused
by the position of the storm center relative to the gaging site. Obviously lag
should increase with the size of the basin, but comparisons of basins of equal
size show that lag time may be as much as three times greater in sluggish
streams than in those with short lag times, called flashy streams. Lag, there-
fore, must be determined by more than drainage area alone, and since it is
presumed to be influenced by the geomorphic framework, it should be related
to morphometry in some discernible way.
There is no question that the time elements of basin hydrology have a
monumental influence on the magnitude of peak discharge during floods. Take,
for example, the progression of the flood crest recorded at several gage stations
in the Susquehanna River basin during the flood in June 1972 (fig. 5.29). Most
of the precipitation spawned by hurricane Agnes entered the basin during the
period between June 19 and June 22. In the minor tributary Bald Eagle Creek,
the flow peaked at 143 cms (5050 cfs) on the night of June 22, indicating a
relatively quick response to the storm. In the larger Juniata River, the flood
crested about 12 hours later with a considerably higher discharge value of
3538 cms (125,000 cfs). Far downstream on the main Susquehanna River, the
flood peak did not occur until 12 hours after the Juniata peak, when discharge
rose to 30,564 cms (1,080,000 cfs). The fact that discharge peaked later on
the main stem, and at a significantly higher magnitude than in the tributaries,
shows the importance of timing in basin hydrology.
200
100
« 80 :
1 60-
40-
20-
Mapleton Depot, Pa.

10
23 24 25 26 23 24 25 26
Time (June 22-26, 1972) Time (June 22-26, 1972)
(A) (B)
40,000 ^^m
20,000 -
E 10,000 _
~ 8000
=
£ 6000 -
<2 4000
2000
Figure 5.29.
Progress of flood caused by
1000 hurricane Agnes in the
23 24 25 26 Susquehanna River basin, June
Time (June 22-26, 1972) 1972 Basins of increasing size:
(A) Bald
Eagle Creek; (B) Juniata
(C) River: (C) Susquehanna River.
As suggested above, lag time and peak discharge are positively related to
basin size. More important, however, is the fact that simple expansion in size
cannot totally explain the observed flow characteristics during a flood. Using
the same data as before and replotting the peak discharge as Q/km 2 , figure
5.30 reveals the interesting hydrologic property that discharge per unit area
is much higher for the smallest tributary than for the massive basin of the
main stem. Had the increase in discharge during the 1972 Susquehanna flood
been only a function of increased basin area, the peak flow at the Marietta
station should have been a simple product of the Q/km value of Bald Eagle
2
Creek times the drainage area above Marietta, or 1.25 cms/km 2 X 67,574
km 2 = 84,468 cms (2,978,454 cfs); but this is almost three times the actual
Chapter 5 190
Figure 5.30.
Discharge per unit area during
flood caused by hurricane Agnes
in the Susquehanna River basin,
June 1972 Note that the main
river has considerably less
discharge/area than tributaries
Mam stem also peaks later and
discharges flood water over a
longer time period, showing the
effect of storing water on
floodplains
23 24 25 26
Time (June 22-26, 1972)
measured value. Somewhere within the basin, a built-in flood control mech-
anism exists that not only holds down the magnitude of peak Q but simulta-
neously maintains abnormally high flow over a longer period of time. Most
lowering of the peak flow is due to the ability of floodplains to store or retard
the movement of larger volumes of water until the flood crest passes a given
channel locality. In addition, differences in flood-crest travel time on the nu-
merous tributaries in the basin and vagaries introduced by variable sources of
runoff may help to retard the downstream peak flow. Thus, a portion of runoff
never adds to the increasing peak flow downstream but shows up in the record
after the crest has passed as part of the recessional limb of the flood hydro-
graph.
Morphometric Relationships We now know that stream hydrology, as de-

fined by the discharge hydrograph and by time elements such as flood fre-
quency and lag, is significantly related to many components of basin and
network morphometry. The interdependence of morphometry and hydrology-
is statistically real but does not necessarily indicate a cause-and-effect rela-
tionship — one factor is not the cause of changes in the other. The high cor-
relation probably exists because both factors vary in a consistent way with the
same underlying climatic and geologic controls. In general, area and relief
factors are closely related to flow magnitude, and length elements to the timing
of hydrologic events. All morphometric types, however, are themselves so com-
plexly woven together that no one factor can be isolated as a completely in-
dependent variable. (Murphey et al. 1977).
Since basin area and peak discharge are highly correlative, we should ex-
pect that many other areal parameters would be similarly related to discharge.
In fact, every factor involving area differs in its success as a predictor of dis-
charge, but one parameter, drainage density, seems to have considerable value
as a gage of peak flow. In a study of 1 5 small basins in the southern and central 500
Appalachians and the Interior Low Plateau region, Carlston (1963) demon-
strated a very close relationship between drainage density and the mean an-
nual flood (fig. 5.31). Notably, the basins in his sample have wide variations
in relief, valley-side and channel slopes, and precipitation characteristics; yet
none of these factors seems to disrupt the flood magnitude-drainage density 10
relationship. Carlston suggests that the general capacity of a terrain to
100 J
infil- 9
11/3
• J •
trate precipitated water and transmit it through the underground system is
a/;b
the prime controlling factor of the density-mean annual flood relationship in 50 -
basins up to 260 km 2 in area. In larger basins, channel transit time plays the 13"/»»2
4 M2
dominant role in the flow character. The rate of base flow, found to be inversely
related to drainage density, also dependent on the terrain transmissibility.
V
Thus, as Horton suspected
is
earlier, high transmissibility (as evidenced by in-

e. h/.
filtration capacity)
sultant
spawns low drainage
low-magnitude peak flood. In contrast,
density, high base flow,
an impermeable surface
and a re-
will
O 10
ir
generate high drainage density in order to efficiently carry away the abundant
runoff; base flow will be low and peak discharge high.

Patton and Baker (1976) found that basin relief and drainage density are
the two variables that best distinguish areas of different flash-flood potential.
Since the ruggedness number is the dimensionless product of relief and drainage 1 Q 2 33 per square
density, this number should be a suitable guide to the potential for flash floods. / mile = 1.3 D
2
Indeed, Patton and Baker found that high-potential basins tend to have greater
ruggedness numbers than low-potential watersheds. It appears, therefore, that
1 5 10
drainage density, by itself or in combination with other morphometric vari-
ables, may be an important guide to how a basin will function hydrologically. Drainage density, D
Dingman (1978), however, cautions that the relationship between drainage

density and flow can be overriden by other effects in the basin such as flood- Figure 5.31.
plain or channel storage. In addition, where saturated overland flow is the Discharge (mean annual flood,
O2.33) controlled by drainage
major source of runoff, drainage density may tend not to be related to the density in 13 basins. (From
efficiency at which a basin is drained. Carlston 1963)
In spite of the apparent difficulties, the possible interrelationships be-
tween hydrology and morphometry are seemingly infinite, and even though
the parameters are so complexly related that equations will not explain all the
variability, the approach has some validity and should not be abandoned in
future research. The hydrogeomorphic approach is especially applicable in

determining regional flood hazards (Baker 1976).
In addition to being hydrologic entities, basins are also geographic compart- Basin Denudation
ments where sediment is manufactured, eroded, and deposited, and from which,
given sufficient time, the debris will ultimately be removed. The amount of
sediment leaving a basin can be readily converted into an estimate of lowering
of the basin surface, called denudation, which is usually expressed as a time
parameter or rate. Denudation seems to have no rigorous definition, but be-
cause it implies removal of basin material, it is commonly used as a synonym
for erosion. The two differ, however, because denudation considers only those
eroded products that are removed completely from the basin, assuming in that
Chapter 5 1 92
consideration that the sediment is derived in equal portions from all subareas
of the watershed. It therefore presupposes equal surface lowering over the en-
tire basin. Denudation rates tell us little about those erosive processes that
simply redistribute sediment within the basin; nor do they indicate that at any
given time some parts of the basin are probably aggrading rather than eroding.
Denudation, then, is the long-term sum of the overall erosive process, and even
though it is analogous to erosion it is not precisely the same. Furthermore,

recognize that the basin surface may not actually lower even with a high de-
nudation rate if active uplift of the basin is proceeding at a greater rate.
Estimates of modern denudation are usually based on measurements of
stream load made at gaging stations or on the volume of reservoir space lost
when sediment accumulates behind dam. All types of load (suspended, bed,
a
and dissolved) are included in the analyses at gage stations, which require that
the weight values of load be converted into volumetric terms. Once the volume
of sediment leaving the basin is determined, it is divided by the area of the
watershed above the gage station to provide the third (or vertical) dimension,
which represents the magnitude of surface lowering (for details see Ritter
1967). Rates are commonly expressed in inches or centimeters per 1000 years.
In most basins the amount of solid load is the predominant type of material
lost. Additionally, as we will see in the next chapter, the type and volume of
solid load has a monumental influence on river behavior. Thus, we must ex-
amine the mechanisms by which slopes are eroded and what factors affect the
mode and rate of the erosive processes.
Slope Erosion and Sediment Yield

A raindrop possesses a considerable amount of kinetic energy, derived from
its mass and the velocity it attains during its fall. Under the influence of gravity,
a raindrop accelerates until its force is equal to the frictional resistance of the
air, the speed at that point being the terminal velocity. As the distance needed
to attain this condition is very short, most rain strikes the surface at its ter-
minal velocity, although the absolute speed varies with wind, turbulence, drop
size, etc. In high-intensity rains, drops usually reach a maximum size of ap-
proximately 6 mm and a terminal velocity of about 9 m/sec. The impact of
such rain can directly displace into the air particles as large as 10 mm in di-
ameter and, by undermining downslope support, can indirectly set even larger
pebbles in motion. The amount of soil moved by splash depends on several
interrelated factors. First, the kinetic energy of raindrops is directly related
to splash movement (Kneale
1982). However, it is interesting to note that the
kinetic energy of raindrops sometimes varies in unexpected ways. For ex-
ample, Mosley (1982) found that rain passing through a forest canopy had
greater total kinetic energy than normal rainfall and that rainsplash was three
times greater under the canopy than in open areas. Second, the type of soil
being struck is extremely important in determining the magnitude of splash

movement. Free (1960), for example, found that splash loss varied as £°- 9 for
a silt loam soil and E XAb for a sand, where E is the kinetic energy. Over a five-
year period the total splash loss from the sandy surface was calculated at 1600
tons/acre, an amount three times greater than the loss from the loam, prob-
ably because the fine-grained soil had greater cohesion. Actually the manner
in which the soil particles aggregate (Luk 1979) and the dispersive properties
of the surface material (Yair et al. 1980; Rendell 1982) are more significant
controls than simple textural composition. Third, the rate and amount of splash
transport appear to be a function of slope angle (Savat 1981; Reeve 1982),
but the precise relationship is quite variable and not easily determined (Bryan
1979).
In addition to direct transportation, splash has several other erosion-in-
ducing effects on the soil. By detaching particles, it destroys the structure of
the soil and breaks apart resistant aggregates of clays. These physical pro-
cesses make the soil much more susceptible to erosion by surface flow. Fur-
thermore, as splash disperses the clays, they tend to form a fine-grained crust
as they settle back on the surface. This crust forms a semipermeable barrier
that reduces infiltration and promotes runoff, thereby increasing soil loss by
overland flow.
Wash Most natural slopes are too irregular to permit a uniform flow of water
over the entire surface; flow is deeper over depressions and shallower over flat
reaches or high spots. The variable depth of flow produces differences in the
eroding and transporting capabilities of the water so that, in detail, wash does
not imply that a regular sheet of debris is being carried continuously down
the slope surface. In areas where sheet flow might be possible, only fine-grained
particles can actually be moved efficiently, and those only if the surface has
been prepared for erosion by rainsplash or weathering processes that reduce
cohesion. In areas of concentrated flow, larger sediment can be moved, but the
ability to erode depends more on the hydraulic force of the water and less on
the condition of the surface.
When rainfall and flow become intense, small shallow channels may be
formed, channels that periodically shift their position so that in the long run
erosion is more or less even across the slope. In fine-grained soils, a set of well-
defined subparallel rills is usually formed. Rills vary in size with the erodibility
of the soil, but normally they are only several centimeters wide and deep.
Heaving and other processes can obliterate these tiny channels in periods be-
tween high rains, especially in highly seasonal climates where rain may be
lacking for months at a time. The periodic destruction of rills allows new chan-
nels to form an entirely different location and ensures less than equal low-
in
ering of the entire slope surface. Some rills escape this spasmodic destruction
by entrenching to greater depths, a difficult task that only a few of the largest
rills accomplish. These "master" rills become relatively permanent and even-
tually evolve into true rivers.
In soils that are sandy or coarser, the channels are usually braided because
the material, although easily eroded, is transported with difficulty. Sediment
commonly accumulates as temporary bars within the channel, and these sub-
divide the channel and the flow into a multitude of small passageways. Most
braided channels are wider (up to 5 m) and deeper (1-10 cm) than rills, but
they also change their position regularly because the bar deposits require a
continuously shifting channel environment.
Chapter 5 194
Sediment Yield (Soil Loss) Assuming that allfactors can be assessed within
reasonable limits of uncertainty, the soil loss by water erosion should follow
the Universal Soil Loss Equation
A = RK(LS)CP,
where A is average annual soil loss, R is the rainfall factor, K is the erodibility
factor, LS the slope length-steepness factor. C the cropping and management
factor, and P the conservation factor (for details see Smith and Wischmeier
1962). The precision of such an equation is questionable because in reality it
predicts total soil movement rather than total soil loss. Along a typical slope
surface, some zones may be experiencing active deposition while others are
being eroded. In addition, the equation sometimes underpredicts soil loss. Haigh
and Wallace (1982), for example, showed that ground losses from strip-mine
dumps in Illinois were twice as great as those estimated by the U.S.L.E. None-
theless, the analysis is widely used, and it probably can provide useful ballpark
data for predicting sediment loss from slopes.
Factors Affecting Sediment Yield A number of interrelated geologic, hy-

drologic, and topographic factors cause the magnitude of sediment yield
to vary widely from region to region. The most important of these are
(1) precipitation and vegetation, (2) basin size, (3) elevation and relief. (4) rock
type, and (5) human activity.
Precipitation and Vegetation Intuitively we would expect the amount of

sediment yielded from any basin to be related in some systematic fashion to
the amount of incoming precipitation. Actually the correlation between the
two is not so direct as we might hope, because precipitation is complexly in-
terrelated with other factors that influence its erosive capability. The amount
of runoff from any given precipitation, for example, varies w ith temperature.
Vegetation, a function of both precipitation and temperature, serves as a pro-
tective screen against the erosion of surface material. Precipitation, then, cannot
be considered as a completely independent variable in the realm of denuda-
tion, even though it may be a dominant factor.
In an important paper, Langbein and Schumm (1958) documented the
relationship between sediment yielded from basins averaging 3900 km 2 in area
and parameter derived by adjusting the magnitude
effective precipitation, a
of precipitation to values expected at a mean annual temperature of 10°C
(50° F). Under those conditions, they were able to show that as precipitation
rises from zero, the sediment yield increases rapidly to a maximum yield value
at about 30 cm of effective precipitation (see fig. 2.17). Any increase in pre-
cipitation above 30 cm promotes a decline in sediment yield because the den-
sityand type of vegetation begin to play an active role in protecting the slopes
from erosion. Vegetation generally begins to exert a control on erosion when
is between 8 percent and 60 percent, values typical in semi-
the ground cover
subhumid climates. Thus, as L. Wilson (1973) suggests, the Lang-
arid and
bein-Schumm curve may be valid for regions with a continental climate but
may not be applicable in other climatic regimes, especially nonseasonal types.
50.000
Figure 5.32.
•
Comparison of sediment yields
and drainage basin areas for all
10,000
major sediment-discharging rivers
(greater than 10 X 10 6 1 yr 1 ).
Open circles represent low-yield
Ganges/
•
Yellow* rivers draining Africaand the
i
• Brahma
1000 Eurasian Arctic. Smaller basins
• •
E have larger yields, although the
S^ Yangtze largest rivers(Amazon, Yangtze,
• Ganges/Brahmaputra, and Yellow)
X^ Amazon. all have greater loads than their
100 basin areas would predict.

ID
^^ •
E
a)
to
10
10 100 1000 10.000

Drainage basin area (10 3 km 2 )
In any case, the demonstration by Langbein and Schumm that the relationship
between precipitation and sediment yield is nonlinear and very complex seems
to be a valid geomorphic observation. Precisely at what precipitation the max-
imum values of sediment yield will occur depends on the specific climatic set-
ting (for example, see Fournier 1960), and total values may relate more to the
seasonality of the climate than to the mean annual precipitation (L. Wilson
1973). Some studies tend to support that contention (Corbel 1959; L. Wilson
1972; Jansen and Painter 1974). It should also be noted that variations in
sediment yield under different climates may be partially offset by an increase
or decrease in dissolved load. Normally, dissolved load will increase regularly
with precipitation, but maximum solution is probably reached at about 63 cm
of annual precipitation, with little additional dissolution resulting from higher
precipitation (Leopold et al. 1964). Regardless, the amount of dissolved load
usually does not exceed solid load even in regions of high precipitation, where
chemical loss should be great (Li 1976; Leigh 1982). This probably results
because the magnitude of solution loss is dominated by rock type rather than
climate or vegetation (Garrels and Mackenzie 1971).
Basin Size A number of studies have suggested that sediment yield de-
creases markedly as the size of the drainage basin increases. This hypothesis
is supported by remarkably high yields in very small basins of the midwestern
United States (Schumm 1963c) and also by data from the major sediment-
discharging rivers of the world (Milliman and Meade 1983; fig. 5.32) The
explanation for this phenomenon seems to lie in several topographic realities:
(1) small basins generally have steep valley-side slopes and high-gradient
stream channels that efficiently transport sediment; (2) in basins filled to ca-
pacity with streams, the drainage density always remains high near the basin
Chapter 5 196
divide, but it may decrease with time in the central part of the basin;
(3) fioodplain area increases as the basin expands, especially in the central
and lower reaches of the basin (Hadley and Schumm 1961).
The integration of these factors leads us to realize that in natural basins
most sediment is produced in the small headward subareas, but during its
downstream transit a significant portion may be stored in the fioodplain system.
How long it will remain in storage depends on the rigor of the geomorphic
and hydrologic processes. It has been assumed that sampling over a long enough
time would show that sediment stored within the basin is flushed rapidly from
the system during episodes of rejuvenation. Inclusion of such spasmodic bursts
of erosion in a long-term sample would tend to temper the variations in yield
that appear to be related to basin size.
There are reasons to believe, however, that the relationship between sed-
iment yield and basin size is less real than it appears because we tend to em-
phasize the extreme erosion in small marginal basins, and we do not fully
comprehend the equilibrium state of the master streams. For example, there
is ample evidence that the downstream decrease in sediment yield occurs during
episodes of accelerated erosion in small basins because storage in stream chan-

nels and valleys is increased (Trimble 1977). This occurs because the main
channel is unable to transport the additional load. Thus, sediment yielded from
the basin mouth decreases during a time of high erosion in the marginal basins,
indicating that rivers are in disequilibrium conditions throughout the basin.
The reverse can also happen; i.e., a reduction of erosion in the headland basins
by sediment control measures starves the main channel load and creates ero-
sion of the previously stored sediment, leading to increased yield measured at
the basin mouth (Trimble 1977). Thus, the apparent basin size-sediment yield
relationship may simply stem from the possibility that different parts of
drainage basins are responding in opposite directions to external erosional
stimuli in the small marginal zones. Modern sediment yield values, therefore,
may be related to recent changes in land use practice rather than an inherent
areal control.
Elevation and Relief Mountainous terrains with excessive elevation and re-
lief are known to produce abnormally high sediment
yields, particularly where
rocks are nonresistant (Corbel 1959; Hadley and Schumm 1961; Schumm
1963c; Ahnert 1970) or affected by recent or current tectonism (Li 1976). For
basins at least 3900 km 2 in area, the greatest denudation rates average about
0.9 m/1000 years and probably occur in mountain belts where relief and el-
evation are greatest and rocks are erodible (Schumm 1963c).
In the arid climate of the western United States, sediment yields are a
function of the relief-length ratio (fig. 5.33). Utilizing this fact and holding
area constant, Schumm (1963c) showed that the relationship between relief
and denudation is definable in quantitative terms. It has also been shown that
Figure 5.33.
Relation between mean annual
sediment accumulation in
reservoirs and the relief ratio for
basins located on the Fort Union
formation in the upper Cheyenne
River basin. (From Hadley and
Schumm 1961, fig. 31)
0.04 0.06 0.08 0.10 0.12

Relief ratio
elevation alone produces disparate denudation rates since low-lying areas in

any climatic regime yield less sediment than higher basins of comparable size
(Corbel 1959).
Significantly, relief and elevation analyses demonstrate clearly that de-
nudation rates are not constant through time. As relief and elevation of a basin
are gradually diminished during its evolution, the rate of surface lowering de-
creases proportionately, and each successive interval of stripping requires a
longer period of time (fig. 5.34).
Rock Type With similar climate and topography, basins underlain by clastic
sedimentary rocks and low-rank metamorphics usually produce abundant sus-
pended loads and so are characterized by higher rates of denudation than re-
gions of crystalline rocks or highly soluble sedimentary rocks (Corbel 1964).

The relationship of lithology and denudation is poorly understood in a quan-
titative senseand may be obscured by other rock characteristics, such as frac-

turing, which cause different lithologic units to behave similarly with respect
to denudation. Even so, properties that are commonly a reflection of lithology,
such as the infiltration capacity, seem to be systematically related to sediment

yield (fig. 5.35), indicating that the lithologic influence is real. At this time,
however, geomorphologists have not been able adequately to reveal its fun-
damental character.
Chapter 5 198
Relief-length ratio
Figure 5.34.
.02 .04 .06 .08
Relation of denudation rates to
relief-length ratio and drainage-
basin relief. Denudation rates are
adjusted to drainage areas of
1500 square miles Curve is
based on the average maximum
denudation rate of 3 feet per 1000
years when relief-length ratio is
05 (From Schumm 1963)
10,000 20,000 30,000

Relief (feet)
Figure 5.35.
Relation between infiltration
capacity and sediment yield,

indicating hthologic influence
(Data from Hadley and Schumm
1961)
^W i i i I <W
1.0
Infiltration capacity (in/hr)
2000
Figure 5.36.
"D „ Changes in sediment yield and
channel behavior in one area

under various types of land use
1000
<u
03
~o 600
200
1800 1860 1900 1960
Land use Forest Cropping Woods and Urban

grazing
Construction
Channel condition Stable Aggradation Scour Stable Scour

Aggradation
The Human Factor Most estimates of denudation are significantly inflated

where human activity disturbs the natural setting. Exactly how much humans
accelerate erosion varies with the type of land use and the particular environ-
ment, but it seems reasonable to propose that human interference has the po-
tential to alter drastically the natural sediment yield (Moore 1979; Dunne
1979; Toy 1982).
Evidence indicates that human activities may increase detrital loads by
at least an order of magnitude (Judson 1968a; Meade 1969); chemical loads
are expanded by pollutants introduced into streams or the atmosphere (Meade
1969). One important contributor to accelerated erosion is the replacement
of mature forest cover by intensely cultivated land (Toy 1982). Studies in the
United States show an increase in sediment yield of 1 to 3 orders of magnitude
when cropping is substituted for the natural vegetation (Ursic and Dendy 1965;
Wolman 1967). With proper soil conservation techniques, however, the effect
can be reversed and sediment yield values will decrease dramatically (Trimble
and Lund 1982).
Construction associated with urbanization causes an even more dramatic
rise in the sediment yield (fig. 5.36), but after construction is completed the
values decrease rapidly because much of the surface is protected from erosion
by our concrete citadels (Wolman 1967). Urbanization also affects the runoff
characteristics within a drainage basin (McPherson 1974); consequently, the
concentration of sediment in streams, as influenced by humans, depends on
variations in both hydrology and sediment yield.
Several workers have suggested that the abnormalities induced by human

activities may be great enough to invalidate the use of sediment yields to cal-
culate rates of denudation in drainage basins (Douglas 1967; Trimble 1977).
Trimble (1977) was able to show that in large drainage basins in the south-
eastern United States, lowering of upland surfaces was proceeding at a rate
of 95 mm/ 100 yr; in contrast, the denudation rate calculated by sediment in
the streams was only 5.3 mm/ 100 yr. Thus, the delivery ratio (sediment yield
Chapter 5 200
Figure 5.37.
Hillslopes
General linkages between
outside
sediment storage sites and
swales
erosional processes Boxes
indicate storage elements; listed
Sheet erosion
below each box are erosional
Soilcreep
processes mainly responsible for Slide-scarp erosion
Debris slides/flows
mobilizing sediment in that Headcut erosion
element Arrows show transfers Slide sheet erosion -
between elements Labels on Soil creep

arrows qualify or restrict location Debris slides/flows
of transfers (From Lehre 1982)
Gully
Channel Channel
banks bed
Gully-wall erosion Bed erosion

Bank erosion
Bank slides
Headcut erosion
as a proportion of upland erosion) was only 6 percent. Therefore, sediment

yield may be a poor indicator of how rapidly the upland area of a basin is
being lowered. The major portion of sediment eroded from the uplands is, of
course, being stored within the basin by deposition.
In addition, Meade (1982) has suggested that the accelerated erosion pro-
duced by early settlers in the eastern United States while clearing the land
for cultivation has been largely arrested by soil conservation and reduced
farming. However, sediment stored in the basins in response to the earlier set-
tlement is now being eroded and is augmenting the present river loads and will
continue to do so for decades or centuries. The point is that modern river loads
may be reflecting erosional events that occurred in the distant past.
Sediment Budgets
It is clear from the above that storage of sediment within a basin is a signif-
icant geomorphic phenomenon. Its importance is further realized with the
concept of a sediment budget (Dietrich and Dunne 1978; Kelsey 1980; Lehre
1982). A sediment budget is a quantitative analysis of a drainage basin that
shows the relationship among erosion of basin materials, discharge of sedi-
ment from the basin, and the associated changes in sediment storage. In es-
sence, it is an accounting sheet of ins and outs of sediment and as such is
similar to the hydrologic budget discussed earlier. Of greater importance,
however, is its and land management because it dis-
significance in planning
tinguishes the dominant erosional processes in the basin and the circum-
stances under which storage of sediment may change. Additionally, the sedi-
ment budget demonstrates the process linkage that we discussed in chapter 1
(fig- 5.37).
Table 5.5 Sediment budget for a small drainage in northern California for the years 1971-1974
Recurrence Production Redistribution Yield: Bed + Bank

Interval of Mobilization to on Bed + Susp. Storage
Year Rainfall Peak Flow on Slopes (1) Channels (2) Slopes (3) Load (4) (5 = 2-4)
M illimeters Years
1971-72 602 1.5 86 148 -71 24 + 124

1972-73 1184 15-20 1219 985 +317 1420 -435
1973-74 1046 3-5 1960 1575 +389 630 + 945
1971-74 2832 3265 2708 +635 2074 + 634
From Lehre 1982
To make a complete sediment budget analysis one must identify and

quantify sediment mobilization (processes that initiate motion and move sed-
iment any distance), sediment production (sediment reaching or given access
to a channel), and sediment yield (sediment actually discharged from the
basin). Having made such measurements, a balance sheet is constructed to
reveal the progression of sediment through the basin. For example, table 5.5
shows the sediment budget for the years 1971-1974 in a small drainage basin
northwest of San Francisco (Lehre 1982). In dry years or years without ex-
treme flow events (1971-1972, 1973-1974), most mobilized and produced
sediment was stored within the basin. In contrast, during the year having a
flow event with a 15- to 20-year recurrence interval (1972-1973), the amount
of sediment mobilized and produced was less than the amount discharged from
the basin. This indicates that sediment was taken out of storage during that
year, presumably by erosion of channel banks and beds. By continuous obser-
vation, Lehre (1982) was able to demonstrate that the variations in annual
sediment budgets shown in table 5.5 were accompanied by different erosive
processes. In the years of low rainfall and/or peak flow, sediment was mobi-
lized by spalling, rainbeat, and minor sliding and was moved toward the chan-
nels by sheetwash. The lack of water, however, assured a minimum transport
distance and most sediment went into storage. In the year of high peak flow
(1972-1973), sediment was mobilized and produced mainly by debris slides
and flows. It was quickly delivered to the channels and because of the coin-
cidence of high channel discharge was removed from the basin.
Sediment budget analysis new way of looking at the inner
is a relatively
workings of the drainage basin, and we do not yet know what its applications
will be. However, it seems to have considerable promise for land management,
especially in small basins that are unstable geomorphically and subject to a

variety of interrelated processes.
Rates of Denudation
Regardless of the problems inherent in denudation-rate analyses, various es-
timates have been made. Table 5.6 presents a random sample of modern de-
nudation rates for basins of varying size in the United States. The values are
imprecise and probably high because in most cases they do not include an
Chapter 5 202
Table 5.6 The influence of geology and climate on suspended-load denudation

in basins of different size in the United States
Average
Annual
Suspended
Load Denudation
Basin Location Area (mi 2 ) (tons x 10 3 ) (in/1000 yr)
Mississippi Baton Rouge, La 1 .243.500 305,000 1 3

Colorado Grand Canyon, Ariz, 137.800 149,000 56
Columbia Pasco, Wash, 102,600 10,300 05
Rio Grande San Acacia. N.M 26,770 9,420 1.8
Sacramento Sacramento, Cal 27.500 2,580 05
Alabama Claiborne. Ala 22.000 2.130 05
Delaware Trenton, N J 6.780 998 08
Yadkin Yadkin College. N.C. 2.280 808 18
Eel Scotia, Cal 3,113 18.200 30.4
Rio Hondo Roswell, N.M, 947 545 30
Green Palmer, Wash 230 71 1.6
Alameda Niles. Cal 633 221 18
Scantic Broad Brook, Conn. 98 7 04
Napa St Helena, Cal 81 63 4 1
Data from Judson and Ritter 1964
adjustment for human impact. Nonetheless, they do indicate the general ten-
dency for denudation rates to fall between 2.5 and 15 cm per 1000 years when
considered on a regional scale. Judson (1968b) recalculated the denudation
rate for the entire continental United States by subtracting the effect of human
occupancy from earlier estimates. His figure of 3 cm/ 1000 years agrees rather
well with the denudation in very large drainage basins that are mostly unaf-
fected by humans (Gibbs 1967) and perhaps represents a reasonable approx-
imation for denudation on a continental scale.
Although methods other than analyses of sediment wedges have been em-
ployed (Eardly 1967; Ruxton and McDougall 1967; Clark and Jager 1969).
most past rates of denudation are estimated from measurements of sediment
accumulation in depositional basins. A valid estimate can be made only if
(1) the volume of sediment derived by erosive processes can be accurately de-
termined, (2) the boundaries of the source area are definable, and (3) the time
interval of sediment accumulation can be ascertained within reasonable limits.
It is very difficult to meet all these requirements: noneroded matter such as

pelagic and volcanic rocks add to the depositional volume; material eroded
from the basin as dissolved load may not be returned to the deposit by chem-
ical precipitation;and absolute dates that bracket the time of deposition are
necessarily imprecise. Nonetheless, some estimates of past rates have been
made for large portions of North America (Gilluly 1949, 1955. 1964; Menard
Table 5.7 Denudation estimates based on solid and dissolved loads delivered to
the oceans by major rivers of the continents.
Denudation
Solid Load Dissolved Load Total Rate
Continents (t/km 2 /yr) a (t/km 2 /yr) ft
Load (cm/1000 yr)
North and Central 84 33 117 4 00

America
South America 97 28 125 428
Europe 50 42 92 3.15
Asia 380 32 412 14 10
Africa 35 24 59 202
Australia 28 2 30 1 03
3 Solid load from Milhman and Meade 1 1983)

"Dissolved load from Garrels and Mackenzie (1971)
1961). It is interesting to note that the rates found in these studies of large
regions are within the same order of magnitude as those based on modern
stream data. Such similarity prompted the hypothesis (Ritter 1967) that when
viewed on a large enough area or over a long time interval, denudation rates
will probably be about the same. Based on current evidence, the average value
will probably fall somewhere between 2.5 cm and 15 cm per 1000 years. That
range is generally supported by estimates of solid and dissolved loads being
delivered to the oceans from the world's continents (table 5.7).
The use of denudation rates calculated for continent-size areas (table 5.7)
can lead to terribly incorrect conclusions. For example, a 3 cm/ 1000 yr rate
suggests that 300 m of surface lowering will be accomplished over an entire
continent in a 10-million-year period. Such a rate might be used as evidence
to support the generally accepted canon of geomorphology that most of the
earth's topography is no older than Pleistocene, i.e., almost all landscapes
formed in the last 2 million years. Therein lies the fallacy of denudation rates
because we know that large regions of Tertiary and older landforms do exist,
especially in the Southern Hemisphere. For example, radiometric dates and

geologic evidence in southeastern Australia show that much of that landscape
was in its present form by mid-Miocene, and some upland surfaces originated
in the Mesozoic (Young 1983). The incompatibility of these observations and
denudation analyses from river sediment arises because denudation rates are
unrealistically spread evenly over entire continents. Actually, the interiors of
continental plates probably experience extremely slow denudation and may
easily preserve old landscapes (Young 1983). In contrast, continental plate
margins where active tectonism is occurring probably have enormously high
denudation rates. Combining the two subareas provides an average rate for
the continent that is indicative of neither. The point here is that a denudation
rate calculated for a large basin or region tells us nothing about the tenor of
erosion occurring in any component part of that basin or region, and the overall
rate should never be used in that sense.
Chapter 5 204
Summary In this chapter we examined a remarkable statistical balance among the spa-
tial characteristics of river networks and the watersheds that contain them.
Because the parameters of this morphometry also relate in a significant way
to the hydrologic and erosional properties of most watersheds, drainage basins
serve as primary units for systematic analyses of geomorphology. Drainage
basins and their river networks probably evolve according to fundamental hy-
drophysical laws, but their ultimate character is conditioned by the geological
framework and the external constraints of climate. An equilibrium condition,
defined in terms of mathematical balance, is probably attained early in the
growth history of most basins. This does not indicate, however, that basins
evolve in an orderly way with time. Geologic catastrophes that upset equilib-
rium tend to be filtered out in a morphometric sense because basinal param-
eters apparently adjust to changes rather quickly.
Water flowing in basin rivers is derived from variable sources, and sedi-
ment reaching stream channels is produced and delivered by numerous erosive
processes operating on basin slopes. Both water and sediment are amenable
to budget analyses, which provide basic data for watershed planning. The
amounts of water and sediment entering stream channels are functions of the
physical properties of the basin and the effect produced by human activities.
Estimates of basin denudation can be made on the basis of total sediment
yield, but difficulties are created by sediment storage within the basin.
Chorley, R. 1969. Water, earth and man. London: Methuen and Co.
J.
Costa, J. E., —
and Baker, V. R. 1981. Surficial geology Building with the
Earth. New York: John Wiley & Sons.
Doornkamp, J. C, and King, C. A. M. 1971. Numerical analysis in
geomorphology: An introduction. London: Edward Arnold Ltd.
Dunne, T. 1978. Field studies of hillslope flow processes. In Hillslope
hydrology, edited by M. J. Kirkby, pp. 227-94. New York: John Wiley
& Sons.
Dunne, T, and Leopold, L. B. 1978. Water in environmental planning. San
Francisco: W. H. Freeman.
Horton, R. E. 1945. Erosional development of streams and their drainage
basins: Hydrophysical approach to quantitative morphology. Geol. Soc.
America Bull. 56: 275-370.
Judson, S. 1968. Erosion of the land. Am. Scientist 56: 356-74.
Patton, P. C; Baker, V. R.; and Kochel, R. C. 1979. Slack water deposits:
A geomorphic technique for the interpretation of fluvial
paleohydrology. In Adjustments of the fluvial system, edited by
D. D. Rhodes and G. P. Williams, pp. 225-53. Dubuque, Iowa:
Kendall/Hunt.
Schumm, S. A. 1963. Disparity between present rates of denudation and
orogeny. U.S. Geol. Survey Prof. Paper 454-H.
Fluvial Processes
6
I. Introduction V. Channel Patterns
II. The River Channel A. Straight Channels
A. Basic Mechanics B. Meandering Channels
B. Flow Equations and C. Braided Channels
Resisting Factors 1. The Origin of Braids
III. Sediment in Channels 2. Anastomosed
A. Transportation Channels
B. Entrainment and Bank D. The Continuity of Channel
Erosion Patterns
C. Deposition VI. Rivers, Equilibrium, and Time
D. The Frequency and A. Adjustment of Gradient
Magnitude of River Work B. Adjustment of Shape and
IV. The Quasi-Equilibrium Pattern
Condition VII. Summary
A. Hydraulic Geometry VIII. Suggested Readings
B. The Influence of Slope
C. Channel Shape
205
Chapter 6 206
Introduction It seems fair to say that fluvial action is the single most important geomorphic
agent. Although other surficial processes are significant, streams are so ubiq-
uitous that their influence in geomorphology can hardly be overestimated. As
discussed in chapter most geomorphic analyses of stream processes in the
1,
early twentieth century were based primarily on logic and qualitative obser-
vation. The few quantitative studies of fluvial mechanics were somehow lost
in the wave of geomorphology that used fluvially produced landforms to re-
construct history. Our understanding

of river mechanics is hampered by the
measurements in a natural channel. Much of our knowl-
difficulty of obtaining
edge of fluvial mechanics has therefore been derived from studies in flumes,
where slope and velocity can be varied and any variable can be held constant.
Unfortunately, flumes have a limited range of discharge and depth, and al-
though the application of flume results to real systems gives an excellent first
approximation, it cannot take into account the complexities of the many ev-
erchanging and interdependent variables in natural streams (see Maddock
1969). We still need much more data about natural rivers before we can hope
to utilize our knowledge in a predictive way. Nonetheless, there has been a
rebirth of interest in the details of fluvial mechanics, and data are now ac-
cumulating faster than one can assimilate their meaning. This explosion of
information is undoubtedly the key to the advancement of geomorphic thinking,
for without it our basic tenets will never be critically evaluated. In this chapter
we will examine briefly much of the current thinking about streams; those
interested in greater detail than space allows here are referred to a number
of excellent books about the fluvial realm: Leopold et al. 1964; Leliavsky 1966;
Morisawa 1968; Raudkivi 1967; Chorley 1969a; Graf 1971; Schumm 1972,
1977; Gregory and Walling 1973; Richards 1982; Hey et al. 1982.
The River Channel Basic Mechanics

The a bility of a river to do geomorphic work represent s^ balance betwee n
driving and resisting forces. It depends on how much potential energy is pr o-
vided to t he flo w ancl how much of that energy is consumed in the system by
the various resisting elements The resistance generated along the channel pe-
.
rimeter or within the flow by the shape of the channel, thfi-Size

is_c ontrolled
and concentration of sediment ,a nd the total volume and physical propertie s

of the w a ter. To illustrate, let us examine what happens to a constant discharge
of water flowing in a long, steeply inclined channel that has no tributaries.
Gravity tends to accelerate the flow downstream continuously unless the in-
crease in velocity is moderated by friction within the water and turbulence
generated along the channel perimeter. If these resisting elements were con-
stant and less than the gravitational force, acceleration would continue along
the entire length of the channel. In natural rivers, however, the intensity of
^resistance is not constant but increa ses with the flow velocity. At any point in
a river, therefore, the velocity~represents IHe ba lance be tween the energy
causing flou and the energy consumed by resistancetoHow!
207 / /Fluvial Processes J
P =w+ 2d
Figure 6.1.
A = wd Cross-sectional measurements of
a stream channel: w= width;
d= depth; A = area
R = hydraulic radius;
P= distance along wetted
perimeter
Resistance is spread through the water by viscous or turbulent processes

and is_proportional to the velocity or its square. In l aminar flow, particles of
water move in straight paths that are not disrupted!)" the movement of neigh-
boring particles. In this type of flow regime Jmernal friction is the dominant
resisting force, and its intensity is proportia&fto the velocity of flow. Because
the c hange in velocity with dep th is linear, r esistance increa se s uniformly from
t he water s urface to the bed, the rate of increase being determined by the
molecular viscosity of the fluid. Most resistance in laminar flow results from
intermolecular viscous forces as layers or particles of the fluid slide smoothly
past one another (Leopold et al. 1964). The controls on viscosity are internal
characteristics of the fluid such as temperature and suspended sediment con-
centration.
Injuxhulfcnt flow , the water does not move in parallel layers; its velocity
fluctuates continuously-anall directions within the fluid. Water repeatedly in-
terchanges between neighobnng zones of flow, and shear stress is transmitted

across layer boundaries in anofber form of viscosity, called eddy viscosity. Eddy
viscosity greatly increases thexpsipation of energy and the flow resistance. In
turbul ent flow resistance is proportional to the square of the velocity. Resis= .
tance and velocity, therefore, do not change uniformly with depth. In _addilioji,
because turbulence generated along the channel boundaries, most resistanc e
is
in this type of flow due to external factors such as the channel configuration
is
and the size of t he bed mate rial.

As depth and velocity increase, the conditions at which laminar flow
changes to turbulent can be predicted by a dimensionless parameter called the
Reynolds number (Re):
Re = VRp \
where V \s the mean velocity, R the hydraulic radius, p the density, and \i the
molecular viscosity. The hydraulic radius R is determined by the relationship
R~ p ,
where A is the cross-sectional area of the channel and P the wetted perimeter
(fig. 6.1). In wide, shallow channels the hydraulic radius closely approximates
the mean depth.
Chapter 6 208
Since the factor — defines a fluid property called kinematic viscosity (t>),
the Reynolds number represents a ratio between driving and resisting forces
because
VRp _ VR _ driving force

'
M v resisting force
In normal situations true laminar flow occurs where Re values are less than
500, and well-defined turbulent flow when Re is greater than 750.
Another dimensionless number used to describe the conditions of flow is
the Froude number (Fr ):
F = r
?r
where d is depth and g is gravity. The Froude number is important because it
can be used to distinguish subtypes of turbulent flow called tranquil flow

(Fr 1), critical flow (Fr = 1), and rapid flow {F,
< 1). >
Flow Equations and Resisting Factors
Flow and resistance have been the concern of hydraulic engineers for centu-
and a number of equations have been derived to express the relationships
ries,
between the factors. Two equations of great importance to students of rivers

are the Chezy equation and the Manning equation Both deriy efrom equating .
djj^jng_and_ resisting forces in nonacCeleratin^ flow, and both have been em-
ployed in a variety of fluvial investigations. The Chezy equation
V = C \/RS ,
derived in 1769, shows that velocity is directly proportional to the square root
p Av-^ of the RS product, where 5 is the slope of the channel. The Chezy coefficient
\\c, X^ (C) is a co nstan t of proportionality that relates to resisting factors in the system.
\.\,^'°C^^ TheJ Manning] equation originated in 1889 from an attempt by Manning
i « to systematjz£-lhg_existing data into a useful form. The equation
yA^
.a ~*~
c° is similar to the ChezyTormula in that velocity is proportional to /? and 5. In
V v
fôw? ^ \f\*° addition, the factor n, called the Manning roughness coefficient, is also a re-
*"
^V\r°^' t K sitting element that is closely related to the Chezv coefficient because as
C.L*"* C(RS) 12 = (~) R1 \S'
then C= 1.49 1^—1 .
Manning's n is presumed to be a constant for any particular channel frame-

work; consequently, it has been used extensively in analyses of river mechanics
1
209 Fluvial Processes
Table 6.1 Manning roughness coefficients (n) for different boundary types.
Manning
Roughness
Boundary n(ft 1/6 )
Very smooth surfaces such as glass, plastic, or brass 0.010

Very smooth concrete and planed timber 0.01
Smooth concrete 0.012
Ordinary concrete lining 013
Good wood 014
Vitrified clay 015
Shot concrete, untroweled, and earth channels in best condition 0.017
Straight unlined earth canals in good condition 020
Rivers and earth canals in fair condition — some growth 025
Winding natural streams and canals in poor condition — considerable moss 0.035
growth
Mountain streams with rocky beds and rivers with variable sections and 0.041-0.050
some vegetation along banks
From Handbook of Applied Hydrology by Ven T Chow Copyright© 1964 by McGraw-Hill Inc Used with permission of
(table 6.1). The U.S. Geological Survey, for example, has developed a visual
guide for rapid estimation of Manning's n (Barnes 1968). It should be noted
that substituting depth for hydraulic radius in calculations using the Manning
equation will introduce considerable error (Tinkler 1982). Therefore, R rather
than d should always be used to calculate velocity or resistance coefficients.
Although resist ance coefficient s are d efined by hydraulic characteristics
(S^R-t-V), they are not independent of other factors because in alluvial chan-
nels their values vary with particle size, sediment concentration, and bottom
corrfiguralion. For example, in thesame reach of a river, different flow con-
ditions (as determined by the Froude number) may mold bottom sediment
into a variety ofbed forms (fig. 6.2), which in turn have a decided influence
on the value of Manning's n (table 6.2). I rregularity of the channel bottom
due to other factors s uch as bars, riffles, and be nds_may have an equally im-
_
portant influence on the roughness (Leopold et al. 1964; Simons and Rich-
ardson 1966). In fact, Prestegaard (1983a) was able to demonstrate that in
gravel-bed reaches of 12 western U.S. rivers, bar resistance accounted for 50
to 75 percent of the total flow resistance. Particle size of bed material also
causes variation of roughness values. Wolman (1955) demonstrated this last
fact in the Brandywine Creek (Pennsylvania), where the resistance-particle

size relationship is defined by the equation
= 2 1o g7
f+ 1.0,
where F is another resistance parameter called the Darcy-Weisbach resistance

coefficient, is water depth, and D H4 is the particle diameter, which is equal
d
to or larger than 84 percent of the clasts on the channel bottom. Increased
particle size should produce a correlative increase in flow resistance.
Chapter 6 210
Figure 6.2.
Bed forms in alluvial channels and
their relation to flow conditions.
F = Froude number, d = depth.
(A) Typical ripple pattern, F << 1 (E) Plane bed, F < 1 and d < 0.4mm
Weak boil
(B) Dunes with ripples superposed F << 1 (F) Standing waves, F > 1
Boil Incipient breaking,

and moving upstream"
(G) Antidunes, F> 1
Breaking antidune wave
Accelerating flow
(D) Washed-out dunes or transition F < 1 (H) Antidunes, F > 1
Table 6.2 Variation of Manning n values with changes in bed form occurring
under different flow conditions.
Manning n
Bed Form (ft")
Lower Regime
Ripples 0.017-0 028
Dunes 0.018-0.035
Washed-out Dunes or Transition 014-0 024
Upper Regime
Plane bed 011-0 015
Standing waves 0012-0.016
Antidunes 012-0 020
From Handbook of Applied Hydrology by Ven T Chow Copyright © 1964 by McGraw-Hill Inc Used with permission of
0.0120
Figure 6.3.
Effect of suspended load
concentration on the Manning
roughness coefficient n
0.0110
2 0.0100
0.0090
5 10 15 20
Mean sediment concentration (grams/liter)
Clearly, external boundary conditions such as channel shape and particle

size generate a large amount of resistance. Some of these factors produce tur-
bulence in the form of eddies and secondary circulation. In contrast, sediment
concentration (the amount of sediment per unit volume of water) affects re-
sistance internally. This modification was first detailed by Vanoni (1941, 1946),
who showed that an increase in th e concentration of suspended sediment t ends
to lo wer the resistan ce (fig. 6.3). As the concentration increases, the turbulent,
effect presumably is reduced because the mixing process within the fluid is
dampened. All other factors being equal, sediment-laden water should flow at
a highe r velocity thanc lear water..
Although most energy in a stream is d issipated by turbulence a small par t is , Sediment in Channels
usedin_the important task of_groding and transporting sediment ^These pro-
cesses, often taken for granted, are extremely complex and poorly understood,
yet they underlie some of our most basic geological concepts. We will briefly
review the more significant ideas about the relationship between river flow and
sediment.
Transportation
In general, fine-grained sediment (silt and clay) is transported suspended in
the water by the supporting action of turbulence. Suspended load usually moves
at a velocity slightly lower than that of the water and may travel directly from
place of erosion to points far downstream without intermittent stages of de-
position. Coarse particles may also travel in true suspension (Francis 1973),
but they are likely to be deposited more quickly and stored temporarily or
semipermanently within the channel. Except for short spasms of suspension,
coarse sediment usually travels as bedload. Bedload refers to sediment trans-
ported close to or at the channel bottom by rolling, sliding, or bouncing. How
long coarse debris remains stationary within a channel depends on the size of
Chapter 6 212
the debris and the flow characteristics of the river; such debris probably is
immobile more than it is in motion. Bradley (1970) showed that gravel can
be stored in channel bars long enough for weathering to drastically weaken
its resistance to abrasion.
Because of fluctuating discharge, at any given time a single particle may
be part of either the bedload or the suspended load. As this makes the dis-
tinction between the two load types unclear, other terms have been devised to
relate sediment more appropriately to river flow. Wash load consists of par-
ticles so small that they are essentially absent on the stream bed. In contrast,
bed material load is composed of particle sizes that are found in abundant
amounts on the stream bed (Colby 1963). While most, if not all, bedload is
bed material load, most bed material load is transported as suspended load.
The relationship between wash load and discharge is poorly defined be-
cause most streams at any flow can carry more fine-grained sediment than
they actually do. In fact, the concentration of fines is a function of supply
rather than transporting power; therefore, it is relatively independent of flow
characteristics. Coarse sediment, on the other hand, is usually available in
amounts greater than a stream can carry, and so its concentration should cor-
relate more significantly with the parameters of flow such as depth and ve-
locity. The problem, however, is that direct measurement of bedload is
extremely difficult because hand-held instruments can sample for only short
periods, and when they are placed on the channel bottom the flow regime is
disrupted. In addition, where bedload has been continuously measured, the
amount of sediment passing a given channel cross section varies significantly
on a time interval as short as one minute (Leopold and Emmett 1977). Fur-
thermore, the amount of bedload varies drastically in different subwidths of
the channel cross section.
Because of the difficulties surrounding direct measurement, most esti-
mates of bedload discharge are made by means of empirical equations relating

flow parameters and bedload transport rates (Meyer-Peter and Muller 1948;
Einstein 1950). These equations, however, are themselves problematical be-
cause small errors measurement of the flow parameters produce large vari-
in
ations in the computed load. In addition, the equations are valid only when
hydrologic and sedimentologic conditions are the same as those that prevailed
at the time the equation was derived (Andrews 1981).
The problems inherent in deciphering the relationship between parame-
ters of flow and bedload transport are difficult but not insurmountable. Some
very good bedload measurements have been made on the East Fork River,
Wyoming (Leopold and Emmett 1976, 1977), where the U.S. Geological
Survey has installed a concrete trough across the channel floor. Sediment
moving along the river bottom falls onto conveyor belts rotating within the
trough and is carried to the channel side where it is weighed (fig. 6.4). In
addition, other measuring techniques and instruments have been used suc-
cessfully (Helley and Smith 1971; Emmett 1980; Reid et al. 1980).
Figure 6.4.
U.S.G.S. bedload sampling
station, East Fork River, Wyo.
(A)View across river showing
suspension bridge and drive
mechanism of conveyor belt
bedload sampler (B) Bedload trap
on streambed visible below
suspension bridge.
Chapter 6 214
Entrainment and Bank Erosion

The temporal and spatial variationsin bedload transport rate noted at the East
Channel bed Fork River station are related to the mechanics involved in moving coarse-
grained sediment. Most large particles do not move great distances during any
d transporting event. Instead, their downstream migration is characterized by
spasmodic bursts of short-distance movement separated by periods during

which they come to and remain at rest. The processes that initiate the bursts
of motion experienced by any particle are collectively known as entrainment.
The amount of sediment entrained depends directly both on the erosive power
Figure 6.5. of the flow and on the size of particles on the bed surface that are in the proper
Component of flowweight exerted
as shear stress on a channel position to be eroded. Two streams with identical flow conditions may have
bottom. The critical shear stress different bedload or bed material discharge if one flows across a fine-sand
is the depth-slope product (dS)
bottom and the other over a cobble bed.
multiplied by the specific weight
of the water and Is the angle of
The term competence refers to the size of the largest particle a stream
slope can entrain under any given set of hydraulic conditions. It should be evident
that the value of competence to geomorphology depends on how we measure
the sediment being moved and, more importantly, how accurately we can de-
termine the flow conditions. Although this may seem simple enough, in prac-
tice it is an excruciating problem for several reasons: (1) particles are entrained
by a combination of fluvial forces including direct impact of the water, drag,

and hydraulic lift, and each of these may be best defined by a different pa-
rameter of flow; (2) flow velocity is neither constant nor easily measured, es-
pecially during high discharge; and (3) sediment of the same size may be
packed together differently or have shape properties that cause abnormal re-
sponses to the same flow conditions. Thus, any investigation into the me-
chanics of competence must settle for only partial success until we can eliminate
or inhibit some of the inherent variability.
Historically, two hydraulic factors have been utilized to represent the flow
condition in the competence relationship. The first, critical bed velocity, is
used to demonstrate the relationship between velocity and entrainment. It has

been long known (Rubey 1938) that the volume or weight of the largest par-
ticle moved in a stream varies as the sixth power of the velocity. The sixth-
power law provides a sound theoretical basis for competence studies, but it is
less satisfying in practice because accurate measurement of bed velocity is
exasperating — if not impossible — in high-energy streams. Mean velocity, an
easily determined substitute in competence work, may not be reliable unless

the depth is also considered in the analysis (Vanoni et al. 1966).
The second factor, critical shear stress (sometimes called critical tractive
force), signifies the downslope component of the fluid weight exerted on a bed
particle (fig. 6.5). It is proportional to the depth-slope product and can be
expressed by the DuBoys equation for boundary shear:
Te = yRS,
where re is the critical shear stress, 7 the specific weight of the water, R the
hydraulic radius, and 5 the slope. In most streams transporting coarse bed-
load, R is closely approximated by mean depth, but as indicated earlier, care
Figure 6.6.
Sediment particles of different
sizes begin to move on
the
streambed values of
at different
mean velocity and depth-slope.
Smallest particles (B) move mainly
as a function of dS while largest
particles move primarily as a
(/-/)
function of the mean velocity.

(From Rubey 1938)
30 40 50 60 70 80
Mean velocity (cm/sec)
should be taken before a substitution of parameters is made (Tinkler 1982).

The use of critical shear stress in competence studies has been criticized (Yang
1973), but the simple reality that depth and slope in a river are easier to mea-
sure than bed velocity makes it an appealing parameter.
how the two methods correlate with each other is not completely
Precisely
understood, but Rubey (1938) presented evidence to suggest that in the size
range between fine sand and pebbles, critical bed velocity becomes more im-
portant in the entrainment process as particle size increases (fig. 6.6). Smaller
sizes move as a function of the dS product and seem to be relatively indepen-
dent of velocity. Thus, the shear stress approach may be completely valid only
for smaller sizes or low-velocity flows, and very fine-grained sediment requires
higher velocities for its entrainment than the sixth-power law would predict.
Wolman and Brush (1961) found similar trends in a later flume study of the
movement of sand-sized debris.
Chapter 6 216
1000
Figure 6.7.
Mean velocity at which uniformly
sorted particles of various size are
eroded, transported, and
deposited
I I I I I III .
to
CM co CM CO If) O
o o o o
o o CM co in
o o o o o o CM co -n
o o o o o o o o o o o o
Size of particles (mm)
The above observations fit rather nicely the curves produced by Hjulstrom
(1939) and shown which relate current velocity, particle size,
in figure 6.7,
and process; as figure 6.8 shows, Rubey's size classes seem to fall along the
trend of the boundary between erosion and transportation, that is, the threshold
velocity needed to initiate motion. The velocity that produces erosion of clay-
sized particles is in some cases as great as that needed to entrain larger ma-
terial. This explains the commonly observed phenomenon of coarse particles
being transported across stationary material of a smaller size.
Unfortunately, flumes are not useful in the study of competence when par-
ticles are larger than pebble size. Most competence investigations of coarser
sediment have therefore been made in natural rivers or canals and, for reasons
explained earlier, employ shear stress as the diagnostic hydraulic variable (Lane
1955; Fahnestock, 1963; Kellerhals 1967; Scott and Gravlee 1968; Church
1972, 1978; Baker 1973b; Baker and Ritter 1975). Shear stress analyses such
as these have produced widely divergent results concerning competence. An-
drews (1983) suggests that much of this variation comes from our failure to
consider sediment characteristics in the competence analyses. He was able to
show that the size distribution of bed material has a significant effect on the
shear stress required to entrain a particle of any given size. The variation in
estimates of competence also prompted Bradley and Mears ( 980) to combine 1
a number of earlier techniques to gage the velocity and depth associated with
entrainment.
Medium Coarse Very Figure 6.8.

Fine sand Granules Pebbles
sand sand coarse sand Portion of Hjulstrom curve with
00
size grades B to H placed on
90 -
80
70
fin
50
1
Erosion (and transportation)
/ • H
/
/
diagram. Compare
6 7 (Adapted from
figures 6.6and
Rubey 1938)
10
'
30
?n
T
Exact boundary depends on
'tea ^*"^
^*"
/ F
/
>
^ .— — — *
• d y
B
10
9
g
|
| Transportation
/
H
7
(after movement has st arted) y/|
6
> 5
4
/ Depo sition
^r
3
/
1
1 2
Diameter of particles (mm)
In a related type of study, Bagnold (1973, 1977) proposed that entrain-

ment and transportation of bedload may be analyzed in terms of stream power.
Stream power is defined as
0) yQS,
where w is stream power and 7, Q, and S are specific weight, discharge, and
slope, respectively. If power is considered per unit area of the stream bed, it
essentially becomes a combination of shear stress and velocity because
<" = -^ = ydSv =
width '
rV,
where Kis mean velocity. Where available stream power is greater than that
needed to transport load, scour of bed alluvium (entrainment) will occur. Bull
( 1 979) suggests that stream power analyses can be used to determine the like-
lihood of a threshold response in rivers.

Chapter 6 218
The processes of entrainment determine the type and magnitude of ero-

sion that occurs on the channel floor. It is incorrect, though, to assume that
the only significant erosion is vertically directed. Bank erosion, which proceeds
laterally, not only contributes to the sedimentary load but also, through its
control on channel width, exerts a direct influence on other channel processes.

A large number of studies have identified the major processes involved in bank
erosion (see Thorne, 1982). Invariably researchers of this phenomenon arrive
at the striking conclusion that bank erosion is rarely, if ever, accomplished by
a single process but instead involves some combination of processes unique to
the individual setting. In general terms, two major categories of erosion have
been recognized: fluvial entrainment and processes of weakening and weath-
ering (Thorne 1982).
Fluvial entrainment promotes bank erosion in two ways. First, sediment
may be entrained directly from the bank surface by the shear stress generated
in river flow, a process usually referred to as corrosion. Second, 'differential
corrasion often produces an overhanging ledge of cohesive sediment because
a noncohesive layer in the bank material has beenmore rapidly eroded (Thorne
and Tovey 1981). The overhanging bank sections, called cantilevers, are mo-
bilized when the undercut cohesive material finally fails and drops to the sur-
face below. A similar process operates where vertical fractures, called tension
cracks, exist in the floodplain sediment. In these cases, lateral trimming of the
bank at the water level intersects the tension crack and produces downward
failure along the fracture plane. This type of movement has been called soil
fall (Brunsden and Kesel 1973), earth fall (Twidale 1964), slab failure (Hag-
erty 1980), and shallow slip (Thorne 1982). The failure is usually associated
with release of river pressure against the banks and constant removal of debris
from the bases of the banks. Normally, slab failure does not involve as much
bank material as other mass movements discussed below, but the process is
very significant because it occurs frequently (Thorne and Tovey 1981).
Weakening and weathering tend to reduce the strength of bank materials
and thereby promote instability and failure. The mechanics of failure depend
on many variables such as geometry and structure of the bank and the phys-
ical-chemical properties of the bank material. The most important control on
weakening of bank material is the soil moisture condition. This condition de-
pends on both climate and bank properties, and it is transformed into bank
erosion by processes that (1) reduce strength within the bank and (2) act on
the bank surface to loosen and detach particles and their aggregates. For ex-
ample, many researchers note a distinct seasonality associated with bank ero-
sion rates, winter and spring rates being considerably greater than those in
the summer. Presumably this results from high moisture contents and the ef-
fect of frost action in the winter and spring months (Wolman 1959; Twidale
1964; Hill 1973; Thorne and Lewin 1979).
Where saturated banks are found in poorly drained, cohesive sediment,
positive pore pressure can decrease the strength of bank material. This is es-
pecially true in high, steep banks after prolonged precipitation or rapid draw-
down of the river level. Under these conditions bank failure may occur by
rotational sliding (see figure 4.21). In some cases, the stratigraphy of the
floodplain sediment plays an important role in bank failures, especially where

cohesive layers rest above and below a noncohesive layer. Usually the non-
cohesive layer consists of permeable sands or silts in contrast to the cohesive
material, which is normally richer in clay. In these cases, the cohesionless zone
often serves as an avenue of pronounced seepage of underground water. Sap-
ping at the base of the overlying unit may occur, or the seepage may actually
transport material from the noncohesive unit in a process known as piping
(Deere and Peck 1959; Hagerty 1980). In either case, internal strength is lost
and failure of the overlying mass usually follows. In addition, the top horizon
of the clay-rich cohesive unit beneath the zone of seepage may become a lu-
bricated surface that serves as a sliding plane for overlying material. This type
of failure, referred to as a planar slide (see figure 4.20), usually functions on
sloping surfaces, but where shear resistance on the plane of sliding is very low,
the movement can occur on a horizontal surface. Planar slides (fig. 6.9) have
been recognized as important factors in the erosion of bluffs along the Mis-
sissippi River (Brunsden and Kesel 1973) and, in combination with other types
Figure 6.9.
of movement, are probably quite common (Varnes 1958). They require, how-
Bank erosion along the Osage
ever, that vertical fractures (tension cracks) exist in the bank sediment. River in central Missouri by lateral
spreading and planar sliding
1 JSJTt' V "^l
^^^^^B^^^St*^_ ^^î^H
Skr9r w
«-
«r*> ^
^^
^H| ^
V. v
-%
1
:
fegN
%
•
-4 JJ
>
J
*** »BL ***
• y
s^ssi^^*
&%*-* .' /* ^ •>*
^^f»
1 '«**
^rV
t
*A
J^^/jr
^. ;
•
"XT'
3l
Chapter 6 220
The important message in the above is that many times riverbank erosion
has nothing to do with rivers. Often mass movement phenomenon con-
it is a
trolled by the texture and stratigraphy of floodplain sediment and triggered
by the movement of groundwater. In light of this, the rate of bank erosion in
alluvial channels can be enormous or minuscule, depending mainly on the
character of the bank materials. In general, banks that are composed of fine-
grained sediment or are densely vegetated (Hadley 1961; D. G. Smith 1976)
have more resistance to corrasion than channels with sandy or gravelly banks.
The actual process of erosion, however, may differ; clay-rich banks usually
retreat by undercutting and failure of large blocks of the bank (Stanley et al.
1966; Laury 1971), while more coarse-grained banks erode by dislodgement
and sloughing of individual particles. Even highly cohesive banks may there-
fore erode rapidly if the dominant process is undercutting and/or mass failure.
Deposition
If the entrainment of sediment logically represents a threshold of erosion, a
similar threshold must exist when sediment in transport is deposited. Sus-
pended rock and mineral fragments tend to settle to the bottom at a rate that
depends on the density of the water, the fluid viscosity, and the size, shape
(Komar and Reimers 1978), and density of the sediment. The distance any
suspended particle will be transported in one event depends on its fall velocity
and on whether its downward settling is offset by turbulent forces in the water.
Coarse particles tend to be deposited during minor fluctuations in velocity (see
figure 6.7). Variations in any of the fluvial properties make the channel floor
a dynamic interface where some particles are being entrained while others
simultaneously are being deposited. The net balance of this activity depends
on local conditions rather than on the average cross-sectional hydraulics: in
fact, local effects may be different from those that occur over a long reach of
the channel (Colby 1964). It is known that scour and fill may be happening
in the same channel reach at any given time, and a number of alternating
events may coincide with fluctuating discharges. A long episode in which less
sediment leaves the bed than is returned results in a distinct period of aggra-
dation, and the converse situation leads to an episode of degradation. Long-
term events are caused by changes external to the channel (climatic or tec-
tonic) that introduce into any stream reach more sediment than the available
discharge can transport. These events will be considered in more detail later
when we examine the relationship of time in channel mechanics.
Fluvial deposition is important to geomorphology in several ways. On a
long-term basis, continued deposition results in landforms that reflect distinct
periods of geomorphic history. The stratigraphy of the associated deposits in-
dicates the types of rivers involved in the aggradational phase (Schumm 977) 1
and therefore provides clues to environmental reconstructions. On a short-term

basis, deposition creates bottom forms such as dunes, bars, and riffle-pool se-
quences that are closely interrelated with channel pattern and the character
and distribution of flow within the channel (for example, see Schumm et al.
1982). These relationships are important parts of river mechanics and will be
discussed again in later parts of this chapter. FinalK. you should recognize
that the short- and long-term mechanics of deposition have implications be-
yond the boundaries of geomorphology. They are clearly basic to sedimen-
tology and stratigraphy and, interestingly, may be key factors in subdisciplines
of economic geology such as exploration for valuable placer deposits (Schumm
1977).
The Frequency and Magnitude of River Work

At this juncture we can logically ask when and how fluvial work is done. Is it
the superevent of very high discharge that happens once in a millennium that
causes rivers to do what they do, or
is it the normal flow that is repeated time
and time again? The answer probably depends on one's perception of geo-
morphic work.
Geomorphic work is usually estimated in one of two ways. First, Wolman
and Miller (1960) suggest that the work done by a river can be estimated by
the amount of sediment it transports during any given flow. They concluded
that in most basins 90 percent of the total sediment load (i.e., 90 percent of
the work) is removed from the watershed by the sum of rather ordinary dis-
charges that recur at least once every five or ten years. There seems to be little
argument concerning that conclusion. It can be justified theoretically and by

measurement; for example, loads transported during some very large floods
(Stewart and LaMarche 1967; Scott and Gravlee 1968) represent a smaller
percentage of the total load removed from their basins than the normal flows
of high frequency and low magnitude.
The second way to estimate geomorphic work, with perhaps greater im-
plication, is to assess the conditions under which rivers make adjustments that
control or maintain their channel morphologies. Primarily on the basis of their
sediment analyses, Wolman and Miller (1960) also suggested that river chan-
nels form and reform within a range of flows between a lower limit set by the
demands of competence and an upper limit where flow exceeds bankfull and
is no longer confined to the channel. That is, channel configuration itself is
probably related to high-frequency discharge, and its precise character is pre-

sumably an indication of river work. This hypothesis also received consider-
able support and, indeed, was reinforced in studies which suggested that
channel morphology is adjusted during flows having a recurrence interval of
between 1.1 and 2 years, i.e., bankfull discharge (Kilpatrick and Barnes 1964;
Dury 1973). Therefore, the discharge that determines the characteristics and
dimensions of a channel is known as the dominant discharge, and its frequency
and magnitude was implicitly accepted to be bankfull discharge.
It seems justified to say that river channel morphology is maintained in
all environmental settings by geomorphic work done during a dominant dis-
charge or within a distinct range of flows. The real question, however, is whether
bankfull discharge is the dominant discharge for all rivers. For example, Harvey
and his colleagues (1979) found that river flows in northwest England redis-
tributed bed material between 14 and 30 times a year and changed overall
channel form from 0.5 to 4 times a year. In addition, the effect of major floods
on channel and floodplain configuration seems to vary with the environmental
setting (Costa 1974; Gupta and Fox 1974; Baker 1977; Moss and Kochel 1978).
—
Chapter 6 222
This prompted the suggestion (Wolman and Gerson 1978) that the Wolman-
Miller principle should be modified to include factors that control the work
of floods in different environments.
One of the major factors controlling flood effect is recovery time, essen-
tially the time needed for a river to recover its equilibrium form after a major
flow event has disrupted the channel configuration. In humid climates the re-
covery time is relatively short, perhaps 1-20 years. In semiarid to arid cli-
mates, however, recovery periods tend to be much longer. The significance
here is that the dominant discharge cannot have a recurrence interval shorter
than the recovery time. This follows because the effects of one flood would not
be removed (and the channel returned to presumed equilibrium form) be-
its
fore another channel-modifying flood occurred. Thus, river channels in humid

climates may be controlled by flows of intermediate frequency and magnitude
(bankfull to 10-year flood), but it seems likely that larger discharges are more
important in the work of arid or semiarid rivers.
The Quasi-Equilibrium Every river strives to establish an equilibrium relationship between the dom-
Condition inant discharge and load by adjusting its hydraulic variables (e.g., channel
width and depth, velocity, roughness, water slope, etc.). This normal fluvial
condition has been aptly referred to as a "quasi-equilibrium" state (Leopold

and Maddock 1953; Wolman 1955) because the flow variables are mutually
interdependent, meaning that a change in any single parameter requires a re-
sponse in one or more of the others. The difficulty involved in understanding
riversbecomes evident when you consider that discharge and load are in con-
tinuous flux, and so all the hydraulic variables must always be adjusting. Ob-
viously equilibrium as a steady-state condition cannot be attained in a river
thus the term quasi-equilibrium.
Hydraulic Geometry
The quasi-equilibrium condition was first demonstrated in a landmark study
by Leopold and Maddock (1953). Using abundant flow records compiled at
gaging stations throughout the western United States, they set out to deter-
—
mine the statistical relationship the hydraulic geometry between discharge —
and other variables of open channel flow. Because every river has wide fluc-
tuations in discharge, any given channel cross section must transport a range
of flows that come to it from the adjacent upstream reach. Discharge, there-
fore, serves as an independent variable at any station, and the changes in width,
depth, velocity, or other variables can be observed over a wide spectrum of

discharge conditions (fig. 6.10). At a station each of the factors (w. d. v) in-
creases as a power function such that
w = aQ b
d = cQf
v = kQ™,
Figure 6.10.
Hydraulic geometry relationships
of river channels comparing
variations of width, depth,
velocity, suspended load,
roughness, and slope to
discharge at a station and
downstream, (From Leopold and
Maddock 1953)
EXPLANATION
Change downstream for
discharge of given frequency
Discharge —
Change at gaging station for
Note: All scales are logarithmic discharges of different frequencies
where a, c, k, b, f, and m are constants. The exponents b, f, and m indicate

the rate of increase in the hydraulic variable (w, d, v) with increasing dis-
charge. Because discharge (0 equals the product of width, depth, and ve-
locity, the relationship can be expressed as
Q = aQb X cQf X kQm

or
Q = ackQ h+ f +m
and it follows that (a c k) and (b + / + m) must each equal 1. Leopold
•
and Maddock found that the average at-a-station values of b, f, and m for a
large number of midwestern and western streams were .26, .40, and .34, re-
spectively. The exponent values, however, will differ with climate and geology,
Chapter 6 224
and average values will not fit any particular stream. Essentially the at-a-sta-
tion exponents tell us what portion of the increase in discharge will be caused
by an increase in each of the component variables.
As discussed earlier, discharge also increases with the expansion of
drainage area, and so on most rivers, it must increase downstream. The ques-
tion is how much downstream increase in volume is due to each of the
of the
variables of width, depth, and velocity. To make this analysis, care must be
taken to ensure that the variables are measured during the same flow condi-
tions. On a given day, for example, a disastrous flood with high w, d, and v
values may be occurring in an upstream reach while flow conditions far down-
stream are normal. A comparison of the hydraulic variables in these two widely
divergent frequencies of flow would be misleading. Obviously the frequency
of the dischargemust be considered for any observations to be valid.
Leopold and Maddock (1953) found that the mean annual discharge is
equaled or exceeded about 20 to 25 percent of the time on a large number of
rivers. The mean annual discharge may not be the flow that determines the
character of the channel; its absolute value is relatively low, being equaled or
exceeded one out of every four days, and the water level is well below the
bankfull stage. Nonetheless, its makes it a
availability in published reports
useful parameter to determine downstream hydraulic geometry. In sum, at-
a-station and downstream hydraulic geometry differ in that one (at-a-station)
compares flows of vastly different frequencies while the other (downstream)
analyzes variables at the same frequency of Q even though the absolute values
i\ in cfs units are different.
Width, depth, and velocity increase downstream with increasing mean an-
nual discharge (fig. 6.10). The average values of b,f, and m for western streams
are and .1, respectively, but they vary from region to region and for any
.5, .4,
particular stream within a region. In general, the rate of change in depth if)
is relatively consistent in both downstream or at-a-station geometry; values
seem to range from .3 to .45. Width usually increases much more rapidly and
with more consistent values downstream than at a station, probably because
the b value for any cross section depends more on bank cohesiveness than on
discharge (Wolman and Brush 1961; Knighton 1974; Williams 1978). Ve-
locity increases more rapidly at a station than it does downstream.
The nonchalance of the statement that velocity increases downstream in-
dicates how much fluvial geomorphology has changed since the Leopold and
Maddock study. The suggestion that mean velocity, and probably bed velocity
(Leopold 1953), increase downstream came as a shock to most geologists, who
intuitively "knew" that small tributaries flowing on steep slopes must be trav-
eling faster than the low-gradient trunk rivers. The surprise at this new inter-
pretation of velocity was probably due mainly to geologists' inclination to

consider slope as the major, if not the overriding, control of velocity. The im-
portance of slope seems to be entrenched in geologic thinking and emerges in
our basic concepts. Further, slope has always been involved in the interpre-
tation of sediment transportation and deposition, and it usually is the easiest
parameter to reconstruct for an ancient fluvial regime. For example, the sur-
face of coarse-grained deposits such as terrace alluvium or fan debris is in
many cases the original stream bottom, and its gradient represents the channel
slope at the time of deposition. Slope, then, is a valuable tool in geologic inter-
pretation and a prime factor of fluvial processes, but as we will see later, it
cannot be singled out as the only adjustable variable in a stream channel or

even the dominant one.
The possibility of a downstream increase in velocity should have been sus-
pected because Manning's equation tells us that depth plays a greater role
than slope in determining velocity. In a stream with a constant roughness, if
depth increases downstream at the same rate that slope decreases, an overall
increase in velocity will occur. Increased depth simply overcompensates for
the loss of velocity due to a gradually decreasing slope.
All geomorphologists, however, do not accept the idea of a downstream
increase in velocity. Carlston (1969), for example, suggested that on large rivers
downstream velocity is probably constant, but on smaller streams it may in-
crease or decrease according to local controls. Mackin (1963) sounded a more

serious objection when he cautioned against the injudicious use of empirical
methods as the basis for sweeping geomorphic conclusions, citing the down-
stream velocity interpretation as a specific example. Indeed, some streams do
decrease in velocity downstream (Brush 1961; Carlston 1969), but as Mackin
points out, these individual cases may be obscured when they are placed on
the same scatter diagram with measurements taken on other streams with
different fundamental controls. The increase in velocity, therefore, may be the
result of a particular methodology; as such it represents the exposition of a
general trend rather than a conclusive rule. It is difficult to find fault with
Mackin's argument, but in defense of empiricism we must recognize that
streams do not lend themselves to a purely rational or scientific approach. The
complex interaction in a river cannot be adequately expressed in mathemat-
ical terms, and the multitude of variables that define the system are constantly
adjusting and readjusting to minor variations in flow. There may always be

an element of indeterminancy in river mechanics that is simply beyond ra-
tional comprehension (Leopold and Langbein 1963). It may be better to strive
for general trends and admit their fallibility in specific cases than to wait end-
lessly for a detailed scientific explanation.
Hydraulic geometry analyses have become standard in fluvial geomorphic

studies for making comparisons of rivers of different climates, physical set-
tings, and size (Wolman 1955; Leopold and Miller 1956; Myrick and Leopold
1963; Fahnestock 1963), even though the values are probably controlled by
local factors (Miller 1958; Hadley 1961; Wolman and Brush 1961; Thornes
1970; Knighton 1974; Park 1977). In fact, the case has been made that the
most probable channel form which adjustments of the hydraulic vari-
is one in
ables to changing discharge are minimal (Langbein 1964, 1965; Knighton 1977;
Williams 1978). This concept, known as the minimum variance theory, sug-
gests that the shape of channels is predetermined by local geomorphic con-
straints that allow adjustments to be made with the least variation in the
hydraulic variables. If no constraints are present, all variables have equal vari-
ance and the exponent values will each be 0.33. Minimum variance is an over-
simplification of reality because it only predicts the most probable exponent
values. Thus, any individual cross section may have values different from the
predicted values.
Chapter 6 226
1,000,000 _
Figure 6.11.
Relation of suspended load to
discharge in Powder River at
Arvada, Wyo. (From Leopold and
Maddock 1953)
100,000
I i 1 1 mi
100 1 000 10,000
Discharge (cfs)
The utility of hydraulic geometry in geomorphic studies has yet to be sat-

isfactorily documented. Park (1977) found that variations in sets of b-f-m
values do not even distinguish rivers in diverse climates. On the other hand.
Rhodes (1977) feels that all rivers can be categorized on the basis of various
ratios of exponent values and/or other river properties (Froude number,
roughness) that are controlled by the exponential values. The groups proposed
by Rhodes reflect basic fluvial mechanics; this suggests that hydraulic ge-
ometry should be useful in predicting how any particular river will work. For
example, one group includes rivers in which the rate of increase in velocity
(m) exceeds the combined changes in width and depth (b + /). Such rivers
should experience a rapid increase in competence with rising discharge, a con-
dition that is probably needed to entrain coarse bedload (Wilcock 1971).
In most streams the amount of suspended sediment at a station increases
rapidly with increasing discharge (fig. 6.11), varying as the simple power
function
L = pQi,
where L is suspended load and p and j are constants. The at-a-station value
of j is normally >
1, indicating that the influx of sediment to the river is
10,000
>.
o
1,000
53
Q. 500 1,000 10,000 100,000
in 5,000 50,000
c
o
Discharge (cfs)
o
CO (B)
T3
Oct. 14,1941 -59,600 cfs

Right
~ 100 Tbank
— 20
o
(D
E
T5
0)
CO
aa)
a 10
T3 CD
C
CD
Q.
(0
3 O
CO
1
nil ti I I I I
500 1,000 100,000 100 200 300

50.000
Distance (feet)
(C)
Figure 6.12.
greater than the addition of water. Interestingly, the dramatic increase in sed- Changes in (A) suspended load.
iment content is not necessarily caused by scouring of the channel floor. Sev- (B) stream-bed elevation, and
(C) water-surface elevation with
eral studies have shown that scouring can occur at different times during the
discharge during flood of
passage of a flood. It may occur at peak flow, during rising flow, or in the September-December 1941, San
waning part of the and Maddock 1953; Foley 1978; Andrews
flood (Leopold Juan River near Bluff, Utah (From
Leopold and Maddock 1953)
1979). It has also been shown that scouring and filling can occur simulta-
neously in close proximity within the same channel reach (Foley 1978; An-
drews 1979).
In those situations where scouring occurs at peak flow or during the reces-
sion phase, deposition may take place in the rising stages of the flood, precisely
when the suspended load is increasing rapidly (fig. 6.12). Scouring takes place
later in the flood when velocities and sediment loads are lower. This obser-
vation would mean that the bulk of sediment added to a river during floods
derives from the valley-side slopes of the watershed, channel bank erosion, or
Chapter 6 228
tributary input. The observation also helps explain the notable variation in
suspended load any given discharge (evident in fig. 6.1 1) because the sed-
at
iment acts as an independent variable that demands compensatory adjust-
ments in velocity and depth. Since the value of j is greater than unity, the
sediment concentration (suspended load/unit volume H 2 0) increases, thereby
lowering the internal resistance to flow, as discussed earlier. The decreased n
prompts an increase in velocity and enhances the river's ability to carry a
greater load. In this way, a high rate of increase in suspended load seems to
initiate the adjustments in hydraulic geometry that are needed to carry that
load through the section.
The Influence of Slope

It was suggested earlier that channel slope has always been recognized as a
prime adjustable property of rivers. Geologists and geographers traditionally
have carefully studied the river gradient and generally have accepted the
proposition that a concave-up longitudinal profile (change in elevation with
increasing length) is the channel form assumed by rivers in equilibrium. There
is abundant evidence to substantiate the importance of slope in a river striving
to maintain balance, but whether the gradient adjustment operates to the ex-
clusion or subservience of other variable changes is a question of considerable
debate in modern geomorphology.
Observations at gage stations show that the slope of the water surface
remains relatively constant during flows of different magnitudes, indicating
that the adjustments to increasing discharge must be made by the other hy-
draulic variables. For example, we cannot call on a dramatic increase in slope
to produce the relatively high rate of increase in velocity (pi). The increasing
velocity must be generated by an increase in depth, a decrease in roughness,
or both. Downstream the channel gradient does exert an influence, because in
most rivers there is a notable decrease in slope. Roughness, however, usually
remains fairly constant (Leopold and Maddock 1953) because of the offsetting
effects of a decrease in particle size (decreases n) and a decrease in sediment
concentration (increases ri). As a result, any increase in velocity downstream
can best be justified by the increase in depth, explaining the low m values in
that direction.
The and other hydraulic parameters reveal
relationships between slope
the complexities of quasi-equilibrium, but they do not explain why slope usu-
ally decreases downstream or what external factors may control the form of
the longitudinal profile. As early as 1877, G. K. Gilbert concluded that slope
was inversely related to discharge, and since Q increases with basin area and
stream length, it is axiomatic that slope should decrease downstream. How-
ever, in most rivers particle size generally diminishes downstream, prompting
many observers to suggest that channel gradient adjusts to the bed material.
Actually both factors are probably involved. Rubey (1952) demonstrated that
if channel shape is constant the slope will decrease with (1) a decrease in par-
ticle size, (2) a decrease in total load, or (3) an increase in discharge. Rubey
concludes that the channel gradient at any point along the river is a function
V
North River
• Limestone areas in the

Shenandoah Valley
Areas of Wissahickon of Jonas
1928 (Gillis Falls)
a Areas of Martinsburg shale

Areas of Brandywine formation
(Coastal Plain)
Areas of Devonian sandstone

* Granodiorite area (Tye River)
+ Calfpasture River
x North River and Dry Branch in

alluvial terrace areas
l
I I I I II I i i i i i i
20 40 60 100 200 400 600 1000 2000 4000 600010,000

Median diameter of bed material (millimeters)
Figure 6.13.
of both sediment and discharge. If Rubey is correct, then slope is dependent, Scatter diagram showing relation
or partially so, on all hydraulic variables because they are also related to dis- between channel slope and
median size of bed material in
charge. selected rivers Maryland and
Many studies subsequently have shown the correctness of Rubey's anal- Virginia (From Hack 1957)
ysis. In one of these studies, comparing stream profiles in areas of differing

geology, Hack (1957) found no consistent correlation between slope and bed-
material size when all sample localities from a geologically divergent region
were considered together. His plot, shown in figure 6.13, reveals a tremendous
scatter; for example, streams with a median particle size of 60 mm have slopes
ranging from 1.1 m/km (6 ft/mi) to 37 m/km (200 ft/mi). It was only after
Hack added a third variable, drainage area, to the analysis that a significant
relationship became apparent (fig. 6.14), and slope could be defined by the
equation
- «
m"
Chapter 6 230
1000
i
800 =
• Limestone areas in the

Shenandoah Valley
• Areas of Wissahickon of Jonas,
1928 (Gillis Falls)
Areas of Martinsburg shale
» Areas of Brandywine formation
(Coastal Plain)
Areas of Devonian sandstone
* Granodiorite area (Tye River)
+ Calfpasture River
* North River and Dry Branch in
40
i i
60
i i 1
100
I
areas
alluvial terrace
1 l 400 600
I i
1000
Ratio, median size of bed material to drainage area
Figure 6.14.
Relation between slope and the where M is the median size of the bed material in millimeters, A is area in
ratio of median size of bed mi 2 and
, S is slope in ft/mi. Since basin area can normally be used as an index
material to drainage area in
selected streams, Maryland and of discharge (Leopold et al. 1964), Hack's study reinforces Rubey's contention
Virginia See figure 6 13 for that both Q and sediment are determinants of slope. It does not indicate which
comparison (From Hack 1957) factor is the principal determinant; indeed, one would expect the relationship
to be defined by different mathematical equations in different physical set-
tings. Nonetheless, Hack demonstrated that streams flowing within a single
geological unit and having the same drainage area should have similar channel
slopes and particle size.
On a local scale, slope is less dependent on discharge and more closely
and bed characteristics. For example. Prestegaard (1983b)
related to sediment
found that components of bed roughness (particle size and bottom configu-
ration) are the major determinants of water-surface slopes in a number of
gravel-bed streams. In rivers such as those, particle size has an important in-
fluence on slope over a range of distances from local (one to three times the
channel width) to an entire reach (100 m-300 m). Bed configuration exerts
an influence only on a scale relating to the entire reach length.
Before assuming that slope seems to be understandable and possibly even

predictable, note once again that Rubey's conclusions were based on the as-
sumption of a constant channel shape. Such an assumption is probably more
than nature will allow because in most alluvial rivers the channel shape is
subject to change. Thus, although slope depends on both sediment and dis-
charge, fluctuations in those variables do not demand an adjustment in the
channel gradient. It is equally possible that changes may be counterbalanced
by variations in the cross-sectional shape of the channel.
Channel Shape
Logic tells us that, unless velocity is completely unrestrained, rivers with a
large mean annual discharge have greater cross-sectional areas than streams
with smaller average flows. This fact has been verified repeatedly by casual
observation and documented by the relationships exposed in hydraulic ge-
ometry. It is significant, however, that many rivers with the same mean annual
discharge have different cross-sectional areas; even when the total area is the
same, the width-depth ratio may vary considerably. Obviously factors other
than discharge alone influence the shape of river channels, and so we look to
the channel sediment for an explanation.
Schumm (1960) presented cogent arguments to suggest that channel
shape, as defined by W/D, is determined primarily by the nature of the sedi-
ment in the channel perimeter. Where perimeters have a high percentage of

siltand clay (particles < 0.074 mm), channels tend to be narrow and deep.
In contrast, wide, shallow channels seem to be characteristic of rivers having
coarse-grained perimeters. Schumm's data, collected from semiarid and arid
climate streams, show that W/D is related to the percent of silt-clay by the
equation
F = 255M-' 08 ,
where F is the width-depth ratio and M is the percent of silt and clay in the
channel perimeter. The magnitudes of the mean annual discharge or the mean
annual flood do not seem to affect this relationship; in fact, only 40 percent
of the variability in channel shapes can be accounted for by discharge alone
(Schumm 1971).
In streams that are not aggrading or degrading, the channel shape is also
related to the load being transported. Bedload is more efficiently transported
in a wide-shallow channel because higher velocities are nearer the channel

floor(Lane 1937). Suspended loads are carried best in channels having lower
width-depth ratios (narrow, deep channels). This prompted Schumm (1963a)
to classify rivers on the basis of their load types; more important, it demon-
strates another viable method by which a river can be adjusted. A channel
reach suddenly burdened with a different type of load than the one it previ-
ously carried may as easily alter its channel shape to accommodate the new
load as change its gradient by deposition or erosion. The repeated changes in
the shape of the Cimarron River in Kansas since 1880 are an excellent ex-
ample of this phenomenon (Schumm and Lichty 1963).
Chapter 6 232
The responses envisioned by Schumm are based on relationships deter-

mined in semiarid to arid climates. In other climates the sediment control on
channel characteristics may be different from those noted above (Baker 1978).
Therefore, widespread application of the equations without consideration of
climates is not encouraged.
Factors other than the character of sediments can affect the cross-sec-
tional shape of channels. It has been shown that root systems of riparian veg-
etation exert a drastic decrease in bank erosion and therefore control the
channel width (Smith 1976). In contrast, large trees that fall into a channel
may bank erosion by diverting flow around a tree jam into an un-
increase
protected bank (Keller and Swanson 1979; Keller and Tally 1979). In arid
regions sediment by precipitation of CaC0 3 may increase bank cohesion and
thereby prevent expansion of channel width (Van Arsdale 1982).
Channel Patterns Our discussion of adjusting mechanics in rivers has thus far centered on the
balancing factors that function in a channel cross section or a series of cross
sections considered in a downstream direction. Rivers also have characteristic
forms extending over long stretches of their total length which, when observed
in plain view, display a distinct geometric pattern. The pattern that a river
adopts is now recognized as another manifestation of channel adjustment to
the prevailing discharge and load. In that sense, channel patterns are another
important aspect of river mechanics.
Patterns are usually classified as straight, meandering, or braided, even
though the boundaries between types are sometimes arbitrary and indistinct.
The distinction between straight and meandering, for example, is based on a
property called sinuosity. Although there is no complete agreement on how
sinuosity should be determined (Leopold and Wolman 1957; Schumm 1963b:

Brice 1964), we will follow Schumm's definition that it is the ratio of stream
length (measured along the center of the channel) to valley length (measured
along the axis of the valley). The transition between a straight and a mean-
dering stream is usually placed at a sinuosity value of 1 .5, but again this value
probably has no particular mechanical significance.
The braided pattern, characterized by the division of the river into more
than one channel, is more easily discerned. The designation becomes vague,
however, when only part of the river's total length is multichanneled. How
much of the river must consist of divided reaches to constitute a braided system
is an individual decision. Another complication is that some single-channeled
riversbecome distinctly braided in times of high flow, requiring that stage be

considered when deciding on the pattern classification. In addition, there is no
reason to expect that a river will display the same pattern for its entire length.
In fact, because patterns are probably a function of discharge and load, minor
variations in those factors downstream may easily generate different patterns.
especially in very large watersheds.
Straight Channels
Most streams do not have straight banks for any significant distance, making
the straight pattern a rather uncommon one. It may seem strange, then, that
straight streams display many of the same channel features as the more
common meandering pattern. As figure 6.15 illustrates, straight reaches often
contain accumulations of bed material called alternate bars that are posi-
tioned successively downriver on opposite sides of the channel. A line con-
necting the deepest parts of the channel, called the thalweg, migrates back
and forth across the bottom. In rivers with a poorly sorted load, the channel
floor undulates into alternating shallow zones called riffles and deeps called
pools.The pools are directly opposite the alternate bars, and the riffles are
about midway between two successive pools. Clearly, a straight channel im-
plies neither a uniform stream bed nor a straight thalweg, and the spacing of
bars, riffles, and pools is closely analogous to that in a meandering channel
(Leopold et al. 1964; Dury 1969).
Sequences of pools and riffles are very important manifestations of how
bedforms, flow, and sediment transport are interrelated in rivers to maintain
quasi-equilibrium. This is true regardless of channel pattern, indicating that
the tendency to develop bars and/or pool and riffle sequences must result from
some fundamental property of moving water. Pool-riffle sequences have dis-
tinct spatial characteristics that are largely independent of the material type
forming the channel perimeter, even if the channel is cut in bedrock (Keller
and Melhorn 1978). Successive riffles in straight channels are usually spaced
about five to seven channel widths from one another and, in alluvial channels,
Figure 6.15.
are composed of more coarse-grained sediment than the intervening pools. Rif- Features associated with
fles are also wider and shallower than pools at all discharges (Richards 1976a). (A) straight and (B) meandering
rivers T = thalweg; B = bar:
At low discharge, riffles are characterized by rapid flow and steep water
= =
R riffle; P pool.
surfaces; in contrast pool velocities are low and surface gradients are gentle.
Assuming that coarse particles cannot be moved under the low flow conditions
existing in the pool, logic tells us that the pools should be destroyed by ag-
gradation when sediment moves downstream from adjacent riffles. Conse-
quently, the question arises as to how the riffle-pool sequence is continuously
maintained. Actually, it has been demonstrated (Keller 1971; Richards 1976b;
Andrews 1979) that as discharge increases, the rate of increase in bottom
velocity and mean velocity is greater in the pools than in the riffles. Thus, a
velocity reversal occurs at some particular discharge, above which the pool
velocity is greater than the riffle velocity (fig. 6.16). As a result, coarse par-
ticles entrained from the riffles during rising discharge can be moved through
the pool when discharge exceeds the point of velocity reversal. In the reces-
sional phase, the largest particles are deposited on the riffles while discharge
is still above the velocity reversal. Ultimately, fine sediment is eroded from
the riffles and trapped in the adjacent pools, leaving a coarse-grained lag on
the riffle and a thin, fine-grained deposit in the pool. Thus, pools and riffles
seem to be formed and maintained by scouring (pools) and deposition (riffles)
Chapter 6 234
Figure 6.16.
Rate of increase in mean velocity
over a pool and riffle with
increasing discharge East Fork
River, Wyoming (After Andrews
1979)
100 500 2000

Discharge (cubic feet per second)
Figure 6.17.
Schematic diagram of convergent
flow and secondary circulation
over a pool (A) and divergent flow
and secondary circulation over a
riffle (B) in a straight channel
under conditions of relatively high discharge. Presumably this discharge oc-

curs with moderate frequency and, therefore, is probably not associated with
flooding.
Pool-riffle distributions are somehow related to the fundamental me-
chanics of open-channel flow that produce secondary water movements. The
secondary motion occurs as cells circulating in planes that are transverse to
the normal downstream component of flow (Leliavsky 1966: Leopold 1982).
The direction of cell circulation over the pools is the opposite of that operating
over the riffles (Keller and Melhorn 1973; Richards 1982). and the surface
portion of the cells can be noted as zones of converging or diverging flow. In
convergent flow, surface water is accelerated, and the maximum velocity oc-
curs near the bottom because isovels in the water are depressed. This facili-
tates scour (fig. 6.17). In divergent flow, surface water tends to spread outward
toward the banks. In this case bottom velocities are retarded and deposition
is common (fig. 6. 1 7). Significantly, convergent flow is common over pools and
divergent flow is normally located over riffles (fig. 6.17).
Meandering Channels
The most common river form by far is the meandering pattern, also shown in
figure 6.15. Meandering reaches contain the same physical components ob-
served in straight channels (pools, riffles, bars), distributed in a similar way.
The thalweg also migrates back and forth across the channel, impinging against
the outer bank of the meander bends and crossing to the opposite side near
the riffles.
Secondary circulation is a prime property of flow in a meandering system.

Presumably, as water is forced against the outer bank of a meander, its slightly
elevated level at that point gives the flow a circulating motion. The water moves
along the surface toward the undercut bank and along the bottom toward the
point bar (fig. 6.18). This corkscrew motion, called helical flow or helicoidal
flow, has traditionally been thought of as a single rotating cell that reaches its
greatest velocity slightlydownstream from the axes of the meander bends at

the position of the pools; the velocity decreases gradually downstream until
the flow approaches or reaches the next riffle and the lateral circulation dis-
appears. From this transition zone to the next downstream meander, the spiral
velocity increases again, but the cellular pattern is opposite to its original di-
rection because the orientation of the meander itself has been reversed (fig.
6.18). The magnitude of the lateral velocity varies but it can be great enough
to influence the transport of sediment. Particles eroded from an undercut bank,
for example, may be dragged into the center of the channel by the bottom
limb of the helical cell. When the motion reverses at the transition, some of
the sediment is returned to its original channel side and deposited on the next
downstream point bar.
We now know that this model of helical flow is probably oversimplified. Figure 6.18.
Traditional model of helicoidal flow
For example, several researchers (Hey and Thorne 1975; Bathurst et al. 1979) at successive bends in a
have observed two helicoidal cells at the meander apex and at the points of meandering river. P represents a
pool and R represents a riffle.
inflection (riffles) in several British rivers. At the meander apex, the two ro-
tating cells meet at the surface in a zone of convergence that promotes scouring
close to the undercut bank (fig. 6.19). At the transition, the flow diverges over
the position of the riffle (fig. 6.19) and tends to induce deposition in that zone.
This suggests that the polarity of the rotating cells must change between the
meander axis and the next downstream inflection point. Such a distribution
requires that an intermediate zone of no secondary circulation must be present
where one type of helicoidal movement is dissipated and another begins (fig.
6.19). The precise relationships between helical flow and sediment transport
are, therefore, still to be understood. The data needed to make such an analysis
are very difficult to collect and require detailed field measurements (Bhowmik
1982).
Meandering rivers shift their positions across the valley bottom by eroding
on the outer banks of meander bends and simultaneously depositing point bars
on the inside of the bends. Even though the location of the river varies with
time, there is no compelling reason to suggest that the shape or hydraulic
properties of the river stray far from average values as long as the prevailing
controls of climate and tectonics remain unchanged. In fact, meanders in rivers
of all sizes are dimensionally similar, with consistent geometric and hydraulic
relationships as in table 6.3 and figure 6.20.
Chapter 6 236
Figure 6.19.
Model showing two secondary
and surface
cells flowlines in a
meandering river.
Figure 6.20.
Geometric parameters of a
meander X = wavelength;
A = amplitude: rm = radius of
curvature
Table 6.3 Empirical relationships between parameters that define meander

geometry
Dependent Relationship Source
Wavelength
X = 66 w099 Inghs (1949)
X = 10 9 w 101 Leopold and Wolman (1957)
X = 4 7 rm098 Leopold and Wolman (1957)
X = 30 0„P S Dury (1965)
Amplitude
A = 186 w099 Inghs (1949)
A = 10 9 w' 04 Inghs (1949)
A = 2 7 w' 1
Leopold and Wolman (1957)
wavelen.;' 'ipiitude (III
width (It) at banklull stage inktull discfi.i-

radius ol
'\
Figure 6.21.
Relation between radius of
curvature, width, and flow
properties Decreasing ratio m/w
A
in (B) causes flow to break away
from inside of bend and create
eddying zone (C and D). R = rm
(After Bagnold 1960)
R/w = 2.0 0\
(C)
The most revealing geometric property is the meander wavelength, which

relates to many other variables including discharge, width, and the radius of
curvature (rm ). The relationships are in many cases almost linear and un-
doubtedly reflect basic mechanical principles. For example, using the equa-
tions in table 6.3 and considering them as linear, we find that where all
parameters are measured in feet,
4.7 10.9
= 2.3.
w 4.7
10.9
Actual measurement of this ratio shows most rivers to have values between 2
and 3 (Leopold and Wolman 1960), suggesting that the relationship probably
is a function of channel curvature exerting an influence on flow. Bagnold ( 1 960)
showed that as the radius of curvature decreases decreasing

—J, the main I
filament of flow tends to shift toward the outer bank, causing a concomitant
decrease in resistance on the inside of the bend. Greater curvature will con-
tinue to decrease resistance until a critical value of —

f
w
is attained, when flow
along the inner bed becomes unstable and breaks away from the boundary
(fig. 6.21). This creates eddy currents along the inside boundary, increasing
the energy dissipation and so effectively establishing a minimum resistance
for the flow. Inmost fluid systems, eddying begins when the curvature ratio
is between 2 and 3, suggesting that the large number of real meanders having
these values probably represents a quasi-equilibrium between flow and ge-

f
ometry. Hickin (1974) found the critical -Rvalue to average 2.11 on the
w
Chapter 6 238
Beatton River in British Columbia. He demonstrated that once a developing

meander attains the critical curvature, it exerts significant control on the sub-
sequent rate and direction of lateral migration.
Before the recent developments in fluvial geomorphology, most geologists
accepted the premise that random diversions of flow by slumped boulders or
fallen trees were the ultimate causes of meanders. Such aberrations undoubt-
edly can start meanders, and once an initial bend develops, the sinuous nature
is transmitted downstream and causes more bends to form (Friedkin 1945).
Helical flow is also propagated from the initial curve. Callander (1969) showed
that any deflection of downstream flowlines on the river surface creates hy-
draulic instability because secondary flow is initiated. Sediment on the channel
floor begins to move toward the inner bank. This results in the formation of
an incipient bar, thalweg migration, and scouring of the bottom. In addition,
two circulating cells are present at this time (Leopold 1982). Thus, it is clear
meandering are already present in straight streams.
that the ingredients of
Although it may be tempting to write off meandering as the result of
random perturbations of flow direction, the change from straight to mean-
dering requires that other factors operate within the system. First, Parker
(1976) showed that secondary flow alone is not the cause of meandering. The
instability associated with the beginning of helicoidal cells will not culminate
in a meandering pattern unless a second process triggers the action. In alluvial
rivers, the necessary process is bedload transport. Second, the meandering pat-
bank erosion. The bank erosion must be a local
tern will not develop without
phenomenon rather than the widespread retreat of both banks, a process that
would lead to creation of a wide-shallow channel.
Third, the development of a meandering pattern requires certain energy
adjustments within the flow. Bank erosion and flow around curves dissipate
available energy, and therefore the transition from a straight to meandering
pattern implies that energy must be consumed in a different way. Most anal-
yses of river energy indicate that meandering streams are probably closer to
an equilibrium condition than straight streams because (1) meandering tends
to dissipate energy in equal amounts along the length of the channel and
(2) under the constraints of (1), meandering tends to minimize the total en-
ergy expenditure (to do the least work) or the rate of energy expenditure. This
is accomplished by adjusting the curvature geometry or the channel gradient.
If a straight stream is prevented from making any of the above adjust-
ments, it is doubtful that it can change into a stable, meandering pattern. For
example, Schumm (1963b, 1967b) showed that sinuosity decreases and
meander wavelength increases when a river transports sand and gravel rather
than a fine-grained suspended load. Presumably this results because the avail-
able river energy must be used to entrain and transport bedload. and therefore
the energy adjustments coincident to meandering (such as lowering slope)
cannot be made.
Braided Channels
A basic part of the formation of the braided pattern is the division of a single
trunk channel into a network of branches (fig. 6.22) and the growth and sta-
bilization of intervening islands. Detailed studies of braided systems in both
flumes and rivers (Leopold and Wolman 1957; Fahnestock 1963; Church 1972;
Rust 1972; N. D. Smith 1970, 1974) show that divided reaches have different
channel properties than adjacent undivided segments. Braided zones are usu-
ally steeper (fig. 6.23) and shallower; total width is greater although each
channel may be narrower than the undivided trunk; and changes in channel
positions and the total number of channels are likely to be extremely rapid
(Fahnestock 1963; Church 1972; N. D. Smith 1974). Fahnestock, for ex-
ample, documented lateral shifting of the channels up to 122 meters (400 ft)
in eight days in the braided segment of the White River in Washington.
The Origin of Braids Any explanation of the origin of braids is necessarily

oversimplified because, like all fluvial processes, it involves the simultaneous
interaction of a number of factors. According to Fahnestock (1963), the most
important of these are the following:
1 . Erodible banks. Most investigators of channel patterns feel that
bank erosion is perhaps the most necessary factor in creating a
braided system. If bank erosion is prohibited by material

Figure 6.22.
cohesiveness or vegetation, it is unlikely that a braided pattern will
Photograph of typical braided
develop. stream pattern. McKinley River,
Alaska
Chapter 6 240
2. Sediment transport and abundant load. Almost every braided river

transports large volumes of bedload, and much of the channel
shifting is prompted by temporary deposition of bars across
entrances to branches of the network. It is incorrect, however, to
assume that braided rivers are overloaded, since many times braids
form when the channel is actively being eroded (Leopold and
Wolman 1957), and at the same time, undivided segments
immediately downstream from an evolving braided reach are being
actively aggraded. Even though the load exceeds the transporting
capacity of the channel, no braiding occurs.
3. Rapid and frequent variations in Q. Fluctuations in discharge tend
to produce the alternating erosion and deposition that seem to be a
necessary part of braiding mechanics. However, some of the
channel shifting and much of the increase in the number of
channels may simply be a matter of reoccupying abandoned
channels during high river stages. It is also significant that
laboratory studies have produced braids under constant discharge,
indicating that discharge may sustain the pattern, not cause it.
Although slope and other channel properties of braided streams are dif-
ferent from those of meanders, they probably are not factors in the origin of
braids. More likely they represent the geometric modifications brought about
by particular sediment and discharge requirements. For example, Parker
(1976) emphasized that slope and the width-depth ratio are important man-
ifestations of the controls leading to braiding versus meandering. If slope and
w/d are high under conditions of dominant discharge, the pattern will prob-
ably be braided. In addition, braids result from a river's tendency to form bars
(Parker 1976), but the type of bars developed may be a significant factor in
maintaining a braided pattern (Church and Jones 1982). Therefore, it seems

safe to say that braids do not necessarily connote instability. The pattern simply
represents another condition a river may establish in response to external con-
trols. It may be maintained for a long period of time and possibly is as close
to true equilibrium as the meandering pattern.
Experimental studies suggest that braid development follows a distinct
sequence of events, illustrated in figure 6.23 (Leopold and Wolman 1957).
During high flow a portion of the coarse load being transported is deposited
because of some local channel condition. This initial accumulation becomes
the locus of an incipient longitudinal bar because reentrainment of the par-
ticles requires a greater velocity than did their transportation and deposition
(fig. 6.7).Continued deposition here allows the bar to grow both upward and
in the downstream direction. As particles move across the reach, they are de-
posited on the lower end of the bar where depth suddenly increases and ve-
locity decreases. These changes occur because the width and discharge remain
constant over that segment of the cross section, and since q = wdv, an increase
in depth requires a decrease in velocity. Most smaller particles move easily
over the growing bar, but some may be trapped in the interstices between the
larger grains.
—
Successive cross sections

Station Station
at station 14
22
I
- 18
i
14
r
10 22 18 14 10
Water surface
I
0.25
o CD
> 2 -0.20
c
r
a> E
-0.15
o j;
a)
o
c -0.10
o Water surface
o
> -0.05
o 2
^ TJ CO -
UJ
> C3 0.25-
6 hours 22 hours 4 hours

EXPLANATION Water surface
Deposit coarser than original
sand
.\ Deposit finer than original 0.25
'"
sand
Island or area nearly out of
10 hours
water
- > Path of principal bed transport Water surface
Riffle, or water flowing in
steep, thin, sheet

0.25
Well-defined edge of bar
9 hours Ill-defined edge of bar

18 hours
_L
1 2 feet
(A) (B)
Figure 6.23.
As the expanding bar begins to occupy a significant portion of the channel Stages in the development of a
area, the channel is no longer wide enough to contain the total flow, and as braid in a flume channel.
(A) Sequential development of the
flow is deflected around the bar, the banks are eroded. Simultaneously, the bar pattern at various times (B) Cross
itself may be trimmed and the channel somewhat deepened. These processes sections at one station along the
combine effectively to enlarge the channel on both sides of the bar, allowing flume. (From Leopold and Wolman
1957)
the water level to be lower at any equivalent discharge. Eventually the bar
emerges as an island flanked by two distinct channel branches. Bars and is-
lands do not necessarily remain fixed in position or shape since they are also
susceptible to lateral erosion (N. D. Smith 1974). In documented cases, how-
ever, vegetation may spread rapidly on the islands, especially if overbank flows
or wind provide silt as a capping layer; the rooting of vegetation tends to in-
The whole process from initiation to stabili-
crease the resistance to erosion.
zation may two years. Bar growth can also begin when a river
take as little as
reoccupies an abandoned channel. In such an event, an erosional remnant may
become a core upon which new bar growth will develop (Eynon and Walker
1974).
Chapter 6 242
Other bar types also are common in braided systems. V D. Smith (1974)
recognizes transverse, point, and diagonal bars in addition to the longitudinal
type. Transverse bars are tabular bodies that grow by downstream migration
of foreset beds developed more or less perpendicular to the current. These bars
form when sand moving along the bottom encounters a shallow depression
where velocity is lowered (Jopling 1966). In response, the sand is deposited
as a "delta" that builds upward former
until the flow velocity increases to its
level, and the top of the bar becomes the channel which sand can
floor across
once again be transported. At low flow, transverse bars may be exposed and
dissected into a series of small channels.
Smith (1971) noted significant differences between longitudinal and
transverse bars in the Platte River in Nebraska. Transverse bars normally are
composed of sand and are better sorted andmore fine-grained than longitu-
dinal bars. In addition, they show well-developed planar cross-beds in contrast
to the crude horizontal stratification found in the coarse sediment of the lon-
gitudinal bars (Rust 1972). A more recent analysis of the macroforms (bars)
and the meaning of their sedimentary structures is given in Crowley (1983).
Anastomosed Channels Another type of multichanneled river is an anasto-

mosed stream. Historically, the terms "anastomosed" and "braided" have been
used interchangeably, but this practice should be discontinued because the
normal characteristics of the two patterns and their formative mechanics are
considerably different. Smith and Smith (1976) define an anastomosed river
as one having an interconnected network of low-gradient, relatively deep and
narrow channels that have variable sinuosities and stable banks composed of
fine-grained sediment and vegetation. The significant point here is that these
rivers have some characteristics of both meandering and braided streams but
overall are distinct from each of those patterns. For example, the pattern is
multichanneled, indicating a relationship with braiding; however, the banks

are not erodible, which as indicated above, is a prime requisite in the for-
mation of braids.
The essential ingredients needed to form anastomosed rivers are (1) rapid
aggradation caused by downstream base level rise and (2) stable, cohesive
banks that retard channel migration. Under these controls, floodplain surfaces
are rapidly elevated by overbank deposition. It is also possible that gravel may
simultaneously fill the channel (Smith and Smith 1976).
The Continuity of Channel Patterns

If channel patterns reflect specific adjustments to fluvial variables, it follows
that boundaries between the patterns should be definable in terms of those
variables. It also follows that each pattern must be stable within certain
threshold limits of the controlling factors. When the limits are exceeded, a
viable fluvial response would be a pattern change. Patterns, then, may be
ephemeral fluvial properties, especially in segments where the values of the
controlling factors are critically close to the threshold condition.
As we saw earlier, straight reaches in suspended- load rivers are rare enough
to be considered as unstable and probably transitional to the meandering form.
This observation is supported by energy analyses and by the fact that the phys-
ical components of straight reaches are analogous to those in meanders. For

example, in straight reaches the spacing of successive riffles is about five to
seven times the channel width. In meanders the relationship
X = 10.9w 101 ,
where X and w are measured in feet, suggests that the two successive riffles
found in one complete wavelength are spaced about the same as those in straight
reaches. Straight reaches, therefore, will probably not remain long in that form
unless the banks are unerodable or a coarse-grained load requires the river to
use all its energy for transportation, leaving none to be dissipated in meander
bends. Even so, special cases may exist. Distributaries in deltas, for example,
are straight and perhaps remain so through a combination of very low gradient
and very low bedload values. In addition, there is no compelling reason to
expect that meandering, once achieved, is forever stable. The mechanics that
produce meandering also cause chutes and cutoffs that return the reach to
straightness (Keller 1972; Lewin 1976), or they proceed to multilooping (Brice
1974) or other complexly woven forms (fig. 6.24).
Figure 6.24.
Complex of loops and abandoned
channels in strongly meandering
river pattern. Rio Grande in
3
Alamosa County, Colo
Chapter 6 244
Figure 6.25.
Relation of slope and discharge.
Lines represent threshold slopes
discharges as
at various
determined in different studies
100
Discharge, cfs
The threshold marking the transition between meandering and braided is
perhaps more precisely, although empirically, defined. Several studies have

suggested a threshold boundary between the two forms based on the relation-
ship between slope and discharge (Lane 1957; Leopold and Wolman 1957;
Ackers and Charlton 1971). Although the limiting values are not consistently
the same (fig. 6.25), it seems obvious that at the threshold an increase in slope
atany given discharge (or an increase in discharge at any given slope) will
change a meandering pattern to a braided pattern. It has also been demon-
strated (Schumm and Khan 1972) that the threshold values between patterns
can be defined as well, if not better, by the slope-sediment load relationship.
This suggestion was reinforced by Osterkamp ( 978), who found that in many
1
Kansas streams the lowest gradient of the braided pattern is best defined when
bed material characteristics are combined with mean discharge to serve as the
independent variable.
In sum, we can say with some assurance that the combined effects of dis-
charge and sediment delimit the stability range for any channel pattern. Slope
is probably not the inducing agent in pattern change but more likely adjusts
as a dependent variable, along with the pattern, to changes in the sediment-
discharge regime.
The physical operations within rivers are driven by their attempts to establish Rivers, Equilibrium,
and maintain the most efficient conditions for transporting water and sedi- and Time
ment. Because every river has a unique combination of these two factors, the
parameters that define the equilibrium state must differ from river to river
and even in various segments of the same river. Every student of rivers un-
derstands that discharge and load are not constant, and thus the element of
time becomes a significant factor in any consideration of fluvial mechanics.
Remember that discharge and load are not in themselves independent vari-
ables because they are ultimately a function of climate, geology, and tectonics.
Furthermore, it is unreasonable to expect either discharge or load to change
independently of each other since both factors are related to the same, more
basic, controls. A
change in climate, for example, will surely alter discharge,
but it may prompt a simultaneous change in the character of the load
also
because vegetation and weathering will likewise adjust to the new climatic
regime.
The importance of time is that the type of fluvial variable most likely to
act to maintain equilibrium may depend on the time span being considered.
Figure 6.26 shows a gradual increase in mean annual discharge provided to a
river over a period of several hundred years. Presumably the load character-
istics are also changing for the reasons stated above. Within the period, normal
variations in discharge and load are accommodated by instantaneous adjust-

Pattern
ments of hydraulic geometry. Major floods may occasionally alter the valley change
topography or divert the river to a new position, but the channel itself will
reorganize according to the average flow and load conditions. At some point,
however, the gradually changing mean values of load and discharge can no Time
longer be balanced by hydraulic variables under the prevailing channel con-

Figure 6.26.
figuration or pattern. One flow event, perhaps not even a major flood, will
Diagram showing gradually
eventually exceed the stability limits of the original channel morphology; a increasing discharge with time.
fluvial threshold is passed and a major rearrangement of the channel pattern Discharge variations are
accommodated by variables of
or its configuration must take place.
hydraulic geometry until threshold
During this long-time interval, the river was in a perpetual quasi-equilib- is reached. At threshold, a major
rium condition if one considers only the instantaneous responses of hydraulic change in the river character,
such as a pattern change, is
geometry (w, d, v, sediment concentration, roughness) to short-lived events
required to carry the increased
such as floods. It is clear, however, that during the same interval, the river was average discharge.
approaching a different equilibrium condition that required a particular
channel morphology or pattern to balance the new mean values of discharge
and load.
discussed in chapter 1, we see, then, that fluvial equilibrium depends
As
on the time scale one uses to define it. Schumm and Lichty (1965) addressed
this problem by recognizing the importance of different geomorphic time spans
and indicating that some variables of fluvial geomorphology are dependent or

independent according to the time span being considered. Channel mor-
phology, for example, is an independent variable during steady time and exerts
a direct control on the hydraulics of flow. During graded time, however, channel
morphology is a dependent variable.
Chapter 6 246
Our interest here is in the adjustments a river might make to counter-

balance changes in discharge and load that occur over a period of tens or
hundreds of years, the time interval known as graded time (Schumm and Lichty
1965). Channel morphology is the main dependent variable on this temporal
scale, largely determined by mean values of the controlling factors. Rivers
during this episode may appear to be quite stable, if stability is judged by
hydraulic geometry. Even the morphology may show little change since its
adjustment may be imperceptibly slow.
Adjustment of Gradient
A common response of channel morphology to changes extending over a graded
time span is the alteration of slope. As discussed earlier, in alluvial rivers the
normal downstream decrease in gradient promotes a concave-up longitudinal

profile. The concavity, however, is usually not perfectly smooth in detail but
is commonly interrupted by perturbations. These can be caused by reaches
where the channel is floored by bedrock or by local zones of erosion or de-
position. Local filling may be initiated by an influx of bed material load that
is too coarse or too great in volume to be transported on the preexisting gra-
dient. For example, incision by upstream tributaries might provide so much
additional load that deposition would occur in the downstream reaches. A
change in particle size of the load may also be impressed on a trunk river if
tributary basins are experiencing a change in fundamental controls. This is
especiallycommon where tributaries drain mountains that are subject to

spasms of glaciation or uplift.
If a coarser load is produced, deposition should occur at the confluence
of the two rivers until the local gradient is increased sufficiently to allow the
bedload to be transported. As the channel raised by deposition, the
floor is
slope of the river upstream is and a wave of filling may

effectively lowered,
spread through the channel network. In contrast, a change that produces finer
or less load may induce a river to entrench its channel. The most apparent
cause of this response is the construction of large dams (Williams and Wolman
1984). These effectively starve the river of bedload in the reaches immediately
downstream from the construction, and the river uses its excess energy to lower
its gradient by downcutting. The entrenchment, however, may be stopped if
the channel floor becomes armored by particles in the alluvium that are too
large to be entrained on a lower slope (Livesey 1965; Hammad 1972).
Presumably, then, gradients will adjust to counteract changes in the load-
discharge relationship. Recognize, however, that even if a change in slope oc-
curs, there is no requirement that it will be propagated throughout the length

of the system. Other responses involving the channel cross section are equalh
able to provide the necessary counteraction away from the point of equilib-
rium disruption (Leopold et al. 1964; Leopold and Bull 1979).
A clear indication of a river gradient undergoing active readjustment is
a a short, oversteepened segment of the longitudinal profile known as a knick-
>- i point. Knickpoints are created by any process that results in the lowering of
base level. Such an event causes pronounced channel incision immediately up-
stream from the site of base level decline because the river is attempting to
Schematic evolution of knickpoints for various resistant and nonresistant bed material
Figure 6.27.
Inclination Models of knickpoint evolution for
various types of bed material r c is
tc >> Ta critical bottom shear stress

needed to initiate erosion r is
\ actual bottom shear stress. The

\\ knickpoint
break in
lip, shown at A, is the
slope where the channel

Uniformly N ^ \ .Te < T„ becomes oversteepened The
very resistant ^vj. knickpoint face at B extends from
the lip to the base of the
knickpoint
Downstream aggradation No downstream aggradation
Parallel Retreat Replacement
tc > r„
\ ':
<
Layered
L.
•t, < < t„ Uniformly \ ". Tc To
resistant- moderately \_ '^
nonresistant resistant
Initial profile
Time 1
Time 2
establish a new equilibrium condition. It is generally accepted that knick-

points will migrate upstream with time and, in doing so, may even initiate a
wave of erosion throughout a river basin. In fact, rates and distance of the
headward erosion are amenable to modeling (Pickup 1977) and have been
expressed in mathematical terms (Begin et al. 1981). Knickpoint behavior can
also be studied experimentally (Brush and Wolman 1960; Begin et al. 1980;
Gardner 1983). Nonetheless, it seems clear that the details of the headward
erosion depend on the character of the geologic and hydrologic setting.
Both flume and field data suggest that in channels formed of cohesionless
material, pronounced knickpoints will be smoothed out after only a short dis-
tance of upstream migration (Brush and 1960; Morisawa 1964). Wolman
Where the channel composed of bedrock or cohesive sediments, the knick-
is
point may retreat for considerable distances and still preserve a vertical
headcut, although Gardner (1983) showed that this does not always occur.
Gardner (1983) suggests that knickpoint evolution can occur in any of
three different modes, depending on the balance between the shear resistance
of the bed material and the shear stress produced in the river flow (fig. 6.27).
In knickpoint inclination there is a uniform change in slope of the knickpoint
face, the details of which depend on the material resistance and whether the
river can transport the sediment away from the front of the knickpoint. Where
the shear stress needed to erode this debris (t c ) is less than the actual shear
stress in the river (r ), aggradation will occur and the fill becomes part of the
Chapter 6 248
readjusted gradient (fig. 6.27). Parallel retreat is characterized by retreat of

the near-vertical knickpoint face without a change in its inclination. This mode
of migration is produced best if layering exists in the parent material such
that a more resistant zone overlies a less resistant zone (fig. 6.27). The third
type of migration, knickpoint replacement, occurs when erosion of the bottom
takes place upstream from the knickpoint lip as well as along the face. It re-
sults in a knickpoint profile consisting of two distinct zones in which the orig-
inal slope has been modified (fig. 6.27).
The changes in channel slope discussed above are responses familiar to
geologists as part of the concept of the graded river. The idea of a graded
condition in rivers finds its roots, once again, in the writings of G. K. Gilbert,
who indicated that "equilibrium of action" in streams consists of a mutual
adjustment between velocity, discharge, slope, and load. Later W. M. Davis
(1902) refined these ideas by introducing the terms "grade" and "graded
slopes" to describe the balanced fluvial condition. Although terminology left
wide latitude for interpretation, the early conceptual models clearly stressed
the importance of slope in the adjusting process, paving the way for the widely
accepted premise that a concave longitudinal profile is the trademark of a
graded river. In addition, Davis tied the graded condition to his cycle of ero-
sion by suggesting that development of the graded profile, which is the op-
timum form to transport sediment, marks the beginning of the mature stage
of the cycle.
The idea of a graded river did not go unchallenged, and in fact it was in
answer to these challenges that Mackin (1948) put the concept into perspec-
tive and provided for the first time a clear definition of a graded river as
. . . one in which, over a period of years, slope is delicately adjusted to

provide, with available discharge and with prevailing channel
characteristics, just the velocity required for the transportation of the
load supplied from the drainage basin. The graded stream is a system
in equilibrium; its diagnostic characteristic is that any change in any of
the controlling factors will cause a displacement of the equilibrium in a
direction that will tend to absorb the effect of the change.
The concept of grade as an equilibrium condition is valuable in under-

standing fluvial mechanics even though it probably overstates the role of slope.
The construction of Hoover Dam, for example, caused a radical decrease of
load in the Colorado River downstream from the dam. The river did not adjust
to the altered load by a slope change; instead, an increase in roughness brought
about the expected decrease in velocity (Leopold and Maddock 1953). Other
studies have shown dramatic changes in the channel shape to be the prime
factor involved in the balancing process (Schumm and Lichty 1963; Knox
1972). Thus, every altered load condition does not have to be countered with
a modification of declivity alone. This has tempted geomorphologists to con-
sider the possibility that all rivers flowing in alluvial channels are graded, ex-
cept that they adjust their slope and/or other channel characteristics to
transport their loads. In that sense, grade becomes analogous to quasi-equi-

librium, and the graded condition represents the most probable state for the
channel configuration and flow properties (Langbein and Leopold 1964).
In spite of the advantages in considering grade and quasi-equilibrium as
equivalent, it is important to remember Mackin's words "over a period of
years," since they may state the true distinction of a graded river (Knox 1975).
It is known, for example, that actively downcutting streams are still in quasi-
equilibrium as defined by their hydraulic geometry; that is, the hydraulic vari-
ables are perfectly adjusted to flow and their measurement would not indicate
that any fluvial response is occurring or, for that matter, that any change re-
quiring a response has occurred. It is also true, however, that the very fact of
progressive entrenchment indicates the river is approaching a different equi-
librium, one established over a period of years. It is not, perhaps, necessary
to require every graded-time adjustment to be made by a change in slope, but
it is important to recognize the time distinction. Graded-time adjustments seem
to be made by the variables of channel configuration and by pattern changes
as alluded to by Schumm
and Lichty (1965). Significantly, the initial response
may appear form of a hydraulic variable (such as n in the Hoover Dam
in the
situation), but this may not be the ultimate response. If the inducing change
is minor, the variables of hydraulic geometry may absorb it, just as they do
during floods. But if the change is major, and especially if it occurs gradually,
the final response may be one involving the channel configuration or pattern,
and changes in those factors take time.
Adjustment of Shape and Pattern

As explained above, one of the more significant advances in recent years is
the growing awareness of fluvial geomorphologists that rivers can respond to

altered discharge and/or load in ways other than cutting or filling of the channel
(for example, see Dury 1964). In a series of papers Schumm (1965, 1968,
1969) pieced together many of the empirical equations we have examined into
a comprehensive, though qualitative, model of possible river adjustments to
altered hydrology and load. The following equations are the basis for what
Schumm (1969) refers to as river metamorphosis:
2h Qi =* — — S-d-
-p^
d~p~
In these equations Q, is the percentage of the total load transported as bed

material load (sand-sized or larger), and Q K can be either the mean annual
discharge or the mean annual flood. The other variables are width (w), depth
Chapter 6 250
Figure 6.28.
Adjustments Murrumbidgee
of the
changes in
River, Australia, to
climate and sedimentology PC2
(paleochannel 2) functioned at
time of greater aridity; PC1
(paleochannel 1) was sinuous and
much larger than present river.
Modern river is highly sinuous and
flows within the channel limits of
PC,. (After Schumm 1968)
Riverine plain
1 Miles
SI
Kilometers
'
PC,
(d), slope (S), meander wavelength (Z.), width-depth ratio (F), and sinuosity
(P). The plus or minus exponents indicate whether the variables are increasing
or decreasing.
To exemplify the use of these equations, let us assume that a large area
is clear-cut of its natural forest cover. We can expect an increase in QM ., be-
cause infiltration rates will be lowered and direct runoff will increase, as well
Q„ because coarse sediment normally stabilized on slopes by

as an increase in
rooting now makes its way to the channel; the coarse sediment also will be
moved more frequently because of the increased peak discharge. With both
Q K and Q, increasing, we can expect increases in width, wavelength, and
W/D and a decrease in sinuosity. Depth and slope may vary in either direction.
Slope will probably increase, however, because the channel becomes straighter,
and depth will probably be constant or decrease since both w and W/D in-
crease.
An excellent geological example of river metamorphosis is found in the
history of the Murrumbidgee River which flows across a large alluvial plain
in New South Wales, Australia (Schumm 1968). As figure 6.28 illustrates,
the present highly sinuous river flows within a floodplain containing large oxbow
lakes and other features that preserve an older and larger channel of the Mur-
rumbidgee (paleochannel 1). Evidence of a still older, low-sinuosity channel
(paleochannel 2) is also present on the plain. The morphologic, sedimento-
logic,and hydrologic characteristics for the three channels are presented in
table 6.4. Pedogenic and geomorphic data confirm that during the tenure of
paleochannel 2 the climate was more arid than at present, and at the time of
paleochannel 1, more humid than now. Using the present river as a norm.
Table 6.4 geometric, and sedimentological data comparing the modern

Fluvial, river with paleochannel 1 and
paleochannel 2, Murrumbidgee River, Australia.
f day)
ft.
(M) Q. Discharge
0) >.
Size
Q cf_ sec) per per
Channel
£ £ o
'3i
o
o>
Bankfull
per
Bankfull
Discharge
day)
Discharge Meander
Wavelength
Median
Silt-Clay
Percent
Q. 3 Velocity
Grain
tip C Sand (tons per Sand (tons
5£
(cfs)
(mm)
or. 55 OS (ft («)
Murrumbidgee
River near
Darlington
Point 57 25 220 21 10 2.0 0.7 3.0 10,000 9 2,000 2,800
Paleochannel 1 16 460 35 13 1.7 0.8 4.2 51,000 45 21,000 7,000
Paleochannel 2
Northern
(Kearbury pit)
Small 055 1.6 600 9 67 1 1 20 52 23,000 90 54,000 18,000
Medium 1,000 12 83 6.3 73,000 140 140,000
Large 1,700 20 90 88 290,000 300 510,000
Central (Kulki pit):
Small 060 3.4 500 8 63 1.1 20 48 19,000 70 35,000 15,000

Medium 800 9 90 5.2 35,000 80 64,000
Large 800 14 57 70 77,000 210 178,000
After Schumrn 1968
Schumm (1968) compared the adjustments in channel parameters that oc-

curred in the Murrumbidgee River under changing climates. In a change to-
ward aridity (present — • paleochannel 2), the decrease in precipitation over
the entire basin should decrease and increase Q,. According to the equa- QK .
tions of river metamorphosis, to transport the changed load with less water,
the channel should become shallower and probably wider (d~, F "). Wave-
4
+
length should probably increase (L ) because sinuosity should decrease (P~)
and slope increase (S + ). Comparison of the Murrumbidgee variables in the
table shows that the actual changes agree with those predicted by Schumm's
equations.
An increase in precipitation (present — » paleochannel 1) would increase
QK but probably cause little change in sediment yield or size because the den-
would also increase (as discussed in chapter 5). In response,
sity of vegetation
the slope and W/D would show little change, but the channel would become
wider and deeper, and the dimensions of meandering (wavelength and ampli-
tude) increase (table 6.4).
One of the more significant observations made by Schumm is that changes
in the Murrumbidgee gradient in response to altered controls were made
without cutting or filling the channel. Most of the slope adjustment was ac-
complished by changes in the length of the river due to variations in sinuosity.
Chapter 6 252
To summarize, our present understanding of rivers requires that we re-

think the meaning of equilibrium and its relation to fluvial mechanics. The
following emerge as prime points for future consideration and should be of
special interest to engineers and geologists:
1. The adjustment of slope in rivers can be made by a major change

in channel pattern rather than by vertical filling or trenching.
2. The initial response of a river may not be the same as its end
response. In addition, regulation of a river may generate responses
that do not remain local but alter channel morphology over great
lengths of the river system.
3. Both discharge and sediment must be considered in the prediction
The greater width associated with channels of
of fluvial mechanics.
low discharge (such as paleochannel 2 in the Murrumbidgee, table
6.4) is contrary to many regime equations relating width and
discharge. The tendency of the river's response to a change in
discharge, therefore, can be counterbalanced by a simultaneous
change in the character of the load.
Summary River action, like all geomorphic processes, behaves according to the driving
and resisting forces built into the system. For example, a river will entrain,
transport, or deposit sediment depending on the driving energy given to the

water by velocity, depth, and slope, and on the amount of that energy con-
sumed by the resistance to flow offered by elements such as channel config-
uration, particle size, and sediment concentration. The work demanded of a
river — the amount of load
it must handle under prevailing discharges is de- —
termined by the geological and climatic character of the drainage basin. Each
river develops a particular combination of shape, gradient, and hydraulic vari-
ables (called the hydraulic geometry) that allows it to accomplish its work
most efficiently. The river will attempt to maintain its high efficiency by ad-
justing the above properties whenever discharge or load vary. Because dis-
charge and load fluctuate continuously, equilibrium as a steady-state condition
can never be attained, and the river variables must perpetually be adjusting.
Nontheless, the normal variations of discharge and load are accommodated
by hydraulic geometry.
The type of channel pattern (straight, meandering, braided) a river dis-
plays and the longitudinal profile are other fluvial characteristics controlled
by the basin environment. Each pattern originates in a specific manner, and
its geometric form is designed to facilitate the work of a river, measured as
the prevailing values of discharge and load. Once established, the pattern will
be maintained as long as the normal variations in load and discharge can be
absorbed by the mechanics of hydraulic geometry.
Major long-term changes in climate or basin tectonics may alter the av-
erage discharge and/or load to a point where adjustments of hydraulic ge-
ometry can no longer maintain the most efficient system. When those threshold
values of discharge or load are reached, major fluvial responses in the form
of pattern changes, degradation or aggradation, or dramatic revisions of the
width-depth ratio will occur to reestablish the greatest fluvial efficiency. An
excellent example of these major reactions has been detailed for the Murrum-
bidgee River in Australia. It is important to recognize, however, that our present
knowledge does not allow us to predict which of the possible adjustments will
occur in response to major changes in the fundamental controls.
The following references provide greater detail concerning the concepts dis- Suggested Readings
Hey, R. D.; Bathurst, J. C; and Thorne, C. R. 1982. Gravel-bed rivers.

New York: Wiley-Interscience.
Leopold, L. B., and Maddock, T., Jr. 1953. The hydraulic geometry of
stream channels and some physiographic implications. U.S. Geol.
Survey Prof. Paper 252.
Leopold, L. B., and Wolman, M. G. 1957. River channel patterns; braided,

meandering, and straight. U.S. Geol. Survey Prof. Paper 282-B.
Leopold, L. B.; Wolman, M. G.; and Miller, J. P. 1964. Fluvial processes in

geomorphology. San Francisco: W. H. Freeman.
Richards, K. 1982. Rivers. London: Methuen and Co.
Schumm, S. A. 1968. River adjustment to altered hydrologic regimen,

Murrumbidgee River and paleochannels, Australia. U.S. Geol. Survey
Prof. Paper 598.
. 1969. River metamorphosis. Am. Soc. Civil Engrs. Proc, Jour.

Hyd. Div., HY 1:255-73.
. 1977. The fluvial system. New York: Wiley-Interscience.

Fluvial Landforms
7
I. Introduction b. Entrenchment and
II. Floodplains Location of
A. Deposits and Topography Deposition
B. The Origin of Floodplains B. Pediments
III. Fluvial Terraces 1. Morphology and
A. Types and Classification Topography
B. The Origin of Terraces a. Size and Shape
1. Depositional Terraces b. Surface Topography
2. Erosional Terraces c. The Piedmont Angle
C. Terrace Origin and the Field d. Slope

Problem 2. Processes
IV. Piedmont Environment: Fans 3. Formative Models
and Pediments V. Deltas
A. Alluvial Fans A. Sediment, Form, and
1 . Fan Morphology Classification
2. Processes, Deposits and B. Dynamics and Delta
Origins Evolution
a. Deposits and VI. Summary
Depositional VII. Suggested Readings
Processes
255
—
Chapter 7 256
Introduction In chapter 6 we examined the basic mechanics of fluvial processes and found
that the activity within a stream channel is generally related to the energy
possessed by the river and to the ways that energy is utilized to carry water
and sediment most efficiently. Rivers, however, are more than natural sluices;
they also mold the geologic setting into discernible topographic forms. They
accomplish this primarily through the erosional capability inherent in the
movement of sediment-laden water, and through the deposition of debris that

occurs when the transporting energy is less than the demands being made on
it. Some fluvial features are purely erosional: the topographic form is clearly-
one of sculptured rock, and little, if any, sediment is associated with the fea-
ture. Others may be entirely depositional. and the exposed topography is formed
by the burial of an underlying surface that existed before the covering sedi-
ment was introduced. In these cases, the bedrock framework may have no
influence on the surface configuration. Many features spring from some com-
bination of both erosion and deposition; the pure cases are probably end mem-
bers of a continuum of possible forms.
If rivers establish or nearly establish some form of equilibrium, itseems
reasonable to expect that fluvial features — the tangible results of river work
will somehow reflect the balanced processes that created them. Here we will
investigate these end products of fluvial action and. wherever possible, docu-
ment how the properties of features may reveal the processes involved in their
origin.
Floodplains Floodplains are perhaps the most ubiquitous of fluvial features, found in the
valley of every major river and in most tributary valleys. However, a precise
definition of a floodplain is more difficult than one might expect. Topograph-
ically and geologically speaking, a floodplain is the relatively flat surface oc-
cupying much of a valley bottom and is normally underlain by unconsolidated
sediment. The sediments of most valley bottoms are not necessarily a function
of the river occupying the valley but may be deposited there by a variety of
geomorphic processes. Nonetheless, to be considered as part of the floodplain,
the surface and the sediments must somehow relate to the activity of the present
river. The definition must also have a hydrologic connotation, since the flood-
plain is a surface subject to periodic flooding. It can easily be defined in terms
of hydrology as the water level attained in some particular stage of the river.
Detailed analyses (Wolman and Leopold 1957) demonstrate that most topo-
graphic floodplains are subjected to flooding nearly every year or every other
year. The recurrence interval of bankfull stage, for example, averages about
1.5, indicating that most rivers leave their channels two out of ever] three
years. If the surface flanking the river has any relief, however, part of the
topographic floodplain will not be inundated by the annual or biannual flood
that marks the hydrologic floodplain.
Engineers may view floodplains in a different sense, being more concerned
with how much of the total discharge moves across the floodplain surface during
high flow events (Bhowmik and Demissie 1982). In addition, engineers and
257 Fluvial Landforms
basin managers are primarily concerned with how much damage a flood will
cause and therefore have little interest in yearly floods. In fact many engineers
use the phrase flood-damage stage —the water level where overflow begins to
cause damage — to mean the flood stage. In general, the damage stage is well
above both the level of bankfull and the average elevation of the topographic
floodplain.
Regardless of how it is defined, a floodplain plays a very necessary role
in the overall adjustment of a river system. It not only exerts an influence on
the hydrology of a basin (lag, etc.) but also serves as a temporary storage bin
for sediment eroded from the watershed. Therefore, floodplains are features
that are both the products of the river environment and important functional
parts of that system.
Deposits and Topography

Floodplains are composed of a variety of sediments that are created by diverse
processes and accumulate in distinct subenvironments within the valley bottom.
Most floodplains can be differentiated into deposits of channel fill, channel lag,
splays, colluvium, lateral accretion, and vertical accretion (Happ et al. 1940;
Lattman 1960). Near the valley sides, colluvium, resulting from unconfined
wash and mass wasting, may be prominent in the floodplain sequence; toward
the axis of the valley, these deposits grade into alluvial-type deposits. Coarse
debris from which the fines have been winnowed are interpreted as channel
lag deposits, in contrast to channel fill, which consists of a poorly sorted ad-
mixture of silt, sand, and gravel. Splay deposits are composed of material
spread onto the floodplain surface through breaks in natural levees and usually
are more coarse-grained than the overbank sediments that they cover. The
most important deposits in the floodplain framework are those of lateral ac-
cretion and vertical accretion, which in some cases can be separated on the
basis of particle size, the laterally accreted sediment usually being sands and
gravels, more coarse-grained than the vertically accreted silts and clays (shown
in the photograph in fig. 7.1). Point bars, however, the most common deposit
of lateral accretion, may sometimes have the same texture as the overbank
sediments (Wolman and Leopold 1957). Therefore, particle size is not an in-
fallible criterion for distinguishing between vertical and lateral accumulation.
Valleys of large rivers such as the lower Amazon and Mississippi have
been actively aggrading for thousands of years. As a result, they may have
incompletely formed floodplains (Tanner 1974) or special features caused by
rapid elevation of the valley floor (Baker 1978). Nonetheless, most valleys of
large rivers are occupied by well-defined floodplains that consist primarily of
laterally and vertically accreted deposits. These are usually associated with
specific depositional environments. In the Mississippi River valley, for ex-
ample, they have been broadly categorized (Fisk 1944) as (1) channel types,
including point bars, chutes and sloughs, and sand ridges of meander scrolls;
and (2) overbank types, consisting of splays, natural levees, and backswamps
(fig. 7.2). The surface of the floodplain may have considerable microrelief that
reflects the fluvial mechanics in the various depositional environments.
Chapter 7 258
Figure 7.1.
Normal sequence of floodplain The floodplain surface is most irregular in azone close to the river where
stratigraphy Lower two-thirds ot
point bars are molded into alternating ridges and swales that Leopold and his
deposit is composed of coarse-
grained, laterally accreted, point colleagues (1964) refer to as meander scrolls. The characteristic scroll to-
bar gravel Yardsticks show how pography may start as a longitudinal bar with a narrow trough to its rear
gravel in lowest unit dips slightly
(Sundborg 1956) or simply as a low ridge of sediment that accumulates on
to the left Upper third of deposit
is accreted silt
vertically the inside of a meander bend during bankfull flow (Kolb and Van Lopik 1958;
deposited during overbank flow Hickin 1974). When the high discharge subsides, the ridge is exposed and
Some gravel at top of floodplain is
rapidly vegetated, becoming, for all practical purposes, the new channel bank.
also overbank debris Sexton
Creek. Shawnee National Forest, The next high flow repeats the process. As the river shifts across the valley b>
southern Illinois undercutting on the outer bank, the successive ridges and intervening swales
that characterize meander scroll topography develop simultaneously (fig. 7.3).
Periodically, flow will break across the point bar surface, often occupying a
particular swale and scouring the surface into more pronounced low channels
called chutes. Wolman and Leopold ( 1 957) measured velocities of up to 3 fps
in chute channels, a flow that is capable of eroding the surface and trans-
porting coarse sand. Chute erosion tends to accentuate the scroll topography.
Figure 7.2.
Map showing complex distribution
of various types of deposits on a
portion of the Mississippi River
floodplam near Grand Tower, III
(Courtesy of S E Harris. Jr.)
ir deposits Channel
Overbank deposits Ponds
Point bar succession

Figure 7.3.
Ridge Water surface Meander scroll topography formed
by point bar deposition in a
Swale laterally migrating river.
Channel boundary (period 1

Channel boundary (period 2)
Present channel boundary
Chapter 7 260
and even though the slough areas may be silted in by overbank deposition, the
scalloped profile of scroll topography (fig. 7.4) can be preserved for hundreds
of years (Hickin and Nanson 1975). (For greater detail about point bar me-
chanics and deposits, see McGowen and Garner 1970.)
In addition to the irregularity caused by scrolls on the inside of meander
bends, some topographic relief near the channel is due to the formation of
natural levees. Natural levees stand along most major rivers as low ridges that
commonly are broader than one might expect, often extending for hundreds
of meters. Levees are usually highest near the active channel and slope grad-
ually toward the valley sides. They owe their character to the retardation of
flow velocity when rivers leave the channel, which results in the largest sus-
pended particles being deposited adjacent to the bank. Levee deposits are
Figure 7.4.
(A)Diagram of two profiles of
meander scroll topography that
has been preserved for
considerable periods of time.
Beatton River, Canada.
(B)Meander scroll topography on
abandoned point bar of the 100 150 200
Mississippi River in Pike County, Meters
III. Local relief between ridges and
swales is about 3 meters
100 200
(A) Meters
therefore more coarse-grained than most other overbank sediment and are ac-
creted more rapidly. For example, Kesel and his colleagues (1974) found a
net addition of 53 cm on natural levees during the two-month overflow of the
Mississippi River in 1973, while only 1.1 cm accumulated in the backswamp
area during the same period.
As one proceeds away from the river, the floodplain topography becomes
much more and its flatness is interrupted only by oxbow channels or

regular,
by splay deposits that have been debouched onto the backswamp surface. The
zone away from the active river is characterized by gradual accumulation of
overbank sediment that tends to subdue any existing relief. Abandoned chan-
nels represented by oxbows or oxbow lakes gradually fill with silts and clays,
leaving clay plugs as the only evidence of the former channel. Old sloughs,
chutes, or cutoffs can also fill with overbank sediment, adding to the general
reduction of relief away from the channel. It is incorrect, however, to assume Figure 7.5.
Large lobe of overbank sand and
that all overbank sediment is fine-grained. Where banks are cohesive, coarse gravel deposited on the floodplain
gravel can be transported onto the surface of the floodplain (McPherson and surface in December 1982 flood
Rannie 1967; Costa 1974a; Ritter 1975; see fig. 7.5). Lobe sediment is 1 .0 to 2 m
thick and 100 m wide Gasconade
River near Mt. Sterling, Mo.
Chapter 7 262
Table 7.1 Rates of lateral migration of rivers in valleys
Approxi-
mate Rate of
Size of Amount Move-
Drainage of ment
Area Move- (feet
(square ment Period of per
River and Location miles) (feet) Measurement year)
Tidal creeks in Massachusetts 60-75 yr

Normal Brook near Terre Haute, Ind ±1 30 1897-1910 23
Watts Branch near Rockville. Md 4 0-10 1915-55 0-0.25
4 6 1953-56 2
Rock Creek near Washington, DC 7-60 0-20 1915-55 0-0.50
Middle River near Bethlehem 18 25 10-15 yr 25
Church, near Staunton, Va
Tributary to Minnesota River near 10-15 250 1910-38
New Ulm, Minn
North River, Parnassus quadrangle. 50 410 1834-84
Va,
Seneca Creek at Dawsonville, Md 101 0-10 50-100 yr 0-0.20
Laramie River near Ft. Laramie, 4.600 100 1851-1954 1
Wyo
Minnesota River near New Ulm, 10,000 1910-38
Minn.
Ramganga River near Shahabad, 100,000 2,900 1795-1806 264
India
100,000 1,050 1806-1883 14
100,000 790 1883-1945 13
Colorado River near Needles. Calif. 170,600 20,000 1858-83 800
170,600 3,000 1883-1903 150
170,600 4,000 1903-1952 82
170,600 100 1942-52 10
170,600 3,800 1903-42 98
Yukon River at Kayukuk River, 320,000 5,500 170 yr 32
Alaska
Yukon River at Holy Cross, Alaska 320,000 2,400 1896-1916 120
Kosi River, North Bihar, India 369,000 150 yr 2.460
Missouri River near Peru, Nebr. 350,000 5,000 1883-1903 250
Mississippi River near Rosedale. 1,100,000 2,380 1930-45 158
Miss
1,100.000 9.500 1881-1913 630
From Wolman and Leopold 1957 See this work for data sources lor individual rivers
In general, the relationship between floodplain deposits and river pro-

cesses would be straightforward if rivers would stay in the same place for ex-
tended periods of time. Actually, as the rates in table 7.1 indicate, most rivers
migrate laterally across the valley bottom quite rapidly, forcing the deposi-
tional environments also to shift their location with time. A backswamp re-
gion, for instance, may include remnant deposits of a channel. The displacement
of one environment by another adds to the complex maze of floodplain de-
posits, emphasizing the point that floodplains are dynamic rather than static
fluvial features.
Figure 7.6.
Progressive lateral erosion and
point bar deposition (cross
sections) in Watts Branch near
Rockville, Md (From Wolman and
Leopold 1957)
5 fee:
Explanation
H Cut
1953
Fill
1955 1956
The Origin of Floodplains

It is now generally accepted that two dominant fluvial processes act simul-
taneously to develop most floodplains. As described earlier, maximum erosion
in meandering rivers takes place on the outer bank just downstream from the
axis of curvature. At the same time, sediment accumulates in point bars that
build up along the inside of the meander bend. Detailed study of these pro-
cesses has documented that bank erosion and point bar accumulation are vol-
umetrically equal during any given period of lateral and downvalley migration
of the meander bends (Wolman and Leopold 1957). In addition, data from
the same study show that the point bars tend to increase in height until they
reach the level of the older part of the fioodplain. It seems clear that a mean-
dering river can shift its position laterally without changing the channel shape
or dimensions.
In channels where coarse sediment is an important part of the load, the
point bars tend to collect sediment that is easily distinguished from that of
overbank origin. During low flow, sediment of all sizes may be temporarily
trapped in the channel floor, but at the peak of bankfull discharges this ma-
terial will be removed along with any debris eroded from the undercut banks.
The coarse sediment is deposited on the point bars, now submerged. As figure
7.1 shows, these deposits commonly display cross-beds dipping into the channel.
Over a period of years, the point bars expand laterally, being progressively
spread across the valley bottom as a thin sheet of sand or gravel (Mackin 1937;
Leopold et al. 1964). If the load is not characterized by coarse-grained par-
ticles, the spreading of lateral accretion deposits proceeds in exactly the same
way (fig. 7.6), but the point bar sediment may be more difficult to distinguish
from the overbank materials.

Chapter 7 264
100
Figure 7.7.
Cross sections of floodplains in
North Carolina and South
Carolina. (From Wolman and
Leopold 1957)
wi <J« ^
Buck Creek
South Tyger
*
m M&JKI ~
%L
North Tyger River
»
HHI
i
Flood plain | Sand o™ 500

feet
Bedrock • Gravel
X Silt and clay
The maximum thickness of laterally accreted deposits is determined by

the depth to which a river can scour during recurring floods. Natural channels
are probably scoured to a depth 1.75 to 2 times the depth of flow attained
during a flood (Wolman and Leopold 1957; Palmquist 1975). This rule of
thumb generally fits the scouring observed in a variety of perennial rivers
(Leopold et al. 1964, p. 229), but where the channel width is prevented from
expanding by bridge supports or extremely resistant banks, the scour depth
may reach 3 or 4 times the water depth. The thickness of lateral accretion
deposits should, therefore, increase gradually downvalley along with the normal
increase in discharge and depth noted on most rivers.
The second dominant process in the origin of floodplains is overbank flow.
As rivers wander across their valley bottoms, they normally leave their channel
confines during periodic flooding and deposit fine-grained sediment on top of
the floodplain surface. Because of this, the floodplain is vertically accreted,
and in most floodplain stratigraphy, a thin layer of overbank silt and clay rests
on the laterally accreted point bar deposits described above (figs. 7.1, 7.6).
Ideally, then, the entire floodplain sequence consists of a relatively thin

accumulation of laterally and vertically accreted sediments that have been
spread more or less evenly across the valley bottom (fig. 7.7). The two types
of deposits can differ in their textural properties (although the process me-
chanics does not require it), but they will be essentially the same age.
An analogous development is associated with floodplains of braided rivers
except that the systems are more dynamic and less regular. Bars and bank
erosion, for example, are not restricted to one particular side of the channel.
Table 7.2 Increment rates of overbank depositon in major floods
Average Thickness of
River Basin Flood Depositon (feet)
Ohio River Jan -Feb 1937 0.008

Connecticut River March. 1936 .114
Connecticut River Sept . 1938 .073
Kansas River July, 1951 .098
From Wolman and Leopold 1957 See this work for references on individual data sources
and the river can shift its position without laterally eroding the intervening
material. Abandoned channels and islands gradually coalesce into a contin-
uous floodplain surface. Normally, floodplain sediments in a braided stream
system might reasonably be expected to be less thick and more irregular. Rec-
ognition of the true floodplain sequence, however, may be complicated by the
fact that braided systems are commonly associated with long-term valley ag-
gradation. In such a case, the floodplain sequence might appear to be enor-
mously thick. However, floodplains relate to the hydraulics of the present river
only,and the sedimentary pile it produces is probably just a thin skim on top
of the fill.
The model of floodplain origin just described raises the question as to which
process — overbank flow plays the dominant role. There
lateral migration or —
is probably no universal answer to this question, because each system obeys
its own unique combination of controlling factors. Nonetheless, evidence sug-
gests that most floodplains result primarily from the processes associated with
lateral migration. Perhaps the most persuasive argument for that conclusion
uses a comparison of the rates involved in the two competing processes. Table
7.2 is a random sampling of sediment increments measured during floods. Al-
though incomplete, it shows that backswamp deposition during overbank flow
tends to be limited, probably because the maximum sediment concentration
in a flood often occurs before bankfull stage is reached. That is, most sediment
is from the system before overbank conditions are attained
transported
(Wolman and Leopold 1957; Moss and Kochel 1978).
Assuming the same vertical increment with each flood, the level of a flood-
plain built entirely by overbank deposition should increase at a progressively
decreasing rate, as shown by the curves in figure 7.8. The initial growth would
be rapid because flooding would occur frequently, and perhaps 80 to 90 per-
cent of the floodplain construction would take place in the first 50 years
(Wolman and Leopold 1957; Everitt 1968). However, as the surface grows
higher relative to the channel floor, the stage needed to overtop the banks is
also increasing. The surface is and the rate of growth

inundated less frequently,
is drastically fact that most floodplains are occupied with water

retarded. The
nearly every year argues against the importance of overbank deposition in their
construction.
Chapter 7 266
Elevation, above low water Elevation in feet

Figure 7.8.
(Little Missouri curve) (Wolman-Leopold curve)
Increase in elevation of floodplam
with time Lower curve from 8
empirical data collected on
floodplam of the Little Missouri
River; upper curve was derived
theoretically for Brandywine Creek
(Pa.) by Wolman and Leopold
(1957) Note different vertical
scales
wm
200 300
Time (years)-
At reasonable increment rates for overbank deposition, a 3 m thick flood-

plain sequence would probably take several thousand years to accumulate.
Assuming this is a valid estimate, it is instructive to note again the lateral
migration rates given in table 7.1, which indicate that most large rivers mi-
grate quite rapidly. This is especially true when their meander geometry is
adjusted for efficient lateral shifting. Hickin and Nanson (1975) suggest that
the rate of lateral migration in meander bends of the Beatton River (British
Columbia, Can.) is greater when r„Jw is about 3 (see chapter 6). At values
higher or lower than 3, the rate of channel migration decreases dramatically.
This observation has been supported by theoretical analyses (Begin 1981).
It seems certain that the magnitude of vertical accumulation depends pri-
marily on the rate at which the river migrates laterally. The total thickness
will approximate the vertical accretion that can be accomplished in the time
the river takes to migrate the entire width of the valley. For example, if a
floodplain is a kilometer wide and the river shifts laterally at a rate of 2 m a
year, it will take 500 years for the river to complete one swing across the valley.
At any given locality, perhaps several meters of overbank sediment will ac-
cumulate in that time, but the entire deposit will be reworked by lateral ero-
sion when the river reoccupies that position. The lateral migration rate thus
becomes a controlling and limiting factor on the thickness of overbank de-
position. The apparent preeminence of lateral processes in floodplain construc-
tion does not mean that overbank deposition is unimportant. Indeed, vertical
accretion may be the dominant process involved during the initial stage of
floodplain development (Schumm and Lichty 1963; Everitt 1968). even though
lateral erosion may subsequently rework the sequence. Furthermore, in some
cases rivers might lack the widespread lateral movement needed to rework the
and portions of the surface could continue unimpeded growth

entire floodplain,
by overbank deposition (Ritter etal. 1973; Kesel et al. 1974; Smith and Smith
1976; Nanson and Young 1981).
In summary, fioodplains appear not onlyto be formed by balanced fluvial
systems but also to serve as integral parts of the system. They are constructed
by simultaneous processes of lateral migration and overbank flooding. Point
bars, the deposits of lateral accretion, are spread in a rather even sheet across
the valley bottom, while overbank deposits accumulate over the entire flood-
plain surface away from the channel. The floodplain acts as a storage area for
sediment that cannot be transported directly from the basin when it is eroded.
Fioodplains are usually considered to be features associated with stable
rivers, but there is no overriding reason why they cannot be present when a
channel is undergoing long-term aggradation or degradation. In fact, the ob-
served frequency of overbank flooding can continue during valley filling if the
channel floor and the floodplain surface are raised at the same rate. Once the
thickness of the valley deposits exceeds the limits of a reasonable scouring
depth, however, the sediment below that depth can no longer be considered as
part of the active floodplain. In a degrading channel, the floodplain becomes
a terrace when channel incision prevents the river from inundating the surface
annually or biannually.
Terraces are abandoned fioodplains that were formed when the river flowed Fluvial Terraces
at a higher level than at present. The surface of the terrace is no longer related
to the modern hydrology in that it is not inundated as frequently as an active
floodplain. Topographically, a terrace consists of two parts: a tread, which is
the flat surface representing the level of the former floodplain, and the scarp,
which is the steep slope connecting the tread to any surface standing lower in
the valley (fig. 7.9). The very presence of a terrace indicates an episode of
downcutting: some change must occur between the conditions prevailing during
formation of the tread and those producing the scarp. Usually the downcutting
phase begins as a response to climatic or tectonic changes, but these are not
always necessary. The tread surface normally is underlain by alluvium of vari-
able thickness, but in a pure sense, these deposits are not part of the terrace.
To avoid confusion, it is better to limit the term to the topographic form and
refer to the deposits as fill, alluvium, gravel, etc.
Types and Classification

Howard and his coauthors (1968) categorize terraces as erosional or deposi-
tional. Erosional terraces are those in which the tread has been formed pri-
marily by lateral erosion. If the lateral planation truncates bedrock, the terms
bench, strath, or rock-cut terrace are commonly used. If the erosion crosses
unconsolidated debris, the terms fill-cut or fillstrath (Howard 1959) have been
suggested. Depositional terraces, the second major grouping, are those ter-
races where the tread represents the uneroded surface of a valley fill. Figure
7.10 illustrates both types.
Chapter 7 268
(A)
(B)
Tread
Scarp
Figure 7.9.
(A) Parts of a fluvial terrace
(B) Terraces and point bars along
the Snake River in western Idaho.
Erosional terraces, especially rock-cut types, are identifiable by the fol-

lowing, rather distinct, properties (Mackin 1937): (1) they are capped by a
uniformly thin layer of alluvium in which the total thickness is controlled by
the scouring depth of the river involved; and (2) the surface cut on the bedrock
or older alluvium is a flat mirror image of the surface on top of the capping
Erosional terrace
alluvium (fig. 7.10). In contrast, the alluvium beneath the tread of deposi-
(A)
tional terraces varies in thickness and commonly exceeds any reasonable
scouring depth of the associated river. Although the tread surface may be fiat,
the surface beneath the fill can be very irregular (fig. 7.10.)
Another classification scheme is based on the topographic relationship be-
tween terrace levels within a given valley, as illustrated in figure 7.1 1. In this
method, terrace treads that stand at the same elevation on both sides of the Depositional terrace
valley are called paired (matched) terraces and presumably are the same age. (B)
If the levels are staggered across the valley they are said to be unpaired (un-
Figure 7.10.
matched) terraces. Most investigators interpret unpaired terraces as erosional
(A) Erosional terrace. Thin alluvial
types, formed by a stream simultaneously cutting laterally and downcutting cover with truncation of
very slowly. Levels across the valley, therefore, are not exactly equivalent in underlying bedrock along smooth,
even surface. (B) Depositional
age but by the amount of time needed for the river to traverse the valley
differ
terrace. Terrace scarp underlain
bottom. Actually, unpaired terraces can also be depositional in origin if the by alluvium that is highest level of
entrenchment between two episodes of valley filling occurs at the valley sides deposited in valley. Note
fill
thickness of alluvium and irregular

rather than along the valley axis (Ritter 1967). Paired and unpaired terraces
bedrock surface beneath the fill.
can be of any origin (erosional or depositional) and develop from either bed-
rock or alluvium. The terms, therefore, are entirely descriptive and carry no
genetic connotation.
The Origin of Terraces
Depositional Terraces The development of a depositional terrace always re-

quires (1) a period of valley filling and (2) subsequent entrenchment into or
adjacent to the fill. This cyclic pattern is necessary because the alluvium at
the tread surface takes form from purely depositional processes. The tread,
its
Paired terraces
in fact, represents the highest level attained by the valley floor as it rose during (A)
aggradation. The initial entrenchment that forms the terrace scarp is pri-
marily vertical, and so the tread surface is virtually unaffected by subsequent
lateral erosion at a lower level (see fig. 7.10).
Valley filling occurs when, over an extended period, the amount of sedi-
ment produced in a basin exceeds the amount that the river system can carry
Unpaired terraces
away. Prolonged aggradation is usually triggered by (1) glacial outwash,
(B)
(2) climate change, or (3) changes in base level, slope, or load due to rising
sea level, rising local or regional base level, or an influx of coarse load because
Figure 7.11.
of uplift in source areas. Where tectonics are ruled out, the balance between Terraces classified on basis of
load and discharge is determined primarily by climatic processes, although it topographic relationships.
(A) Paired terraces have treads at
may be driven by glaciation and may be complexly interrelated with sea level
same level on both sides of the
changes, etc. (see discussion in chapter 2). Although entrenchment has been valley (B) Unpaired terraces stand
considered theoretically (Foley 1980a) and studied experimentally (Shepherd at different elevations on either
side of the valley.
and Schumm 1974), details of the mechanical processes involved are still not
understood. Nonetheless, incision, like filling, can be triggered by tectonic
events or climate change.
Chapter 7 270
Figure 7.12.
Physiographic and geologic
controls of stream piracies in
Mountai
piedmont regions Main river
coming from mountain carries
coarse-grained load on high
gradient. Tributaries that head in
piedmont region carry fine-grained
load on gentle gradient. At an Main Valley Tributaries
equal distance upstream from A'
juncture of the two rivers, the
stand at lower elevation
tributaries
and are position to capture the
in
poised master stream
Cross-section A-A'
Elev. 5000'
Spontaneous filling and cutting can also result from physical processes
that have no relationship to tectonics or climate (see Foley 1980b). For ex-
ample, small streams that rise in the plains surrounding high mountains often
have gentler valley gradients than do the larger rivers that head in the moun-
tains. This trait develops best where the piedmont area is underlain by easily
eroded siltstones and shales. Streams originating there adjust their gradients
to the fine-grained sediment released from the weakly resistant rocks. The
mountain rivers, however, must transport coarse bedload derived from the re-
sistant rocks in the mountain

and they do so most efficiently by deve-
core,
loping a steeper channel gradient. Because of this unique physical control, the
main river stands at a higher elevation than its tributaries at an equal distance
upstream from their confluence (fig. 7.12). It is well established that such a
lithologic and drainage distribution leads to repeated stream captures when
headwardly eroding tributaries intersect the position of the master stream (Rich
1935; Mackin 1936, 1937; Hunt et al. 1953; Hack 1960b; Denny 1965; Ritter
1967, 1972). The mountain stream is diverted into a lower tributary valley
and is contained there until the process functions again.
The sudden influx of coarse load into the valley of the capturing tributary
produces an untenable fluvial condition because the master stream cannot
transport its oversized debris on the low valley gradient established by the
tributary. The obvious result is filling of the valley until the gradient increases
to an incline capable of transporting the mountain load under the prevailing
discharge. Subsequent downcutting, often along the valley side, produces a
depositional terrace (Ritter 1972). The enigma of depositional terraces formed
in this manner is that the eroded surface beneath the gravel was formed b>
one river (the tributary) while the filling was caused by another (the mountain
stream).
Figure 7.13.
Stages in the development of a
rock-cut terrace. (A) During low
water stages, fine sediment (f) is
deposited during normal flow and
coarse sediment (c) is deposited
at the end of a high water event.
(B) High water stage entrains all
the channel sediment and scours
the underlying bedrock before
coarse detritus is deposited again
on the channel floor (Source:
Mackin 1937 The Geological
Channel
Society of America)
gravel
(A) (B)
Erosional Terraces One sometimes wonders if any aspect of fluvial processes

escaped the genius of G. K. Gilbert. The following statement is contained in
his remarkable discussion of the origin of floodplains:
. . . The deposit is of nearly uniform depth, descending no lower than

the bottom of the water-channel, and it rests on a tolerably even
surface of the rock or other material which is corraded by the stream.
The process of carving away the rock so as to produce an even surface,
and at the same time covering it with an alluvial deposit, is the process
of planation. (Gilbert 1877, pp. 126-27)
Clearly Gilbert presupposed the process of lateral erosion long before any
detailed understanding of mechanics existed. Certainly he provided a the-
its
oretical base for the early analyses of fluvial terraces, and his thinking prob-
ably represents a cornerstone in the classic model of rock-cut terraces developed
later by Mackin (1937) and illustrated in figure 7.13.
As we have said, erosional terraces are those in which lateral erosion is
the dominant process in constructing the tread. Mackin (1937) presented an
excellent description of the terrace origin (see that work for details of the pro-
cess; the mechanics were described in the previous section). Briefly, as rivers
migrate across the valley bottom, they erode one bank while simultaneously
depositing point bar debris near the other. The bar sediment later becomes
the capping terrace alluvium. It is usually thin and of constant thickness, and
it sits on a flat surface eroded across the underlying bedrock or sediment. The
buried surface is carved during floods when scouring penetrates the debris
lying on the channel floor. For this to occur, the scouring depth of the river
must be great enough to remove the entire pile of channel alluvium and expose
the suballuvial material to short-lived erosion (fig. 7.13). Continual shifting
of the channel position back and forth across the valley, combined with the
occasional scouring, creates beneath the alluvium the bevelled surface that is
a mirror image of the plane surface on top of the deposit.

Chapter 7 272
The sheet of alluvium is almost always present in an erosional terrace, but

it is not a prerequisite and is certainly not the paramount characteristic of the
feature. That role falls to the laterally eroded surface. Thus, it is probably
acceptable to ignore the alluvium and consider the cut surface to be the ter-
race tread. Like any approach to terraces, this may or may not lead to diffi-
culties in the field, depending on the particular situation.

Erosional terraces are normally thought of as the "equilibrium" model of
the terrace line. It would seem that the development of a tread surface by
lateral planation should require not only time but also a long period of stability
during which base level and channel functions are constant, and no vertical
disruptions by filling or cutting occur. Nonetheless, even this logical rule of
thumb has exceptions. For example, near Pyramid Lake (Nevada), the Truckee
River has formed six erosional terraces during a period when its base level,
represented by the lake, was rapidly declining (Born and Ritter 1970). The
highest and oldest terrace formed sometime between 1925 (when its level was
beneath the lake), and 1938, when aerial photography showed it to be a well-
developed landform. It now stands approximately 10 m above the Truckee
River, which is rapidly downcutting to keep pace with the declining lake level.
Apparently each terrace was formed during one major flood when the river,
in high flow, was able to erode laterally at a dramatic rate into its unresistant
banks. The terrace levels are carved into noncohesive lake sediments recently
exposed because Pyramid Lake, in hydrologic imbalance, has dropped almost
25 m in this century. Thus, where banks are easily eroded, time and stability
are not essential factors in the formation of erosional terraces. In fact, it is
difficult to imagine any geomorphic setting in greater disequilibrium than the

Truckee River system near Pyramid Lake.
Terrace Origin and the Field Problem

Understanding local terraces and establishing a regional pattern of terrace
development are not only basic in historical geomorphology, they are also useful
in providing information for regional planning, land management, water supply,
and locating sand and gravel for building materials. Acquiring such knowl-
edge, however, is a painfully slow procedure requiring field study and corre-
lation of surfaces within a valley or between valleys. Determining terrace
origins is not so easy as we would like to think. Although we postulate guide-
lines for recognizing terraces of different origins, terraces in the real world
develop in such a variety of ways that the exceptions almost become the rules.
Basically, terraces are terraces are terraces — and we should probably not gen-
eralize about features that defy generalization. Each terrace sequence must
be examined according to its own geologic, climatic, and tectonic setting
without preconceived ideas about its origin.
Using terraces to interpret geomorphic history is a monumental task for
several fundamental reasons. First, terraces are rarely preserved intact along
the length of a valley but instead are segmented into isolated and physically
separated remnants, often kilometers apart. Reconstruction of the original
longitudinal profile of the terrace surface requires correct correlation of the

remnants, and every method used in that procedure is burdened with funda-
mental assumptions that may be invalid in certain situations (for details see
D. W. Johnson 1944; Frye and Leonard 1954).
Second, more than one terrace can result during a period of downcutting.
This indicates that entrenchment, representing the response to a threshold-
exceeding change, is not a continuous, unidirectional erosional event that re-
sults in a lowered river gradient. Instead, as discussed in chapter 1, the re-
sponse complex (Schumm 1977). It involves pauses in a downcutting phase
is
during which the river may form erosional terraces by lateral planation (Born
and Ritter 1970; Ritter 1982) or depositional terraces by valley alluviation
(Womack and Schumm 1977). The point here is that complex response results
in multiple terraces formed during the adjustment to a single, equilibrium-
disrupting event. This complicates the historical interpretation.
Third, it may seem that erosional terraces are really not that much dif-
ferent from depositional terraces. After all, both are usually covered with al-
luvium, and if one walked across that alluvial surface there would be nothing
to indicate what type of terrace lay beneath. Nonetheless, a real and very im-
portant difference does exist — one that cannot be disregarded or minimized.
When an erosional terrace forms, the capping alluvium is deposited at the
same time that the underlying surface is eroded. In significant contrast, the
surface beneath a depositional terrace was present before the influx of the
alluvial fill; a finite time gap separates the deposition from the cutting of the
underlying surface. Failure to recognize this subtle distinction between ero-
sional and depositional terraces can lead to drastically different reconstruc-
tions ofgeomorphic history. For example, the differing interpretations of the
terrace sequence in. the Bighorn Basin of Wyoming show the problem well
and demonstrate the difficulty of obtaining sufficient field data for interpretive
purposes.
Mackin (1937) divided the Cenozoic history of the Bighorn Basin into
two major phases: (1) a long period of basin filling throughout most of the
Tertiary, followed by (2) rejuvenation and basin excavation that has continued
to the present. Within the basin are a series of terrace levels standing at el-
evations of from 330 to 6 meters above the present rivers. Mackin felt that
each level represented a rock-cut bench formed when downward excavation
ceased and allowed lateral erosion to become the dominant fluvial process. The
evidence supporting this interpretation seemed to be clear. Where later en-
trenchment exposed the terrace gravels, they were thin, constant in thickness,
and resting on a flat, truncated bedrock surface (fig. 7.14A). All the ingre-
dients of a rock-cut terrace were observed, and to interpret them as such was
certainly reasonable.
In a later study, however, Moss and Bonini (1961) were able to gain ad-
ditional information about the subsurface framework of several key terraces
by running seismic profiles across the features perpendicular to the axis of the
Shoshone River valley. Instead of the expected flatness, the bedrock surface
beneath the alluvium showed considerable relief, and in places the gravel
Chapter 7 274
Figure 7.14.
Interpretations of the Cody
Terrace near Cody, Wyo (A) As
rock-cut terrace, based on
observed alluvial thickness (B) As
depositional terrace, based on
seismic profiles across the terrace
(Moss and Bonini 1961)
Thickness from
seismic profile
Observed by Mackin by Moss and Bonini
(A) (B)
Figure 7.15.
Difficulty in interpreting terrace
origin from field data If
downcutting exposes only part of
fill, terrace may appear to be rock-
Appears to be
cut Data across the terrace are
rock-cut bench
needed to determine true
thickness of the fill.
Shows fill of
depositional terrace
thickness was well beyond a reasonable scour depth for rivers of this type (fig.
7.14B). They interpreted these characteristics to mean that the surface be-
neath the gravel represented the valley topography that existed before it was
buried by the influx of a later fill. They concluded that the terraces are de-
positional and that the was outwash from glaciers in the nearby Absaroka
fill
Range. If all the terraceshad this origin, the history of the basin would change
significantly. The general excavation phase, in this interpretation, was period-
ically interrupted by glaciofluvial filling of the valleys, not by valley widening.
In addition, considerable time elapsed between erosion of the underlying bed-
rock surfaces and creation of the terrace treads.
It is instructive to note that the Moss and Bonini interpretation was made
possible by techniques and data not available to Mackin in 1937 or, for that
matter, to most of us today. What Mackin observed was the edge of a fill where
it intersected an eroded valley (fig. 7. 15). The much-needed third dimension
across the terrace could be reconstructed only with the proper field equipment
and approach.
The topography of almost every region reflects an adjustment between dom- Piedmont
inant surficial processes and lithology. When the rocks have diverse resis- Environment: Fans
tances, geomorphic processes tend to maximize the relief between regions of and Pediments
greatest and least resistance. Nowhere is this more apparent than in areas
where mountains and plains adjoin, especially where the climate is arid or the
region has undergone recent tectonism. Aridity serves to buffer the smoothing
effects of vegetation; vertical tectonic activity accentuates relief by bringing Mountain
more resistant basement rocks toward the surface, where they are commonly
etched into the cores of topographic mountains.
The sloping surface that connects the mountain to the level of adjacent
plains is the piedmont. It extends from the mountain front to a floodplain or
playa, either of which can mark the base level for geomorphic processes that
function on the piedmont surface (fig. 7.16). Piedmonts consist of a number
of geomorphic landforms, but most commonly they are composed of eroded
bedrock plains called pediments and depositional features called alluvial fans.
Figure 7.16.
The piedmont area occupied by either of these
relative percentage of the total
Physiographic components of a
features probably depends on the unique combination of local geomorphic mountain-basin geomorphic
variables. system
Alluvial Fans
Alluvial fans have been investigated most extensively and in great detail in
regions of arid or semiarid climate. This does not mean that fans are absent
in other climatic zones. On the contrary, humid climate fans and their deposits
have been examined in such diverse settings as humid-glacial (Boothroyd and
Ashley 1975), humid-periglacial (Ryder 1971a, 1971b; Wasson 1977), humid-
tropical (Mukerji 1976; Wescott and Ethridge 1980) and humid-temperate
(Hack and Goodlett 1960; Williams and Guy 1973; Kochel and Johnson 1984).
Fans developed in every climatic setting are linked together by a similar plan-
view geometry, but other aspects of morphology and depositional processes
may vary considerably (table 7.3).
Schumm (1977) suggests that fans can be of two major types. Dry fans
are those created by ephemeral flow, and wet fans are developed by perennial
stream flow. Clearly, the mode of formation has a climatic connotation be-
cause ephemeral flow is normally associated with low groundwater tables and
spasmodic rainfall, conditions typical in arid climates. Wet fans are obviously
more dominant in humid where perennial stream flow is the norm.
climates,
In spite of their recognition in all environments, fans so dominate the
piedmont zone in arid climates that most of our detailed understanding of fan
development derives from studies in that setting. Thus, unless specified, the
treatment that follows refers to arid climate (dry) fans.
Alluvial fans are one end of an erosional-depositional system, linked by a
river, in which rock debris is transferred from one portion of a watershed to
another. Fans are largest and most well developed where erosion takes place
in a mountain and the river builds the fan into an adjacent basin. Deposits
tend to be fan-shaped in plan view and are best described morphologically as
Chapter 7 276
Table 7.3 Generalized characteristics of all uvial fans formed in different environments.
Parameter Arid Fans Humid-Glacial Humid-Tropical Humid-

Virginia
Fans Fans Temperate Fans
Fan Morphology
Plan View Broad fanlike Broad fanlike Broad fanlike Broad fanlike to
symmetrical symmetrical symmetrical elongated
Axial Segmented Smooth Smooth Segmented

Profile (20-100m/km) (1-20m/km) (40-100m/km)
Thickness Up to 100'sm Up to 100s m Up to 100s m 5 m to 20 m

Area Small Very large Large Small
Depositional Processes
Major Debris flow Braided stream Braided stream Debris flow

Processes Braided stream Debris flow (avalanche)
Sheet flood
Sieve flood
Return 1-50 yr Discrete events 0-few days Seasonally Seasonally constant to 3000-6000 yr Discrete
Interval constant discrete events
Fan Area 10-50% 80-100% 30-70% 10-70%

Activated
Triggering Heavy rain Meltwater Heavy rain Heavy rain

Processes Snow melt Outwash Monsoon Hurricane
Discharge Flashy Seasonal Seasonal Flashy
After Kochel and Johnson 1984 Used with permission of Canadian Society of Petroleum Geologists
a segment of a cone radiating away from a single point source (fig. 7.17). The
point source represents the spot where the master river of the watershed
emerges from the confines of the mountain; it doubles as the apex of the con-
ical shape. The point source can also shift away from the mountain front to
a position well down the original fan surface if that surface has been en-
trenched at some time during its development. In those cases, the mountain
stream, still occupying a confining channel, traverses a portion of the older
fan material. The stream eventually emerges downfan as the point source for
a still younger fan. Adjacent fans often merge at their lateral extremities: the
individual cone shape and a rather nondescript deposit is formed cov-
is lost,
ering the entire piedmont. These coalesced fans are commonly referred to as
bajadas, alluvial aprons, or alluvial slopes.
The surfaces of fans can often be subdivided into major zones (Denny
1 965, 1 967) called modern washes, abandoned washes, and desert pavements.
which reflect their participation in modern fan processes (fig. 7.18). On the
Shadow Mountain fan in Death Valley, washes make up about two-thirds of
the surface area, but only a few contain unweathered gravel and accommodate
present-day floods. These modern washes are the primary areas of deposition
on the fan surface. Most washes have scrub vegetation and gravel coated with
desert varnish in their channels, indicating that they are abandoned channels
't$ .h
IV
7 \
Figure 7.17.
and have not been flooded for considerable time, perhaps several thousand
Map view of topography on
years. Desert pavements are surfaces of tightly packed gravel that armor, as typical alluvial fan Note
well as rest on, a thin layer of silt, presumably formed by weathering of the downslope deflection of contours
indicating convex cross-fan
gravel. Pavements are the primary erosional areas of a fan. They have not
profile
received sediment for a long time, as evidenced by the thick varnish coating
the pebbles, the pronounced weathering beneath the silt layer, and the striking
smoothness of the surface, due to obliteration of the original relief by down-
wasting into depressions. Pavements characteristically are cut by gullies that
head within the pavement area and may be meters or even tens of meters deep.
Because the gullies carry a locally derived fine-grained load, they often meander
and, importantly, may stand at lower elevations than adjacent modern washes
that head in the mountains.
The distribution of pavements and their gullies in relation to modern
washes creates a geomorphic situation quite similar to that discussed earlier
for the creation of terraces in a piedmont region (see fig. 7.12). Flow in the
modern washes is periodically diverted into the gullies, changing part of the
desertpavement area back into an active wash and shifting the position of the
modern washes. Simultaneously, the lower segment of the captured wash is
abandoned and, over a long period of time, imperceptibly converts to a desert
pavement.
Chapter 7 278
Figure 7.18.
(A) Large alluvial fans at the base
of the Panamint Range in the
north end of Death Valley, Inyo
County, Cal. (B) Components of
fan Death Valley region.
in
(Adapted from Denny 1965)
Ste^fc
:
Gravel in abandoned washes.
:
Varnish cover.
Gravel in desert pavement areas.
Weathered and varnished.
Gravel in modern washes.
Unweathered and no varnish.
Evaporites on floor of Death Valley.
Undifferentiated gravel.
Sedimentary and metasedimentary

rocks of mountains.
600
Figure 7.19.
Straight segments of several
on an alluvial fan
radial profiles
Concave profile stems from lower
gradient on each basinward
segment of the fan. (After Bull
400 1964)
200
mmmmmmmi^m
4 6 10
Scale (miles)
Fan Morphology The longitudinal slope of an alluvial fan generally de-

creases downfan even though its precise value at any point depends on the
load-discharge characteristics of the fluvial system. Near the mountain front,
slopes are commonly very steep, although they probably never exceed 10°
(Cooke and Warren 1973). Fans gradually flatten to their lower extremity,
called the toe, where gradients may be as low as 2 m per kilometer (« 10 ft/
mi). The steepest gradients are usually associated with coarse-grained loads,
low discharges, high sediment production in the source area, and transport
processes other than normal streamflow (Blissenbach 1954; Bluck 1964; Bull
1964a; 1964b, 1968; Hooke 1967, 1968; Hooke and Rohrer 1979). These fac-
tors often conflict in the same region. In Fresno County, California, for ex-
ample, fans derived from basins underlain by mudstones or shales are 33 to
75 percent steeper than fans of the same size related to sandstone basins (Bull
1964a, 1964b). The low gradient expected because of the small particle size
is by a high rate of sediment production. Fan gradients may also be
offset
related to parameters of drainage basin morphometry (Melton 1965a); how-
ever, most statistical measurements probably reflect only the more basic con-
trolling factors. Absolute slope values at any given point, then, may represent
a myriad of controls in the erosional-depositional system and other factors
such as its position relative to the fan axis (Hooke and Rohrer 1979).
Two slope characteristics deserve special attention. First, the gradient of
most fans near the mountain front is approximately the same as that of the
mountain river where it merges with the fan apex. Deposition on the upfan
surface, therefore, is not initiated by a dramatic decrease in gradient as the
master river passes from the mountain onto the fan. Second, although fans
are concave-up from the apex to the toe, their longitudinal profiles are usually
not a smooth exponential curve. Instead, on many fans the concavity stems
from a junction of several relatively straight segments, each successive downfan
link having a lower gradient (fig. 7.19).
Chapter 7 280
Figure 7.20.
Relation of fan area to drainage-
basin area for anumber of fans in
Californiaand Nevada See Bull

1968 for sources of data and
equations of the regression lines
0.01
1.0 0.1 10 100 1000
Drainage-basin area (A D ), in square miles
The changes in fan slope, represented by individual segments, are genet-

ically related to changes in the channel of the trunk river upstream from the
fan apex. For example, in many fans intermittent uplifts have increased the
stream gradients, and in response to each event, a new fan segment has formed,
gradually adjusting its slope until it approximates the newly formed steeper
slope of the trunk river (Bull 1964a, 1964b; Hooke 1972). Under this partic-
ular control, each segment as one moves upfan is steeper and younger, and its
deposits are graded to the level of the next lower segment. Segmentation may
also result from climatically induced changes in the load-discharge balance,
and the segments do not always become steeper toward the apex. Thus the
overall longitudinal profile may be very sensitive to historical changes in the
balance between the erosional and depositional parts of the system (Bull 1964.
1968).
It is now firmly established that the area of a fan is statistically related
by a simple power function to the area of the basin supplying the sediment
such that
Af = cA d ",
where Af is the area of the fan and Ad is the area of the drainage basin. The
exponent n is the slope of the regression line in a full logarithmic plot of the
two variables; it measures the rate of change in fan area with increasing
drainage basin area. The coefficient c indicates how much the fan "spreads
out."
Bull (1968) showed that the relationship is generally very similar for a
group of fans representing a variety of environments in the western United
States, and the mean value of n is approximately 0.9 when A, and A d are mea-
sured in square miles (fig. 7.20). The coefficient c, however, seems to varj
widely, reflecting the effect on fan dimensions of geomorphic factors other than
drainage basin size. Chief among these are climate, source rock lithology, tec-
tonics, and the original space available for fan growth in the collecting basin.
For example, Hooke and Rohrer (1977) suggest that c probably relates to the
competition for space in the depositional zone. A particular fan that receives
a largevolume of sediment from its drainage basin would tend to thicken faster
than neighbor and spread outward at the expense of the neighboring fan
its
area. Thus, even where drainage basins in the same region have equal areas,
the areas of their fans may differ by as much as an order of magnitude if these
other characteristics differ greatly.
Several specific examples demonstrate the effect of these factors and the
significance of fan morphometry. Bull (1964) showed that fans derived from
basins underlain by fine-grained sedimentary rocks are almost twice as large
as those derived from sandstone basins of equal size. The regression lines have
approximately the same slope (« = 0.91 and 0.98), but the effect of particle
size shows up in the value of the coefficient c, which varies from 0.96 for sand-
stone to 2.1 in the mudstone drainage basins. Actually, a high density of joints
and fractures is probably more important than the texture of the rocks. Highly
fractured shales will weather into small chips and will be easily eroded. This
produces the large loads and high values of c (Hooke and Rohrer 1977). The
effect of tectonics is revealed in the fans of Death Valley, where eastward tilting
of the valley permitted fans on the west side of the valley to grow larger while
those on the east side were stunted by burial beneath the playa (Denny 1965).
The c values are 1 .05 for fans on the west side of the valley and 0. 1 5 for those
on the east side.
Processes, Deposits, and Origins Any model that proposes to explain the
geomorphic meaning of alluvial fans must be based on discernible facts con-
cerning both fan morphology and the processes that function in the fan system.
Evidence that some of the pertinent facts are still missing can be found in the
unresolved controversy about the stability or instability of modern fans. Al-
though fans in the California-Nevada region have been studied in greater de-
tail than anywhere else, little agreement exists about their equilibrium
conditions. Some authors view fans as steady-state forms, neatly adjusted and
in a continuing dynamic equilibrium (Denny 1965, 1967; Hooke 1968). Others
see them as actively growing (Beaty 1970) or being dissected (Hunt and Mabey
1966), processes that presumably indicate that fans may be approaching an
equilibrium condition but have not sensibly attained it. Bull (1975b) cites fans
as features that do not attain a steady-state condition but instead develop under
the control of allometric change (see Bull 1975a, Bull 1975b). Still another
interpretation is that fan characteristics can only be explained by cyclic changes
in processes initiated by climatic fluctuations (Lustig 1965), and so every fan
has vestigial properties unrelated to the present conditions. Facing these widely
divergent viewpoints, it seems logical to briefly consider the evidence that led
these scientists to such varied conclusions.
Chapter 7 282
Deposits and Depositional Processes The movement of sediment from source

areas to depositional sites involves a variety of flow types, ranging from highly
viscous debris or mud flows to normal water flow. The type of flow during any
given event depends primarily on the lithology of the basin and its degree of
weathering, and secondly on the magnitude of the precipitation causing the
flow. The ephemeral nature of flow in arid regions results in spasmodic rather
than continuous deposition, and the depositional site changes repeatedly. Only
a limited portion of the fan surface is occupied by flow and undergoing de-
position at any given time.
As flow leaves the confines of the trunk channel, deposition is initiated by
changes in hydraulic geometry, not by a sudden decrease in gradient (Bull
1964). Generally, when the flow becomes unconfined on the fan surface, the
width increases so dramatically that both depth and velocity decrease to a
level where the flow can no longer transport the load. In segmented fans, how-
ever, areas close to the apex are commonly occupied by channels called fan-
head trenches that are incised 7-12 m below the level of the fan surface (Bull
1964a, 1964b). These may connect to the trunk river in such a way that flow
remains contained and is transmitted far downfan before it is freed to increase
its width. The effect of changing hydraulic geometry is reinforced by a loss in
discharge if the fan surface is permeable and water seeps into the underlying
deposits. It is not unusual, for example, for the entire flow to disappear un-
derground before it can traverse the length of the fan.
The deposits of any single flow usually form as narrow tongues, possibly
up to several kilometers long, but normally only 120-700 meters wide (Bull
1968). The length of each deposit probably depends on the viscosity of the
flow, the permeability of the surface, and the distance downfan that the flow
is held in a distinct channel. In many cases, the flow at first follows well-
established channels, but at some point along its length overtops the banks
and spreads outward as diffuse flow. In the case of water flows, lateral shifting
of the loci of deposition allows the braided stream system to deposit a sheet
of poorly bedded sand and gravel in which individual beds can be traced lat-
erally for only short distances. This sheetlike configuration may be interrupted
by thicker deposits that represent an occasional channel entrenchment into
the fan surface and subsequent backfilling. Deposits within these larger chan-
nels are generally more coarse-grained. However, even within incised, active
washes a microtopography may exist that is directly related to variations in
textural properties of fan deposits (Wells 1977). Channels, which are posi-
tioned in topographically lower portions of the wash, are floored with coarse
sediment. Higher parts of the wash consist of berms that flank the channel
and are composed of fine-grained sediment.
Debris flows or mudflows usually follow more well-defined channels be-
cause the confining limits of the channel ensure the depth of flow needed to
offset the high viscosity of the fluid. During transport, however, debris flows
also may overtop banks and spread out as sheets (Bull 1963). Debris flows are
so dense and viscous that only the very largest particles can settle from the
mass during flow. Nonetheless, they are capable of transporting extremely large
boulders for considerable distances on lower gradients than normal streamflow

would require. Their high viscosity, however, effectively restricts the distance
of transport (in comparison with water flow), and their forward movement
may simply stop even though they are still confined in a channel. Therefore,
deposits from debris flow are poorly sorted with boulders embedded in a fine-
grained matrix; in contrast to water-transported sediment, they are usually
lobate and have well-defined margins often marked by distinct ridges.
Some fans are built almost entirely by debris flows (Beaty 1970), al-
though their deposits may be reworked almost immediately by normal stream-
flow. Beaty (1963) reports debris-flow deposits being dissected within 48 hours
of their deposition by streamflow that began in the same storm but continued
after the debris sediment had been deposited. Beaty (1970) also found that in
the Milner Creek (White Mountains, California) fan, debris was derived from
the floor of the trunk channel. Accumulation of 3-7 m of sediment is needed
in the channel to provide the volume found in each major debris-flow deposit
on the fan. Sediment production in the drainage basin, therefore, must be rapid
enough to collect sufficient amounts on the channel floor before successive high-
magnitude storms remove it. If sediment production is too slow, not enough
particulate matter will accumulate within the recurrence interval of major
floods to produce a debris flow, and the discharge delivered to the fan surface
will be in the form of normal streamflow. Some debris flows are initiated by
landslides (Johnson and Rahn 1970).

In fans composed of coarse-grained deposits, large discharges may infil-
trate before crossing the entire fan (Hooke 1967). Under these conditions,
coarse sediment may be deposited in lobate masses called sieve deposits. These
resemble debris-flow deposits but lack primary fine-grained material and are
highly permeable. Sieve material can be deposited when the flow is confined
or unconfined, but only when the surface is permeable and the flow does not
contain fine-grained sediment.
Entrenchment and Location of Deposition It is axiomatic that lateral mi-

gration is involved in developing the convex cross-profiles and the plan-view
shape of alluvial fans. It is equally important to recognize that the loci of de-
position also migrate along radial lines during fan development. This longi-
tudinal shifting accomplished by entrenchment or backfilling of the main
is
channel that extends from the source area onto the fan. Trenched channels
have lower gradients than the fan surface, and so they are deep near the fan
apex and become progressively shallower downfan until they finally emerge
at the surface. In segmented fans, the trenches tend to incise until they have
the same gradient as the adjacent downfan segment (fig. 7.21). Entrenchment
is significant in that it provides an explanation for abnormal distribution of
particle sizes on the fan, since even coarse sediment can be placed far downfan
if entrenched channels confine the entire flow. In addition, entrenchment tends
to enlarge fans because moderate flows that could not traverse the fan uncon-
fined are transported farther in well-defined channels, resulting in deposition
and fan construction in downslope areas.
Chapter 7 284
Scale (miles)
Figure 7.21.
4
Fanhead trenches in segmented
fans Entrenched channels are
400
attaining the same gradient as the
adjacent downfan segments
(From Bull 1964)
200
ID
o
0}
Vertical exaggeration x 211
ii i i I I I I I wmmi^^^mm
5 10
Scale (miles)
Entrenchment can be either temporary or permanent, and distinguishing

between the two seems to be critical in the analysis of fan origin. Many fan-
head trenches appear to be rather ephemeral; that is, they have evidence of
alternating episodes of trenching and filling. Temporary entrenchment some-
times occurs when debris-flow deposits plug the channel, and the flow shifts
laterally to a new position where entrenchment begins again. It also may be
that alternating trenching and filling are expected results of fan processes as-
sociated with changes in rainfall intensity (Bull 1964a, 1964b) or normal al-
ternations of debris flows and water flows (Hooke 1967). In any case, temporary
trenching seems to be a common process on most active fans and can probably
be explained in terms of local conditions. For example, experimental studies
and field observations of wet fans have increased our understanding of tem-
porary trenching (see discussion in Schumm 1977). After the source area
channel and the surface of a wet fan are accordant, water and sediment are
spread as a sheet over most of the area near the fan apex. Downfan deposition,
however, occurs in numerous braided channels. The depositional setting is in-
terrupted when the fan slope near the apex reaches a threshold condition and
incision begins. This results in a fan-head trench, and flow becomes confined
within that channel rather than being widely disseminated over the apex zone.
As entrenchment migrates upstream into the source area, the trunk river is
rejuvenated, and the increased load derived from incision of the trunk river
channel is deposited in the fan-head trench. This in-filling raises the floor of
the fan-head trench until the threshold is reached again, and the process re- Figure 7.22.
Dissected fan in Rock Creek
peats itself. Thus, as Schumm
(1977) points out, the entire fan may continue
valley, Beartooth Mountains, Mont
to grow with lime, but the fan head experiences spasms of entrenchment during Grassy areas represent original
which sediment is reworked and moved farther down the fan. tan surface Tree-covered zones
are in portions of the fan that
In permanent entrenchment, channels are incised to depths that cannot
have been entrenched when the
be easily backfilled, often being cut to levels greater than 30 m below the fan master river (Rock Creek) cut to
surface. In addition, the impetus for the downcutting may be outside the fan lower level.
system and so cannot be explained by local fan processes. On the upper

itself
fan slope, permanent entrenchment may be caused by accelerated erosion, but

it also can occur naturally in a uniform environment if the trunk river con-
tinues to downcut within the mountain. In either case, the fan surface standing
above the trench is no longer involved with active fan processes; soils may
develop on the alluvium, and incipient drainage networks may be established.
If the basin of deposition is open and base level for the fan is the floodplain
of a through-flowing river, downcutting of that river may initiate a wave of

fan incision that is propagated upslope from the toe of the fan. In addition,
lateral migration of the master river can erode the toe of the fan and reju-
venate the streams crossing the fan surface. Eventually the entire fan is dis-
sected when the incision reaches the apex and captures the trunk river as it
emerges from the mountain. This process is especially effective on small valley-
side fans of glaciated valleys (fig. 7.22) and commonly relates to climatic fluc-
tuations and the glacial cycle (Ryder 1971a, 1971b).
Chapter 7 286
Pediments
Since Gilbert first described "hills of planation" in the Henry Mountains of
Utah, geomorphologists have been intrigued with their origin, and this fea-
ture, given the name pediment, has been discussed endlessly during the last
century. Our discussion of these interesting features will necessarily be brief,
but excellent reviews of the topic are available: Tator 1952, 1953; Tuan 1959;
Hadley 1967; Cook and Warren 1973; Twidale 1978. Definitions of the term
"pediment" are as numerous as the workers who have studied this feature,
and like the landform itself, descriptions range from general to rather graphic
and precise. For example, Denny (1967, p. 97) employs the term in reference
"to the part of the piedmont that is more or less bare rock surface." On the
other hand, R. U. Cooke (1970, p. 28) suggests that "pediments are composed
of surfaces eroded across bedrock or alluvium, are usually discordant to struc-
tures, have longitudinal profiles usually concave upward or rectilinear, slope
at less than 11°, and are thinly and discontinuously veneered with rock de-
bris."
In order to understand better what it is that we are discussing, and in

deference to clarity, it may be better to look briefly at those characteristics
that are universally recognized as salient properties of pediments rather than
to adopt a formal definition:
1. Pediments are erosional surfaces that abut against and slope away
from a mountain front or escarpment.
2. They are entirely erosional in their origin and commonly form in a
direction that diverges from the trend of the regional structures.
3. The surfaces are usually, but not necessarily, cut on the same rocks
that make up the mountain. They may truncate both bedrock and/
or alluvium, but they are best developed and preserved on bedrock,
especially resistant types such as granite or related crystalline
rocks.
4. Pediments may or may not have a thin covering of sediment which
presumably represents load that is in transit. This characteristic
has traditionally created problems because the question arises of
how much alluvium can be tolerated before it must be recognized
as a fan, younger in age than the pediment surface and therefore
divorced from the processes of pedimentation. It is applaudable,
then, that Cooke (1970) restricts the pediment to only that part of
the eroded surface not continuously covered by alluvium (fig. 7.23).
The erosional surface beneath the continuous debris cover is called
the suballuvial bench, a term first used by Lawson (1915), and the
cover itself is referred to as the alluvial plain. The pediment, then,
is bounded upslope by the mountain front and downslope by the
alluvial plain (fig. 7.23).
5. Pediments are usually found in arid regions, although most workers
would not restrict the processes of pedimentation to that climate.
Note that we said they are found, not formed, in arid climates.
Field data suggest that some pediments in the Mojave Desert may
be relict features that formed under a more humid. Tertiary
Watershed
Figure 7.23.
Landforms in the mountain-basin
geomorphic system: = M
mountain area; Mf = mountain
front; P = pediment; PT =
piedmont plain; A = alluvial plain;
BLP = base level plain
Mountain/ piedmont
junction
Suballuvial bench
Bedrock Alluvium
climate, and the surfaces are at present undergoing virtually no

expansion under modern desert conditions (Oberlander 1972,
1974). L. C. King (1953) suggests that pedimentation may be a
basic geomorphic phenomenon, present in all hillslope
developments regardless of climate.
Morphology and Topography It is ironic that in spite of the singular atten-

tion devoted to pediments, a multitude of untested hypotheses exist concerning
the processes of pedimentation but an amazingly skimpy pool of reliable data
to support them. After a century of study, there is still confusion and lingering
disagreement about every aspect of pedimentation. Cooke and Warren (1973,
p. 188) express this succinctly in their description of the topic as "a subject
dominated by almost unbridled imagination." Cooke (1970) suggests three

reasons for the failure to resolve these differences of opinion: (1) we have not
viewed pediments as part of an erosional-depositional system but have studied
only the pediments to the exclusion of other related forms; (2) we have not
collected the precise data needed to understand the system; and (3) we have
been overly concerned with general, evolutionary hypotheses of pediment for-
mation and, in many cases, have deduced processes from the genetic model
rather than employing direct measurement and observation. Regardless of these
past sins, we can make some definitive remarks about pediment morphology
and topography.
Size and Shape Pediments vary in size from less than 1 square kilometer
to hundreds of square kilometers, probably depending on fundamental geo-
morphic controls. Shape is also variable, with pronounced irregularities when
the rocks cut by the pediment surface have wide differences in resistance to
erosion (Hadley 1967). Generally they tend to be fan-shaped in plan view,
narrowing toward the mountain front and widening downslope (D. W. Johnson
1932; Rich 1935; Gilluly 1937). Across the pediment, the shape can be either
convex or concave.
Chapter 7 288
Figure 7.24.
Surface Topography Contrary to lay opinion, pediments are not monoto-
View across pediment surfaces to
Bear Peak, Boulder County. Colo .
nous, smooth, flat surfaces but are dissected by incised stream channels and
February 29. 1972 dotted with residual bedrock knobs, called inselbergs, that stand above the
general level of the pediment itself. Inselbergs have been investigated repeat-
edly with regard to their relationship to pedimentation (Twidale 1962, 1978;
Kesel 1973, 1977; Twidale and Bourne 1975; and many others). In some cases,
these residual hills might be the last unconsumed vestigages of a landscape
that has been totally pedimented. More likely, however, most inselbergs rep-
resent areas of rock that are more resistant to weathering and erosion (Kesel
1977; Twidale 1978).
The frequency and size of both incised valleys and inselbergs seem to in-
crease toward the mountain front, sometimes giving the topography the aspect
of gently rolling hills and valleys (Gilluly 1937). Some of the channels and
other depressions may be filled with alluvium up to 3 m thick (Cooke and
Warren 1973), giving the false impression that the bedrock surface is smooth
and perfectly planed.
The Piedmont Angle The upper boundary of the pediment is usually marked
by an abrupt change from the steep slopes of the mountain front to the low
declivity of the In plan view the boundary is usually linear,
pediment surface.
but embayments major valleys of the mountain front can give the trace
into
a rounded or crenulate appearance. The angle formed by the junction of the
two surfaces is the piedmont angle (fig. 7.24), and its development and main-
tenance have traditionally been cited as evidence in theoretical models of ped-
iment origin.
In detail, the piedmont angle can take the form of a narrow zone of intense
curvature rather than a distinct angle. Twidale (1967) reports mountain front
slopes of 22° changing to pediment gradients of 3° over a transition zone

100 m wide. Both the magnitude of the piedmont angle and the sharpness of
the angular relationship are probably related to structural or lithologic control
(Denny 1967; Twidale 1967; Cooke and Reeves 1972), but other processes
including weathering and several forms of corrasion have been suggested as
contributing or dominant factors. The early idea of Bryan (1922) that the
angular relationship represents an ajustment of the two slopes to the size of
debris they are required to transport cannot be accepted without qualification
sinceMelton (1965b) was unable to demonstrate a significant correlation be-
tween slope angle and the size of weathering products. At the present time,
no one set of processes adequately explains the origin and development of all
piedmont angles.
Slope The longitudinal profile of almost all pediments is slightly concave-

up, although local convexities do occur. Overall longitudinal convexities have
been suggested as a theoretical possibility if the suballuvial bench is also con-
sidered (Lawson 1915), but available observation and geophysical data
(Langford-Smith and Dury 1964) have not demonstrated the actual presence
of such a form. Slope angles on pediments range from 0.5° to 11° but seem
to average about 2.5°.
A prevailing perception concerning pediment slopes is that they are con-
trolled entirely by the size of the material they are required to transport, and
so it is widely accepted that they are "slopes of transportation" (Bryan 1922).
It is true that some studies have demonstrated a strong correlation between
pediment slope and the particle size in debris mantling the eroded surface
(Akagi 1980). However, detailed measurements show that the relationship is
not as straightforward as previously supposed (Dury 1966b; Cooke and Reeves
1972). Where particle size decreases in an orderly way downslope, the rate of
decline in pediment gradients is often greater than the rate of reduction in
size. Cooke and Reeves (1972) also found that only the largest particles showed
a consistent decrease with distance from the mountain front. Other statistical
parameters of size varied incoherently with distance and slope decrease. They
attribute these anomalous relationships to differing amounts of in situ sedi-
ment being added to the total load at any given sampling locality.
It might also be logically assumed that slope should be related in a sig-
nificant way to the area of the mountain drainage basin or to the length of
the pediment. It has been noted, for example, that pediments have lower gra-
dients where they are associated with large rivers or canyons. (Bryan 1922;
Gilluly 1937). Although data are limited, neither of these assumptions could
be substantiated in more recent morphometric studies of pediments in Cali-
fornia and Nevada (Mammerickx 1964; Cooke 1970). In addition, Twidale
(1978) showed that in sequences of pediments having the same source, ped-
iment slopes may decrease with time. The oldest and highest pediments have
steeper gradients than the younger pediments. Twidale stressed that in order
for this to occur the mountain front scarp must be notably stable. He also
demonstrated that slopes are related to structural and lithologic controls.
Chapter 7 290
Processes Any viable model of pediment genesis must explain the morpho-
logic and topographic elements of the pediment association. The pediment as-
sociation includes the pediment, themountain area adjacent to it, and the area
show considerable variation, it seems
of the related alluvial plain. Because these
unlikely that any one combination of processes will produce all pediments or
that any single evolutionary model will suffice. Over the years, a number of
models utilizing a few basic processes have emerged as the prime hypotheses
for pediment origin. However, pediment processes are not easy to study over
large areas or during short time periods, and so most of the proposed me-
chanics of formation are based on intuition rather than solid observational
data.
There is no doubt that water flows across pediment surfaces in several
different forms. The presence of drainage patterns on undissected surfaces
provides unmistakable proof that true streamflow occurs on pediments. Many
authors have suggested dominant process of pedimentation (Gilbert
it as a
1877; Paige 1912; D. W. Johnson 1932; Rahn 1966; Warnke 1969), especially
when the streams migrate laterally across the surface while simultaneously
planating the rocks. Although most experts recognize the efficacy of lateral
planation in developing part of the eroded surface, many believe this process
is incapable of producing and maintaining the piedmont angle
in areas remote
from the main drainage lines. Streamflow erosion of these interfluve regions
would require that rivers emerging from the mountains occasionally flow per-
pendicular to the pediment slope —
a maneuver that defies the law of gravity,
as Lustig (1969) reminds us.
Besides normal river flow, unconcentrated flows in the form of sheet and
rill or floods also traverse pediment surfaces. The phenomenon of sheet-
wash
was first observed by McGee (1897), who saw the flow as the erosive
flooding
mechanism in the formation of pediments. Although this concept was adopted
by some later workers (Lawson 1915; Rich 1935), it was refuted by others
who felt that sheetflow acting alone could not create large pediments, espe-
cially across resistant lithology, since the smooth surface itself is a necessary
prerequisite for the development of unconcentrated flow. In fact, observations
of storm runoffs in Arizona (Rahn 1967) indicate that large discharges on
pediments occur as streamflows rather than sheetflows. It may not, in fact, be
particularly important to know whether the flow at the head of the pediment
is an unconfined type such as sheetwash or whether it moves in small rills or
channels. What is important is that flow at the base of the mountain front
appears to be a capable transporting agent and therefore plays a significant
role in pedimentation by preventing the accumulation of debris there and so
perpetuating the character of the piedmont angle. This fact has been docu-
mented beyond question in the development of miniature pediments where the
underlying rocks are relatively nonresistant (Schumm 1962).
The processes of weathering also have been noted, with different degrees
of emphasis, as important factors in pedimentation. The objection that sheet-
wash may be ineffective as an erosive agent is overcome if the initial
partly
strength of the pedimented rock is lowered by weathering. Weathering profiles
of considerable thickness have been recognized on pedimented rocks (Mab-

butt 1966; Twidale 1967; Oberlander 1972, 1974), leading to the suggestion
that the rock surfaces beneath the weathered mantle, and even some subal-
luvial benches, may be produced by weathering rather than by water erosion.
Mabbutt ( 1 966) suggests that pediments developed on granite and related rocks
may be formed primarily by continuing subsurface weathering that, because
of slight differences in resistance, produces an uneven bedrock surface beneath
the mantle. The surface on top of the mantle is kept flat by temporary allu-
viation. In other rock types, weathering may be more rapid at the surface, and
the products are removed by sheetwash to maintain the relatively flat pedi-
ment surface. Some evidence exists to support the conclusion that subsurface
weathering is most pronounced at the junction of the pediment and mountain
front (Mabbutt 1966; Twidale 1967). Thus, headward extension of the ped-
iment and the maintenance of the piedmont angle may be intimately related
to the type of rocks involved and the efficiency of the weathering processes.
Formative Models Cooke and Warren (1973) point out correctly that the
criticalboundary in piedmont areas lies between zones that are primarily de-
positional and those that are predominantly erosional. The position of this
boundary and the presence or absence of pediments are determined by the
amount of sediment produced in the mountain relative to the ability of pro-
cesses to transport the material across the piedmont zone. If supply exceeds
transportation, the boundary may abut against the mountain front, and no
pediment will be found. In the opposite case, the boundary may be well down
the piedmont slope, and pediments will be present. If equilibrium exists be-
tween rates of supply and removal, the boundary will be stable and not nec-
essarily parallel to the mountain front. Once established, the boundary can
change its position in response to climatic or tectonic alterations. Pediments
may be modified after their formation by regrading, or they may be isolated
from the processes that are adjusted to the new piedmont setting, as when the
surface is is buried beneath an alluvial cover. With these pos-
entrenched or
mind, Cooke and Warren (1973) categorized situations under which
sibilities in
pediments form, and their designations are presented here with some modi-
fication and addition.
In one situation, which we can call headword pediment extension, ero-
sional processes allow the pediment to expand into the area of the mountain.
The boundary may remain constant, or it may move toward the moun-
alluvial
tain in phase with the migratingmountain-pediment junction, depending on
whether the sediment derived from headward erosion can be transported from
the system. Cooke and Warren (1973) recognized three formative models as
promoting headward growth: the lateral planation hypothesis, the parallel re-
treat hypothesis, and the drainage basin hypothesis. The lateral planation hy-
pothesis was developed most fully by D. W. Johnson (1932), who believed that
rock planes (pediments) are a natural consequence in arid regions. According
to Johnson, rivers near the mountain front are essentially graded and so heavily
loaded that they cannot cut vertically but tend to migrate and erode laterally.
Chapter 7 292
Mountain
Figure 7.25.
Retreat of pediment surface
parallel with expanding plain
Piedmont angle at V is
maintained Ratio of hillslope
erosion (T) to pediment erosion (t)
must be constant, and the

rectilinear profile is perpetuated
Extreme migration is presumed to trim back interstream bedrock spurs,

forming and maintaining the piedmont angle. In response to changes imposed
on the system, the rivers will perpetuate their equilibrium condition by re-
grading the pediment slope.
The parallel retreat mechanism was introduced by Lawson (1915) and
championed by Rich (1935). It requires that after a mountain front achieves
its diagnostic slope, weathering and erosion will maintain that declivity by
making the surface retreat parallel to itself. The bedrock bench, produced as
the front recedes, is swept clean by rill and sheetwash, processes that keep the
angle between the mountain front and the original piedmont slope at a con-
stant value. A slightly different mechanism of parallel retreat occurs in gra-
nitic regions where subsurface weathering may continue to level the rock
surface at depth (Ruxton and Berry 1961; Mabbutt 1966). Mabbutt called
this process mantle-control planation. It includes slight back-trimming at the
hill base that stabilizes the piedmont angle and extends the pediment toward
the mountain. It may also construct rectilinear profiles across the mountain
front-pediment boundary (Ruxton and Berry 1961), which may also be pre-
served by back-wearing (fig. 7.25).
The drainage basin hypothesis rises from the apparent failure of either
of the above models to explain the total morphology along most mountain
fronts. There appears to be no dramatic change in gradients where master
channels cross the mountain front onto the pediment surface, indicating the
viability of lateral planation there and negating the necessity of parallel re-
treat. Contrariwise, the interfluve areas that do exhibit the pronounced pied-
mont angle are far from the main valley and, for reasons stated before, probably
cannot be trimmed back by lateral corrasion. The drainage basin hypothesis,
therefore, represents a compromise between the two extremes of the other the-
ories. It recognizes that lateral planation will be dominant along the main
drainage line, while interfluve areas evolve by weathering and wash. The rec-
ognition that planation, weathering, and wash all are involved in pediment
mechanics is not new. It was first espoused in early investigations (Bryan 1922,
1935; Gilluly 1937; Sharp 1940) and more recently by Lustig (1969).
A second situation may exist
where there is equilibrium between the suppl\
of debris to the piedmont and removal to the base level zone. In this case,
its
piedmonts consist of both depositional features and pediments that are con-
tinuously and simultaneously being formed and destroyed. Theoretically, the
amount of sediment accumulated in fans in one part of the piedmont is equal

to the amount removed somewhere by pedimentation. Stream piracies,
else
due to the contrasting gradients between mountain and piedmont streams, re-
peatedly shift the zones of erosion and deposition. Gully erosion by streams
heading in the piedmont area and flowing perpendicular to the mountain front
forms the pediment surface. It may subsequently be buried by alluvium after
piracy diverts the mountain stream onto the pedimented surface. This basic
process has been employed to explain piedmont geomorphology in a variety
of climatic regimes (Hunt et al. 1953; Hack 1960b; Denny 1965) and has been
formalized as a widespread, pediment-producing mechanism by Denny (1967).
In the third case, erosion on the piedmont exceeds the debris supplied
from the mountain, so the pediment-alluvial plain boundary is moved away
from the mountain front, and new pediment zones develop farther downslope.
We can call this process basinward pediment extension. Pediment zones al-
ready present before the boundary shift may be modified during the transition;
the most common evidence cited for basinward migration of the alluvial
boundary is the truncation of weathering or soil profiles developed on the ped-
iment surface. A significant corollary to this concept is necessary. Since a ba-
sinward shift of the alluvial plain may expose older pediment surfaces by
removing their sediment cover, some pediments may be truly relict features
that have no sensible relationship to present conditions. Cooke and Warren

(1973) refer to this as the exhumation hypothesis and cite Mabbutt (1966)
and Tuan (1962) as most recent proponents.
its
The and time in the meaning of pediments cannot be

effects of climate
ignored. Oberlander (1972, 1974) presented a thought-provoking argument
that the granite pediments of the Mojave Desert were formed in their entirety
before the region became as intensely arid as it is today. He concluded that
boulders included in the mantles covering the pediment were originally iso-
lated as corestones in very deep chemical weathering and have reached their
present position by subsequent stripping of the weathered zone. The evidence
for this conclusion is quite strong. The bouldery mantle can be traced into a
well-developed weathering profile, including corestones in a grus matrix, which
is preserved beneath basalts dated at > 8 m.y. The pediment surface, there-
fore, most likely represents the weathering front formed under a semiarid Ter-
tiary climate. The mantle cover is not material in transit but is the remaining
part of a weathered profile that was progressively stripped after the region
became more arid in the late Pliocene and Quaternary. Oberlander's expla-
nation differs from others in that the pediment surface is not an exhumed,
rock-cut, suballuvial bench. The pediments, in fact, were not cut in rock but
were formed under a soil cover by the erosion of regolith that was being de-
veloped continuously at the weathering front. Slope retreat and maintenance
of the piedmont angle were probably the results of wash processes with par-
allel rectilinear back-wearing similar to that described by Ruxton and Berry
(1961). It is even possible that stripped pediment surfaces possess consider-

able relief and may be "born dissected" as was proposed earlier (Gilluly 1937;
Sharp 1940).
Chapter 7 294
Oberlander's proposal merits special attention as geomorphologists grope

for all-inclusive models. His work demonstrates that all landforms do not nec-
essarily yield significant relationships based on analyses of process and form.
The failure of pediments to reveal any morphometric consistency may be at-
tributed to the fact thatsome pediments developed under different morpho-
genetic conditions than those of the present. They are not equilibrium forms
but, in fact, may be relicts from the distant past that are disequilibrium freaks
in their modern surroundings. Oberlander states this message most clearly:
. . . granitic landscapes in the Mojave Desert have been evolving in a
sequential manner since the late Tertiary. This evolution has been
triggered by climatic change, and reflects the instability of a landscape
determined under conditions that no longer exist. Consequent changes
in morphology will cease to be diagnostic of climate change only when

all the weathered residuum inherited from the Tertiary morphogenetic
regime is stripped from elevated portions of the landscape and

transferred to adjacent basins of sedimentation. (Oberlander 1972,
p. 19)
The ideas proposed by Oberlander apply only to granitic terrains in the Mo-
jave Desert and may even be invalid in other parts of the arid, southwestern
United States (Kesel 1977). They do, however, demonstrate once again the
irrefutable importance of climate and geology in geomorphic systems and res-
urrect the spectre of time in considering equilibrium. Granites in an arid cli-
mate may require imponderable time spans before their external form reflects
an adjustment between processes and geology. Thus, it may be that the ab-
solute values of graded time (Schumm and Lichty 1965) are eminently de-
pendent on the systemic components.
Deltas Because sediment being transported by must ultimately come to rest,

rivers
deposition at or near a river mouth represents
a bona fide component of the
fluvial system. The most important geomorphic feature produced in that en-
vironment is called a delta. Deltas are also important sedimentary entities,
and geologists continue to study the depositional sequence and complex facies
relationships included in the deltaic mass. Our interest in deltas here is only
in the geomorphic processes that develop their form. The details of delta sedi-
mentation will be left to the sedimentologists.
The term "delta"
is usually applied to a depositional plain formed by a
river at its mouth, where the sediment accumulation results in an irregular

progradation of a shoreline (Coleman 1968; Scott and Fisher 1969). The fea-
ture was first named 2,500 years ago by the historian Herodotus, who noted
that the land created at the mouth of the Nile River resembled the Greek
letter A (delta). Modern deltas, however, display a great variety of sizes and
shapes. At the apex of a delta, the trunk river divides into a number of ra-
diating branches, called distributaries, that traverse the delta surface and de-
liver sediment to the delta extremities. In plan view some deltas look like alluvial
fans and in fact, a fan-delta generally means an alluvial fan prograding into
Delta plain
Topset beds
K
^ "\ A
7/
^-^<^/o.
°/>te
Figure 7.26.
Primary depositional environments
and their associated layering in a
Prodelta classic delta.
> ^
Foreset beds >r
Bottomset beds
(marine or lacustrine)
a body of standing water. Sedimentologists are keenly interested in portions

of the fan-delta that are continuously submerged and therefore have studied
in detail the marine deposits and processes effecting the feature. Geomor-
phologists are more interested in the feature as being the result of fluvial pro-
cesses and usually concern themselves more with the subaerial portion (for
example, see Morton and Donaldson 1978). Thus, even though fans and deltas
are related in form and process, they differ in several important respects:
(1) Deposition on deltas is due to a reduction of river velocity as the flow enters
a body of standing water; the standing water can be the ocean or lakes of any
size or origin. (2) Delta expansion in a vertical sense is finite, the base-level
water body being the approximate limit of upward growth. (3) The gradient
on the delta surface is notably flatter than that on most fans.
Sediment, Form, and Classification

Sediment deposited in deltas is usually fine sand or silt but, depending on con-
trolling variables, may occasionally contain gravel or clay. Deltaic sedimen-
tation and evolution were first described by G. K. Gilbert in his study of Lake
Bonneville. As envisioned by Gilbert and illustrated in figure 7.26, the feature
in its classic and unmodified state consists of a deltaic plain standing partly
above and below lake level. The plain is fronted by a delta slope that connects
its surface to the basin floor over which the delta is advancing. The basin floor,
called the prodelta environment, is composed of fine-grained marine or lacus-
trine sediments that were swept in The delta
suspension beyond the delta front.
plain is composed of a complex of nearly flat layers (topset beds) that truncate
the strata of the delta slope (foreset beds) as the delta progrades. The dip of
foreset beds varies widely. In marine deltas formed by large rivers, the beds
rarely dip more than 1 °, making it very difficult to recognize them in the field.
In small lakes, foreset beds may approach the angle of repose, and the se-
quence is easily discerned.
Most major rivers of the world develop deltas, and each has its own unique
properties reflecting some balance between the fluvial system, the climate, tec-
tonic stability, and the shoreline dynamics. Deltas, therefore, come in a mul-
titude of plan-view shapes, but in general several types serve as model forms
and are used for classification purposes. One or more of the controlling factors
is dominant in each of the major categories of deltas. High-constructive deltas
develop when fluvial action is the prevalent influence on the system (Scott and
Fisher 1969). As figure 7.27 shows, these deltas usually occur in one of two
—
Chapter 7 296
High-constructive deltas
Lobate
Lafourche
(Mississippi) Elongate
type
Modern
Mississippi
type
Miles 10
Distributary channel, levee, Delta plain (marsh, swamp, take, Delta front (including channel
Prodelta
crevasse splay interdistributary bay) mouth bar and sheet sands)
High-destructive deltas
Tide-dominated
Gulf of Papua type
Wave-dominated
Rhone type
Miles 10
Tidal .Channel and Delta plain (flood basin and

| Channel tidal flat
I
'channel-Shelf '
meander belts marine coastal basin)
». Delta plain Tidal Tidal Channel Coastal

(non-tidal) sand bar channel deeps mouth HI barrier Prodelta | fj Shelf
bar Strandplain
Figure 7.27. forms: an elongate type exemplified by the modern birdfoot delta of the Mis-
Classification and geomorphic
characteristics of basic delta
sissippi River, or a lobate type exemplified by the now-abandoned Holocene
types deltas of the Mississippi River system. Both types have high sediment input
relative to the marine dynamics. Elongate deltas have a higher mud content
and tend to subside rapidly when they become inactive, thereby preserving
their facies. Lobate deltas sink slowly upon abandonment, and much
upper sand
of the sand thatwas prograded in the upper zones is reworked by marine pro-
cesses (Scott and Fisher 1969).
']/ Newmid-
—if—channel
Figure 7.28.
Mid-channel
M„/ Thread of
Second
branching-;
\K .
'
2
' H
shoal
Stages in bifurcation
and creation
of trunk river
of distributary
'
/ .
maximum
shoal
Submarine h ,fr turbulence
channels '
\ i f
Original
channels of a delta.
natural levee (at depth)

levee of ';>
/
f .- branches
mid-channel—
Natural Natural
shoal (island-
levee levee
forming) %?|
(confining
bank) MB
Original branching of a delta channel Later stage of channel subdivision
1
(A) (B)
High-destructive deltas originate where ocean or lake energy is high, and

much of the fluvial sediment is reworked by waves, etc., before its final de-
position. Figure 7.27shows two tvpes of these. In wave-dominated types, such
as those of the Nile and Rhone rivers, sediment is accumulated as arcuate
sand barriers near the mouth of the river. In tide-dominated types, tidal cur-
rents arrange the sediment into sand units that radiate linearly from the river
mouth. Muds and silts accumulate inland from the segmented bars where ex-
tensive tidal flats or mangrove swamps evolve.
Dynamics and Delta Evolution

A sediment-laden river entering a body of standing water behaves much like
a free jet of flow (Bates 1953). The
form of either an axial
jet flow is in the
jet in which mixing is three-dimensional or a plane jet in which two-dimen-

sional mixing prevails. Which flow type develops depends on the relative den-
sities of the two water bodies. If the inflowing water is denser (hyperpycnal
flow) because of its cold temperature or high sediment concentration, a plane-

jet flow occurs in the form of a turbidity current moving along the basin floor.
If the densities are nearly equal {homopycnal flow), axial-jet flow results, and
complete mixing occurs very close to the river mouth. Almost all the river
sediment is deposited immediately after entering the standing water. Homo-
pycnal flow is most common in freshwater lake deltas, and its mechanics leads
to the classic Gilbert-type construction (see Born 1972, for example). In the
third possibility, where rivers flow into more dense ocean water (hypopycnal
flow), mixing is rather slow and the river water spreads out laterally in a plane-
jet flow.
The depositional pattern that develops at any individual river mouth de-
pends on the intensity of spreading and turbulence and how those factors are
modified by tides and waves (Wright 1977). In high-constructive deltas, dis-
tributaries develop at the mouth of the in-flowing river where longitudinal
bars are deposited because bedload cannot be transported when the velocity
suddenly decreases. The initial bar exerts an influence on the flow and, as
figure 7.28 shows, causes the river to bifurcate into two channels immediately
upstream from the bar crest (for details see Russell 1967b). The distributary
channels are lined with natural levees that may begin beneath the surface
Chapter 7 298
Figure 7.29.
Sequence of the development of
the subdeltas that comprise the
Mississippi River deltaic plain
OF Mexico
Sale Cypremort > 4600 yrs. B.P. Lafourche ca. 1000-300 yrs. B. P.
Cocodrie ca. 4600-3500 yrs. B.P. | Plaquemine ca. 750-500 yrs. B.P.
ess? Teche ca. 3500-2800 yrs. B.P. ^ Balize < 550 yrs.
St. Bernard ca 2800-1000 yrs. B.P.
through slow accretion of suspended load as it spreads laterally under plane-

jet flow(Morgan 1970). With continued deposition, combined with minor
channel scouring, the levees and bars emerge (fig. 7.28), and the distributary
channels extend farther into the basin. The process may be repeated fre-
quently, giving the basin a veinlike appearance and prograding the delta front.
As the delta progrades, shorter routes to the ocean become available. These
pathways often begin far inland from the delta front, usually developing when
the river is diverted through a breach in the levee called a crevasse. The new
river course shifts the locus of sedimentation and begins to form a new deltaic
lobe. The Mississippi River deltaic plain, for example, actually represents the
coalescence of seven major lobes that were built at different times and in var-
ious positions during the last 5000 years (fig. 7.29.)
The modern birdfoot-delta growth is only a minor portion of the entire
deltaic area. We have good reason to believe that a new lobe of the Mississippi
delta is being developed at the present time by the Atchafalaya River. This
distributary breaking off the Mississippi channel upstream from Baton
river, a
Rouge, Louisiana, carries about 30 percent of the Mississippi River flow.

During flood events considerable load is transported down the Atchafalaya,
which has progressively filled in shallow lakes in the lower Atchafalaya basin.
Since the early 1950s most of the sediment has been reaching Atchafalaya
Bay (approximately 160 km west of New Orleans), where it is actively building
thenew deltaic lobe (Shlemon 1975; Rouse et al. 1978). Landsat satellite im-
agery shows clearly that the lobe is growing (fig. 7.30), with approximately
299 Fluvial Land forms
Figure 7.30.
Emergence of new land in
Atchafalaya Bay as part of

development in a modern lobe of
the Mississippi delta.
(A) (B)
6.5 km 2 of new land having emerged from the bay each year since the early
1970s (Rouse et al. 1978). Most scientists believe that total diversion of the
Mississippi into the Atchafalaya is inevitable because that route to the ocean
is about 300 km shorter than the present course. This gives the Atchafalaya
a distinct advantage, and capture will occur unless humans use absolutely he-
roic measures to maintain the status quo. When
abandoned during
lobes are
shifting of the river course, they are no longer fed by incoming river debris
and immediately become vulnerable to erosional attack by the ocean. Thus,
as a new lobe develops, older lobes are being destroyed.
The processes that produced the major lobes of the Mississipi deltaic plain
are obscured by modifications since their abandonment. Within the modern
delta, however, the genetic mechanics is known in considerable detail. Here
crevasses in the levee system have repeatedly shifted the site of deposition.
The breaks in the levee begin because of overtopping during a major flood and
gradually increase in size through scouring associated with succeeding floods.
Sediment diverted through a breached levee progressively builds a subdelta
from deposits known as crevasse splays. Four such subdeltas have formed in
historic times (fig. 7.31). As channels in the subdeltas bifurcate and prograde,
their gradients decrease and they lose their ability to transport load. When
their gradients approach that of the main trunk channel, it is no longer ad-
vantageous for the river to utilize the crevasse system. Sedimentation in the
subdeltas ends, and the ocean begins to inundate the area as subsidence and
compaction lower the subdelta surface. At the same time, however, a new sub-
delta may be developing somewhere else within the birdfoot system.
Chapter 7 300
Figure 7.31.
Subdeltas of the modern birdfoot
delta of the Mississippi River.
Itseems clear that delta formation can be viewed on several time scales.
On a short-term basis, only a limited area (subdelta) receivesany sediment,
but the position of the accumulation shifts repeatedly. On a longer time scale,
the entire active delta (lobe) has a periodic migration. What we observe as
the Mississippi River delta is in fact a monstrous area created by the co-
alescence of a number of major lobes over a long period of time. Active de-
position occurs on only one lobe at any given time, and on only a minor portion
of that lobe. Even so, the evolutionary processes on any time scale are probably
similar in that they involve channel bifurcation, levee development, and cre-
vassing.
Even though this evolutionary model is probably correct, it is valid only

for deltas of the high-constructive type. The model
known in such detail is
only because geologists have studied the Mississippi delta for many years. Other
delta types have not been so closely examined, and their mode of origin is not
nearly so well known. Furthermore, we are far from understanding precisely
how changes in climate, tectonics, or any of the other fundamental controls
affect the mechanics of delta evolution.
The character of each major fluvial landform is due to the manner in which Summary
a particular set of geomorphic controls influences the river mechanics. The
features can be predominantly erosional, predominantly depositional, or de-
rive from the combination of both types of processes.
Floodplains originate by lateral migration of meanders and by periodic
overbank flooding. The sediment composing a floodplain sequence is mainly
laterally accreted point bar deposits accumulated as the river shifts its position
across the valley bottom. The point bar material is usually capped by a thin
layer of silt and clay deposited as vertically accreted sediment during over-
bank flooding of the river. The total amount of vertical accretion is probably
controlled by the rate of the river's lateral migration. Because many rivers
move across the valley floor rapidly, most floodplains are developed by lateral
accretion, but overbank deposition may be important in the initial phase of
construction.
Terraces are merely abandoned floodplains. They form when entrench-
ment places the river at a lower level, thereby removing the former floodplain
surface from the river's hydrologic activity. The origin of a terrace usually
refers to how its flat tread (the former floodplain level) was formed. The tread
of an erosional terrace is produced by the lateral migration of a river and is
capped by the thin point bar and overbank deposits associated with floodplains
formed in that manner. In contrast, the surface of a depositional terrace rep-
resents the upper level of the sediment deposited in an episode of valley filling.
The use of terraces to reconstruct geomorphic history demands some knowl-
edge of their origins, because the salient properties of the various terrace types
develop through different sequences of events.
Piedmont regions are characterized by depositional features (alluvial fans)
and plains of erosion (pediments). Both features manifest some accommo-
dation between the amount of sediment derived from a source area and the
ability of the river to transport the sediment across the piedmont zone. The
processes that develop fans and work on their surfaces are so complex that
little agreement exists regarding their origin. Pediments also seem to defy ge-
netic generalization. They are usually interpreted as being the result of lateral
planation, weathering, and rill wash, or some combination of the various pro-
cesses. As with fans, however, little agreement can be found concerning the
origin of pediments;some may be relict features that are completely unrelated
to modern geomorphic controls. It is clear that piedmont landforms cannot be
Chapter 7 302
placed into all-inclusive genetic models. Their properties vary too much to be
explained by one mode of origin. Because of this, every piedmont region must
be examined and interpreted according to local tectonics, climate, geology,
and geomorphic history.
Deltas represent the accumulation of sediment as a transporting river en-
ters a body of standing water. At its mouth the river bifurcates into distrib-
utaries and constructs levees. This allows debris to be transported farther into
the ocean or lake basin and permits the delta to expand by prograding into
the basin area. The form and size of the delta, however, depend on the balance
reached between river flow and the counteracting energy of currents and waves
in the ocean or lake. Detailed studies of deltas reveal a very complex growth
history in which the site of active sedimentation shifts periodically through

crevasses in the natural levees. Active sedimentation occurs on only a small
part of the feature at any given time.
cussed in this chapter and also contain more extensive bibliographies on the
various topics.
Bull, W. B. 1968. Alluvial fans. Jour. Geo!. Educ. 16:101-06.

Cooke, R. U., and Warren, A. 1973. Geomorphology in deserts. London:
Batsford Ltd.
Denny, C. S. 1967. Fans and pediments. Am. Jour. Sci. 265:81-105.
Hadley, R. F. 1967. Pediments and pediment-forming processes. Jour. Geol.
Educ. 15:83-89.
Howard, A. D.; Fairbridge, R. W.; and Quinn, J. H. 1968. Terraces,
fluvial — Introduction. In Encyclopedia of geomorphology, edited by R.

Fairbridge, pp. 1 1 17-23. New York: Reinhold Book Corp.
Leopold, L. B.; Wolman, M. G.; and Miller, J. P. 1964. Fluvial processes in
geomorphology. San Francisco: W. H. Freeman.
Morgan, J. P. 1970. Deltas— A resume. Jour. Geol. Educ. 18:107-17.
Oberlander, T. M. 1974. Landscape inheritance and the pediment problem
in the Mojave Desert of southern California. Am. Jour. Sci.
274:849-75.
Schumm, S. A. 1977. The fluvial system. New York: Wiley-Interscience.
Twidale, C. R. 1976. Analysis of landforms. Sydney. Aus.: John Wiley &
Sons.
Wolman, M. G., and Leopold, L. B. 1957. River floodplains; some
observations on their formation. U.S. Geol. Survey Prof. Paper 282-C.
Wind Processes and Landforms
8
I. Introduction VI. Deposits and Features
II. The Resisting Environment A. Ripples
III. The Driving Force B. Dunes
IV. Wind Erosion C. Fine-Grained Deposits
A. Processes VII. Summary
B. Features Produced by VIII. Suggested Readings
Erosive Action
V. Wind Transportation and
Deposition
303
Chapter 8 304
Introduction Scientists have not agreed on the role of wind in geomorphology. Irrefutable
proof that wind is capable of significant geomorphic work is simply not avail-
able in many instances, because definitive, quantitative data concerning eolian
processes and features are woefully few. It has been described as being ab-
solutely dominant in arid regions to being only a minor perturbation on fea-
tures formed almost entirely by fluvial or slope processes. Disavowing these
extreme views, it is probably safe to say that wind can be an effective geo-
morphic agent under certain physical conditions. Regions having sparse veg-
etation and unconsolidated sediment not tightly bound by rooting systems are
most susceptible to wind attack. Extensive evidence of eolian processes is
therefore found in those areas, such as modern deserts, where vegetation growth
is stunted by lack of water or immature soil development. As is true of any
process, however, the proficiency of wind to do geomorphic work depends on

whether the driving force can exceed resistance of the surficial material. Thus,
assigning wind processes to one particular environment alone is a gross per-
version of the principles governing geomorphic processes. The emphasis placed
here on deserts is simply a pedagogical tool, not a geomorphic necessity.
Detailed reviews of the physical basis of wind action and its geomorphic
results are available (Bagnold 1941; Chepil and Woodruff 1963; Cooke and
Warren 1973), and much of the following has been liberally excerpted from
those excellent treatments.
The Resisting Contrary to popular thought, deserts are not the barren tracts of shifting sands
Environment depicted in Hollywood epics. In fact, only one-fourth to one-third of most desert
surfaces is occupied by sand, which usually occurs in large, sandy plains called
ergs (see I. G. Wilson 1973). Normally, deserts display a variety of erosional
and depositional landforms imprinted on a diverse topography that ranges from

flat plains to rugged mountains (table 8.1 and fig. 8.1). Deserts also exist in
different temperature zones; polar deserts are common, although relatively
unstudied. Specific processes differ in relative importance in polar deserts and
hot deserts, and so the similarity of landforms that sometimes exists between
the two regions is not infallible proof of an identical genetic history. In fact,
the same external form may be a function of a myriad of basic processes com-
bined in slightly different ways. Unless stated otherwise, the discussion fol-
lowing refers only to hot desert conditions.

Deserts are by definition arid.They receive less than 25 cm of annual
precipitation and have enormous evaporation rates, commonly 15 to 20 times
greater than the precipitation (Stone 1968). Although deserts have a meager
plant cover, the diversity of vegetal types is surprising, ranging from shrubs
and grasses to true woodlands. This diversity is created by minor variations
in soil moisture that relate to elevation. It is important because humus content
and soil binding differ with particular species and because some plants are
annual and others are perennial. The type of flora influences the resisting set-
ting; woodlands, for example, are significantly denser than other vegetation
305 Wind Processes and Landforms
Major
desert salt lakes
Figure 8.1.
Map of Australian arid regions
Table 8.1 Area and percentage of total desert region of major components in the showing diverse physiography of
arid zone of Australia. (Compare with map in fig. 8.1 .) deserts
Percentage
Km 3
of Arid Zone
Mountain and piedmont deserts 930.000 175

Riverine desert 210,000 4.0
Stony desert 640,000 120
Desert clay plains 690,000 130
Sand desert ,680,000 31.0
Shield desert ,200,000 22.5
Total 5,350,000
From Mabbult 1971 Used with permission ol the Australian National University Press
Chapter 8 306
types and protect more surface area (Whitaker et al. 1968). In general, moun-
tainous zones with stabilizing vegetation and thicker weathering mantles are
less susceptible to wind attack than is the lower level of the desert environ-
ment.
Weathering and soil-forming processes are intimately involved in the
character of a land surface and how it resists or succumbs to wind erosion.
The angular, unaltered debris so prevalent on desert surfaces is presumed to
be the residuum of rock disintegration, leading to a general concurrence that
mechanical weathering predominates in the desert environment. This does not
mean that mechanical weathering functions to the total exclusion of decom-
position. Chemical processes, in fact, do function in the desert, as evidenced
by the ubiquitous desert varnish and the common red coloration of desert soils,
both of which probably require some degree of ion mobility. In addition, as

explained in chapter 4, many processes of disintegration, such as exfoliation
and grus formation, are abetted by chemical reactions that produce volume
expansion.
In general, however, even though decomposition does function, maturely
developed soils are rare in desert environments unless relict profiles formed
during more humid intervals of the Quaternary are preserved.
Desert soils (even in their immature state) are important determinants of
A horizons are quite thin, mechanically
the facility of wind action. In general,
weakened, and more permeable than the lower horizons (Cooke and Warren
1973); case hardening at the surface may in places counteract this tendency
(Lattman 1973). Nevertheless, regardless of how slowly it proceeds, illuvia-
tionmay cement B horizons to the extent that they become nonpermeable,
highly indurated zones (Gile et al. 1965, 1966; Gile 1966). The long-term
result of desert processes is to create a surface that is susceptible to wind ero-
sion and a subsurface that is eminently resistant. Stripping of the surface ho-
rizon will expose the lower, resistant zones, which are preserved as stable
platforms for extended periods of time. It should be noted that stripping at
one place is commonly balanced by deposition of eolian debris somewhere else
in the area.
Surfaces armored with a thin layer of stones that protects an underlying

horizon of sand, silt, or clay are common in many environments. They are
particularly evident, however, in hot desertswhere they are given a variety of
names such as stone mantles, gibber, hammada, reg, and desert pavement.
Desert pavement, as mentioned in chapter 7, is associated with alluvial fans
and other deposits of unsorted alluvium. The surface armor is usually only one
or two stones thick and consists of whole clasts or disintegrated parts of the
coarse fraction found in the underlying alluvium. In most cases, a thin, fine-
grained layer, which is probably a product of desert weathering (fig. 8.2). ex-
ists beneath the pavement. The pebbles at the surface are most likely a lag
concentrate formed as wind action progressively removes the fine-grained al-

luvium.
In contrast to the desert pavements, some parts of the desert are covered
with abundant fine-grained debris, and there deflation is a viable erosive pro-
cess. Playas, for example, are notable for theirand clay mantles; it has
silt
often been suggested that they are subject to extensive degradation by defla-
tion (Blackwelder 1931). There is no doubt that deflation of fine sediment
occurs (fig. 8.3), but it may not be as pervasive in the desert environment as
once supposed (Hooke 1968). Although most deserts do contain fine material,
it is usually fabricated into a cohesive crust by a number of processes (Dury

1966a; Cooke and Warren 1973). Crustation not only gathers individual par-
ticles into aggregates larger than the wind can entrain, but it also protects
Figure 8.2.
Percentage of soil composed of
particles less than 2 mm in
diameter Analyses of soils in the

Lahontan Basin, Nev.
(Reproduced from Soil Science
Society of America Proceedings,
Col. 22, 1958, p. 65, by permission
of the Soil Science Society of
America)
Figure 8.3.
20 40 60 80 100 20 40 60 80 100
sediment
Deflation of fine-grained
% whole soil < 2mm diameter % whole soil < 2mm diameter atPyramid Lake, Nev. Bottom
sediment has been exposed by
recent decline in the lake level.
Chapter 8 308
unconsolidated fines immediately beneath the surface. The greatest resistance

occurs when and clay are mixed together in proportions of clay 20
sand, silt,
to 30 percent, silt 50 percent, and sand 20 to 40 percent (Chepil and

40 to
Woodruff 1963), but establishment of the most resistant size blend is not the
only determinant of resistance. Variables of moisture, humus, and chemical
composition may give considerable resistance to size blends that are usually
susceptible to deflation. Calcium carbonate, on the other hand, may actually
induce erodibility because it tends to inhibit the formation of aggregate clods
(Cooke and Warren 1973, p. 241).
In summary, the geomorphic effect of wind action is regulated to a sig-
nificant degree by the properties of the resisting framework. For wind pro-
cesses to have any geomorphic consequence, the surface must be unvegetated
and littered with noncohesive sediment smaller than gravel-size. Any region
producing sediment of that type, and lacking the climatic, topographic, and
geomorphic conditions needed to protect it, may be open to pronounced wind
attack. Because hot deserts are most susceptible to wind action, they are taken
as the model for the discussion of eolian processes. However, coastal regions,
unvegetated semiarid or subpolar zones, and areas in front of active glaciers
also often are modified bywind processes. In any situation, the amount of
geomorphic work actually accomplished by the wind depends on the second
—
prime variable the character of the wind itself.
The Driving Force Global wind systems are related to the large pressure differentials associated
with worldwide circulation patterns. In contrast, local winds are generated by
anomalies in the physical characteristics within a specific area. For example,
in many cases winds reflect a difference in thermal properties of surface ma-
terials,and so we can expect the greatest wind activity in those environments
where large temperature variations exist in the air layer immediately above
ground level. Such conditions are common in deserts, along seacoasts, or in
areas with pronounced diversity in elevation, such as the juncture of mountain
ranges and the low plains fringing them.
Certain attributes of the wind —
mainly its direction, velocity, and degree
of turbulence —
are responsible for most geomorphic effects. In areas with large
thermal contrasts, wind direction is more or less predetermined by the tem-
perature gradient of the near-surface air. However, temperature gradients often
change from daytime to nighttime because of variations in rates of cooling
and heating. In deserts, heating of the bare desert surfaces incites winds to
move toward them from surrounding areas that have a vegetal cover. At night,
however, the desert surface cools more rapidly than the other regions, and the
winds reverse their direction. A similar diurnal direction change occurs along
coasts because the ocean warms less slowly than the adjacent land, prompting
on-shore winds during the day. The ocean, however, gives up stored heat less
readily at night, and winds blow off-shore. The importance of wind direction
is most evident in the development and preservation of eolian bed forms con-
structed from loose, transportable sand.

100
Figure 8.4.
E t f° The relationship of wind velocity,
30 height above surface (z) and
/V. = 26 shear or drag velocity (I/-).
10
I
cj
/
fk
I
3
l
z B
1
i
V
Sv- =
i
l
61
0.3 \
j
0.1
f
X 0.03
0.01 of
i
0.003 K
I
n nm I
100 200 300 400 500 600 700 800

Wind velocity (cm/sec)
Wind velocity is important because it is the prime determinant of what

material will move under wind attack and what will remain stationary. Wind
velocity increases with height above the ground because it is slowed at the
surface by friction. The change in velocity with height can be expressed in a
number of ways exemplified by the equation (Bagnold 1941)
V. = 5.75 V. log z/ fo
where V. is the velocity at any given height z, V, is a parameter called drag

velocity, and A: is a constant relating to surface roughness. V,, originally des-
ignated as a drag phenomenon, is actually a shear velocity and relates to shear
stress. The parameter k is the height of a thin zone immediately above the
surface in which velocity is zero. The thickness of this zone depends on the
surface roughness, but on a granular bed
it tends to be about 1/30 of the
flat,
diameter of the surface grains. The general relationship between height and
velocity plots as a straight line on semilogarithmic paper (fig. 8.4).
As the wind blows harder, V. increases and exerts a greater stress on par-
ticles exposed at the surface. This follows because the magnitude of shear
(drag) across any unit area of surface is related to V. such that
where r is drag/unit area and p is the density of the air. Thus, for a surface
with constant particle size, a stronger wind will increase the value of V., and
therefore the height-velocity relationship is defined by a number of straight
Chapter 8 310
lines, each having a different V, value (fig. 8.4). Drag velocity in this sense
represents the rate at which velocity increases with height and can be thought
of as a velocity gradient. The value of V. can be determined as the tangent of
the straight line, as shown on figure 8.4, divided by the constant of propor-
tionality, 5.75. For example, the tangent of line OD on figure 8.4

AC
is -pr= or
C
150
:
log 10
— . Therefore,
V* = ^ I 5 75
-
= 26 cm /sec.
Interestingly, all lines of different V* values merge at the ordinate where ve-
locity is zero. This occurs simply because A: is a function of surface conditions,
such as particle size, rather than wind velocity.
Once the value of V, is known, it is possible to calculate the wind velocity
at any height above the surface. For example, if V. = 26 cm/sec, the velocity
at 4 cm above the surface is
V = z
5.75 V. log z/k
= 5.75(26) X log z - log k

= 150 [0.6 - (
- 1.82)]
= 363 cm/sec.
Assisting velocity in wind attack is the factor of turbulence, which occurs

in the form of eddy currents. Turbulence has a direct influence on the en-
trainment process, although it is probably capable of lifting only particles
smaller than sand (Bagnold 1941; Chepil and Woodruff 1963). Sharp (1964),
however, suggests that grains up to 0.125 mm may be affected by wind tur-
bulence. Most likely, turbulence helps in the molding of desert landforms, but
the types of eddy motion are so numerous and mechanically complex (Lumley
and Panofski 1964), that precise measurement of turbulence is precluded, and
its relationship to specific bed forms in not clear. At present little data are
available to generalize about velocities and turbulence in wind processes but.

in ways that are still not quantitatively clear, both factors must bear directly
on the origin of wind-produced landforms.
Wind Erosion Processes

As wind blows composed of loose sediment, a critical drag
across a surface
which particle motion begins. The ease of entrainment, how-
velocity exists at
ever, depends not only on the size of the particles and the wind velocity but
also on complicating factors such as soil moisture, packing, etc. Recognizing
the inherent complexity, the velocity that initiates movement is called the fluid
threshold (Bagnold 1941) and can be estimated by the equation
r
; = A 7" gD
Figure 8.5.
60
Relation of particle size to
1 threshold velocity.
jS
1
50
t
\
I Fluid threshold,/^
40 \
\
\
Impact threshold
\
C 30 \
\
\
\
20 \
zy
10
/
r i i IHHHHfl
.4 .6 .8
Grain diameters (mm)

1 1.5 2.5
where V., is the threshold value of drag velocity, p„ is the density of the air,
p 5 the density of the sediment, D is the particle diameter, g is gravity, and A

is a constant of proportionality which for air is 0.1. Most desert sands have a
threshold velocity of about 16 km/hr, but the precise value for any size varies
with other factors such as particle shape (G. Williams 1964), sorting (Wood-
ruff and Siddoway 1965), and surface roughness. Roughness is a function of
particle size but is also controlled by vegetation. Rough surfaces reduce wind
velocity; as high, dense vegetation increases roughness drastically, it lessens

the erosive effect of the wind.
The influence of moisture on the critical threshold velocity was docu-
mented by Belly (1964) who suggested that a surface moistened by liquid or
vapor would require a significantly higher velocity for sediment entrainment.
For example, Calkin and Rutford (1974) found that when Antarctic dunes
were moistened by summer melting or condensation, the threshold velocity
needed to entrain their sands doubled.
The fluid threshold equations clearly indicate that entrainment velocity
varies directly with particle diameter; nevertheless, as figure 8.5 shows, the
relationship becomes invalid when size is less than 0.1 mm because of the
greater interparticle cohesion and low roughness associated with finer parti-
cles (Bagnold 1941; Smalley 1970). Furthermore, grains greater than .84 mm
are moved with great difficulty, and that size may represent a logical upper
limit for unaided wind entrainment. Once motion begins, however, the surface
is subjected to a continuous rain of moving particles that, on impact with sta-
tionary grains, produce entrainment at a velocity lower than the fluid threshold.
This reduced threshold velocity, called the impact threshold (fig. 8.5), be-
comes progressively more and particles greater
significant as size increases,
than .84 mm may be entrained even when the velocity is well below the fluid
Chapter 8 312
Figure 8.6.
Diagram of the saltation process.
Moving gram strikes the surface
and dislodges a particle to
elevation h The particle moves
downwind and repeats the
process C represents the
distance of travel of the dislodged
particle before striking the
surface
threshold. Indeed, an exceptionally strong 1977 windstorm in southern Cali-

fornia carried particles as large as 7 mm at a height 240 cm above the surface
(Sakamoto-Arnold 1981). It appears, therefore, that when grains begin to move
at the fluid threshold, the entrainment process is self-generating because the
wind speed is already above the impact threshold. Each moving grain is not
only a product of the system but also an integral component of the entrain-
ment mechanics through the process of saltation as depicted in figure 8.6.
Features Produced by Erosive Action

Geomorphic features manufactured by wind action are due to abrasion and
deflation, the two main erosive processes. Abrasion results when sand particles
carried by the wind act as grinding tools to physically wear away exposed
surfaces of solid rocks or rock fragments. It is most effective where relatively
weak rocks, bare of vegetation, are blasted by high-velocity winds that carry
an abundant amount of hard particles (Suzuki and Takahashi 1981). In fact,
Sharp (1980) has demonstrated that the abrasion rate increases significantly
when more load is made available for wind transport. Thus, maximum abra-
sion is determined by some unique combination of velocity and particle con-
centration. Normally this will occur within 2 m of the ground surface.
Perhaps the most frequently cited evidence of wind abrasion is the de-
velopment of faceted stones called ventifacts (fig. 8.7). Ventifacts are pro-
duced on pebble surfaces that are oriented perpendicular to the prevailing
wind. The bombardment of wind-driven sand gradually cuts a smooth
incessant
face into the windward portion of the original surface and forms a sharp edge
between it and the leeward side of the pebble. The slope on the faceted surface
is normally between 30° and 60°, but enough variability exists in these angles
to suggest that some evolutionary sequence is involved in their development.
It is equally feasible, however, that various heights of pebbles above the sur-
face relative to various levels of maximum abrasion potential in the wind can
produce a number of facet inclinations and shapes (Sharp 1964). It is also
possible that some faceting is accomplished by dust-sized particles (silts and
clays) carried in suspension, which would make ventifact sculpturing a func-
tion of the unique aerodynamics associated with each individual situation
(Whitney and Dietrich 1973; Whitney 1978).
Facets do not form on surfaces that parallel the wind direction, and there-
fore the ubiquitous presence of ventifacts with more than one face and edge
(fig. 8.7) has created some interpretive furor. It is tempting to regard multi-
faceted ventifacts as evidence of a shift in the prevailing wind direction, but
Figure 8.7.
Pebbles and cobbles pitted and
it is equally possible that the stones themselves are occasionally turned over
faceted by wind-blown sand.
or rotated (Sharp 1964, 1980). Overturning, of course, would result in a new Ventifacts were all found in
portion of the original pebble being subjected to wind erosion, and any number Sweetwater County, Wyo.
of facets could form in a unidirectional wind.
Abrasion affects rock outcrops by shaping or grooving the surface. The
most common feature etched by the wind, a yardang, is an elongate ridge,
normally less than 10 m high but occasionally as high as 30 m, which is aligned
parallel to the prevailing wind. Yardangs are most prominently developed in
regions underlain by relatively soft rocks and tend to occur in groups in which
individual ridges are separated by round-bottom, wind-eroded troughs (Haynes
Chapter 8 314
1982). Blackwelder (1934) emphasized the importance of eroding the inter-

vening "yardang trough" rather than the ridge He believed that most
itself.
erosion occurs at the lowest level, decreasing inmagnitude and efficiency to-
ward the ridge crest. As a result the troughs are eroded more rapidly, so that
sometimes the ridge sides are slightly undercut and the ridge crests are left
rather ragged.
Grooving of bedrock surfaces is also a common eolian phenomenon. The
abrasive processes that result in parallel furrows seem to operate at different
orders of magnitude, ranging from tiny elongate flutes on ventifact surfaces
measured in centimeters (Sharp 1964), to giant hollows more than a kilometer
wide and up to 10 m deep. Some question lingers as to whether the mega-
grooves are demonstrably erosional or merely a depressed zone between par-
allel ridges of deposited sand. Although the possibility of misinterpretation
remains, some large erosional grooves have been reasonably well documented
(D. King 1956; Shawe 1963; Stokes 1964). In some cases, grooving may have
a distinct influence on regional drainage patterns (Stokes 1964; Beaty 1975b;
Hallberg 1979), but care must be exercised in making such an interpretation
(Rahn 1976).
The process of deflation has been suggested as the prime mechanism in
the genesis of many enclosed desert basins. These basins range in size from
small deflation hollows to vast expanses measured in hundreds of square ki-
lometers. Usually the small depressions are easily related to wind erosion since
they are often strung out parallel to the wind and may be associated with
dunes (Flint and Bond 1968; Sharp, 1979). A deflation origin for the larger
areas, called pans, is much more difficult to prove since they have no clear
orientation. Arguments about their origin, therefore, are often negative in the
sense that they merely eliminate other processes and leave deflation as a vi-
able, but untested, alternative. Very large pans, however, must be relict to
some extent since wind erosion, even if proven, functions so slowly that an
enormous time span would be necessary to accomplish their formation.
Wind Transportation The entrainment of sediment releases particles to the processes of transpor-
and Deposition tation and deposition, and the result is the array of surface forms so diagnostic
in areas under wind attack. The type of transporting process is determined
fundamentally by the size of the debris made available to the eolian force;
silts and clays are carried in suspension, fine and medium sands travel by sal-
tation. Coarse sand, and sometimes gravel (Sakamoto-Arnold 1981), move by

rolling or sliding motions collectively referred to as surface creep. The most
striking eolian features are made with sand because, in the absence of unusu-
ally strong storms,few winds can move larger grains, and sediment entrained
by turbulence and carried in suspension often is so diffuse as to preclude its
accumulation in explicit surface forms. In addition, entrainment of fine-grained
sediment is not so easy as one might expect, because the surface is usually
smooth and the small grains tend to retain moisture and have stronger inter-
particle chemical bonds. Grains larger than 0.1 mm probably cannot be trans-
ported in suspension (Bagnold 1941;Kuenen 1960; Sharp 1963), and particles
315 Wind Processes and Land forms
Figure 8.8.
Change in drag velocity caused
by saltation Dashed lines show

drag velocity before saltation
begins. Solid lines represent drag
velocity after saltation starts.
.003
.002
Wind Velocity (m/sec.)
both larger and smaller than sand often resist entrainment. For those reasons,
our discussion will center on depositional features composed of sand, and on
the mechanics of saltation.
Figure 8.8 shows that when entrainment occurs the moving sand directly
influences the characteristics of the wind itself because the saltating grains
exert extra drag on the system. As discussed earlier, the drag velocity (K.) is
a measure of the velocity gradient of the wind, i.e., the rate at which velocity
increases with height. The stronger the wind blows, the greater the drag on
the surface and the greater the divergence of the drag velocity curves away
from the vertical. The slope of each line on the figure, therefore, reflects its
rate of increase in velocity with height (given as drag velocity, K.); the higher
V. values indicate a more rapid increase in velocity with ascension above the
surface.
The solid lines on figure 8.8 represent the changes impressed on the system
after saltation begins. Note that under saltating conditions k rises dramati-
cally, and it no longer represents a zero velocity but assumes a constant ve-
locity value at the height k' where all the gradient lines merge. The wind
velocity at any height (z) is then defined by the equation
V. = 5.75 V.' log z/k' + V„

Chapter 8 316
where V.' is the drag velocity under conditions of saltation, k' is the new height
of the focus, and V, is the threshold velocity (impact threshold) measured at
the height k'.
The effect of saltation is obvious from the graph. For example, a wind
with a velocity gradient V, = 62 blowing across a completely resistant surface
has a velocity of about 970 cm/sec at a height of 2 cm. Once motion begins,
however, the velocity at that level, read along the gradient line V.' = 62, has
decreased to approximately 530 cm/sec. It is important to stress again that
the value of k or k' is dependent on the roughness components of the surface,
and so figure 8.8 defines curves that are valid for only one particle size. In
addition, very few measurements have been made in the natural setting, and
although estimates of the quantitative relationship between k values and par-
ticle diameter have been made (Cooke and Warren 973, p. 259), enough vari- 1
ance exists in these for us to be cautious about their use.

Saltating grains travel in a characteristic path in which particles dis-
lodged or bounced from a surface move at first in a predominantly vertical
direction (fig. 8.6). The subsequent travel path is probably controlled by ( 1 ) the
particle's initial upward velocity and (2) the velocity gradient above the sur-
face.Bagnold (1941) suggested that most of the forward momentum is gained
from the wind when a grain is near the top of its path, and from then on the
actual trace of the particle movement depends on the magnitude of the wind
thrust and the settling velocity of the grain. Precisely how high a grain will
risefrom the surface depends, of course, on the size of the particle involved
and on what kind of a surface it is bouncing from. It is clear, however, that
the dense curtain of saltating grains is restricted to a thin near-surface zone.
Most observations (Bagnold 1941; Chepil 1945; Sharp 1964) indicate that an
overwhelming percentage of the total load is carried within 2 m of the surface,
although minor amounts may rise above 10 m.
Even though most sand moves by saltation, some of the load never enters
the realm of airstream mechanics, but moves as surface creep. Creep results
when the impact of saltating grains spasmodically shoves or rolls surface par-
ticles forward without displacing them upward. This type of motion becomes
progressively more important as the grain size increases (Sharp 1964) and is
probably the dominant mechanism of forward motion in very coarse sand and
gravel. Nonetheless, in average wind-blown sand, ranging from 0.15 to
0.25 mm, creep seems to account for 7 to 25 percent of the total load.
Theoretically, it should be possible to estimate the total wind-traction load,
because the discharge of sand must be related to the drag exerted on the wind
by the saltating grains, and drag is dependent on V. (drag velocity) and air
density. Indeed, a number of equations have been derived to make these es-
timates. It is doubtful, however, that any equation will be very precise, for
other uncontrolled factors such as particle shape (G. Williams 1964) and the
ambiguities of grain motions in flight must introduce a finite uncertaint} to
any formula. This is perhaps best demonstrated by figure 8.9 where, for ex-
ample, a drag velocity of 50 cm/sec may produce a significant difference in
Q depending on which equation is employed.
200
Figure 8.9.
Relation of sand movement to
drag velocity Zone between
100 curves A and B contains a family
of curves derived from numerous
studies
E 50
20
0.03 0.05 0.1 0.5
Q (g/cm sec)
The most striking features associated with eolian processes occur in vast,
sandy Deposits and Features
deserts called sand seas (or erg in theSahara Desert of North Africa). These
areas have an enormous supply of sand and are marked by varied assemblages
of sand dunes that give a wavelike configuration to the surface. Complex dune
fields, however, are not the only component of sandy deserts since large por-
tions are often occupied by tabular bodies of sand, called sand sheets, which
have little surface topography, or by relatively flat interdune areas of various
types (for discussion see Ahlbrandt and Fryberger 1982). The largest sand
sea in the Western Hemisphere is the Nebraska Sand Hills, an area of eolian
features that was formed in the Holocene and covers almost 57,000 km 2 of
northwest Nebraska (Ahlbrandt and Freyberger 1980).
On the basis of wind data and Landsat imagery, it seems certain that sand
seas form where there is a significant reduction of wind energy along the di-
rection of sand drift (Fryberger and Ahlbrandt 1979). Zones of reduced en-
ergy leading to the development of sand seas are produced in several ways:
(1) Topographic barriers, such as high plateaus or mountains oriented across
the direction of sand drift, tend to reduce energy by dispersing prevailing winds
into variable flow directions. They also physically block the advance of drifting
sand. For example, dunes piled up in the Great Sand Dunes National Mon-
ument, Colorado, were probably positioned there by the blocking action of the
Sangre de Cristo Range. (2) Large bodies of water, such as lakes or oceans,
also serve as interceptors of drifting sand. In these cases, eolian deposition may
be prograded into the standing water. (3) Wind energy may decrease parallel
to the direction of sand drift because of zonation in the regional climate. In
North Africa, for example, alternating zones of climatically controlled low

and high wind energy are common. Sand that drifts out of regions of high
energy accumulates in the lower energy zones.
Within the sand seas, surface features commonly referred to as bed forms
are spaced with pronounced regularity and, like erosional features, come in a
variety of sizes ranging from tiny ripples to giant forms called draa, which
apparently are restricted to the North African deserts. Intermediate in scale
between ripples and draa are dunes, the most common depositional form in
Chapters 318
Table 8.2 Eolian bed forms and their geometry and possible origin
Wave- Possible Suggested

length Height Orientation Origin Name
300- 20- Longitudinal Primary aerodynamic
5500 m 450 m or transverse instability Draas
3- 0.1- Longitudinal or Primary aerodynamic
600 m 100 m transverse instability Dunes
15- 0.2- Longitudinal Primary aerodynamic Aerodynamic
250 cm 5 cm or transverse hnstability ripples
0.5- 0.05- Transverse Impact Impact
2000 cm 100 cm mechanism ripples
1- 005- Longitudinal Secondary Secondary
3000 cm 100 cm vortices ripple
sinuosity
From I G Wilson 1972 Used with permission of Sedtmentology. Blackwell Scientific Publications
Figure 8.10.
Relation of coarse-grained portion
of sediment load (P20 = the
coarse twenty-percentile) and
wave-length of bedforms (A =
ripples. B = dunes. C = draas)
Note that there are no transitional
forms between the groups
regions susceptible to wind attack. No clear dimensional boundaries separate

the geometric types, which are described in table 8.2, and it is not unusual for
the smaller features to be superimposed on parts of the larger ones. I. G. Wilson
(1972), however, demonstrated that the dominant members of the bed form
hierarchy can be differentiated by plotting their wavelength distance against
a parameter of their constituent particle size (fig. 8.10). It seems clear from
Wilson's work that for any given feature (ripple, dune, draa) a larger particle
size is associated with a greater wavelength, and because grain size is a func-
tion of velocity, the expanded wavelength is also a reflection of a stronger
formative wind. In addition, at any given particle size (and presumably wind
velocity), all the types may be forming simultaneously. This strongly suggests
that different process mechanics are associated with each geomorphic form.
The fact that no transitional forms exist between the major varieties strengthens
this supposition (Wilson 1972).
Ripples
Wind ripples range in amplitude from .01 cm to 100 cm and may be spaced
up to 20 m apart. Their dimensions depend primarily on the wind velocity, the
particle size (Sharp 1963), and the type of ripple. Some ripples are formed
purely by the shear stress of the wind acting on the surface (aerodynamic
ripples), but the overwhelming majority of ripples result from the surface
bombardment of saltating grains and its associated creep.
Bagnold (1941) suggested that the process of forming impact ripples is

which saltating grains strike the surface. Since
closely allied to the angle at
saltating grains descend in a nearly uniform manner, most variations in the
angle of incidence are caused by minor topographic irregularities of the sur-
face (fig. 8.1 1). The intensity of creep in any surface area is proportional to (A)
the number of impacts produced by saltating grains. Thus, in the natural sur-
face hollow depicted in figure 8.11, the number of grain collisions is much
greater on the windward face of the depression (BQ than the lee side (AB).
This results in the transport of more grains up the slope BC than are replen-
ished by movement down the slope AB into the depression. Not only is the
(B)
hollow maintained, but grains also begin to accumulate at point C because
they are delivered there faster than they can be removed on the adjacent level
Figure 8.11.
surface. Eventually the accumulation at C produces a second lee slope CD Process of creep in the
which reinitiates the mechanics that operated in the original hollow. Repeti- development of ripples in a
natural surface hollow (A) Minor
tion of this sequence inexorably propagates the ripple form downwind. Ripple
topographic irregularity increases
crests, however, are not always perpendicular to the direction of the wind cur- the incidence of saltating grains
rent (Howard 1977). Because of this distinct formative process, Bagnold (1941) on the windward face BC
feels that surfaces covered by unprotected sand and having no bed forms are (B)Creep builds ripple crest at C
and a second lee slope CD.
probably unstable.
The process just described should create an ever-rising ripple crest and a
gradually deepening hollow. Actually, the height of the ripple is effectively
limited because it eventually rises to a level where greater wind velocity pre-
cludes deposition. At that height, grains are transported directly over the crest
and deposited in the next downwind hollow, where the wind velocity is much
lower. In this way the ripples assume a consistent geometry that, for well-
sorted sands between 0.19 and 0.27 mm, has a height/wavelength ratio that
is normally only 1:70 and never lower than 1:30 (Bagnold 1941).
In nature, however, coarse grains are often concentrated at the ripple crest
(Bagnold 1941; Sharp 1963), allowing the ripple to grow higher than expected
because the large grains continue to be deposited while the smaller particles
flow over the crest into the hollow. This commonly results in a height/wave-
length ratio as low as 1:10, and an asymmetric ripple shape in which windward
slopes are notably more gentle than lee slopes. The asymmetry, however, may
be lost if a high percentage of the ripple is composed of very large grains.
These "granule ripples" (Sharp 1963) have very large wavelengths and may
be burdened by a high proportion of grains larger than sand size.
Dunes
Of all desert and wind phenomena, sand dunes have received the greatest sci-
entific attention. Dunes attain a characteristic equilibrium profile that can be
logically divided into three components, as in figure 8.12: the backslope or
windward surface, the crest, and the slip face or lee slope. Measurements show
that the backslope declivity, normally between 10° and 15° (McKee 1966;
Sharp 1966; Inman et al. 1966), is in stark contrast to the slip face, which
always stands near the angle of repose for sand, between 30° and 34°. The
crest, separating zones of erosion and deposition on the dune, is usually convex-
up, but on very large dunes the pronounced convexity may be lost (Hastenrath
1967).
Chapter 8 320
Figure 8.12.
Cross-profile of normal dune
showing common geomorphic
components BaC
10°-15°
Dune height increases until it stabilizes along with the equilibrium form.
Most dunes range in height from less than 3 m to 100 m, but in rare cases
they have been observed to be as high as 500 m (Walker 1982). The equilib-
rium height no doubt depends partly on the upward velocity gradient in the
but most workers now seem to believe that height is mainly controlled by
air,
some poorly understood wave motion within the wind. This belief is reinforced
by the fact that dunes normally occur in groups with distinctly regular spacing
rather than as randomly placed individuals. Within any given dune field the
wavelength is quite consistent, although from region to region it can show
considerable variation (table 8.2).
The pronounced regularity in dune spacing hints at some prevailing at-
mospheric motion that is capable of maintaining dune forms and their spacing
in an equilibrium state. For example, formation of an initial dune possibly
interferes with the airflow in such a way that it creates eddy currents or, al-
ternatively, fixes a regularly spaced pattern of turbulence downwind from the

intruding dune (Cooper 1958). Eddy currents unquestionably exist in the lee
of dunes (Hoyt 1966), but considerable question remains as to whether they
are modified by other flow mechanics and, more important, whether they pos-
sess the erosivepower needed to regulate the spacing or shape of dunes (Hoyt
1966; Inman et al. 1966; Sharp 1966; Glennie 1970).
Bodies of sand, which serve as the birthplace of dunes, form wherever
local conditions favor deposition of sand that is moving under wind transport.
This will occur when V. (and therefore r) is lowered, and thus deposition is
commonly prompted by topographic irregularities of the surface. Once de-
posited, however, a sandy patch on an otherwise resistant surface becomes an
integral part of the system mechanics.
Equilibrium in sand dunes represents a balance between the volume of
erosion and the volume of deposition occurring on the feature (Howard et al.
1977). This balance is maintained through adjustments of the backslope and

slip face angles, which in turn are controlled by particle size and velocity gra-
dients. Theoretically, the crest zone should exist as a sharp angle formed by
the equilibrium surfaces of the backslope and the slip face. Actually, the wind
does not respond immediately to a change in slope, and so the locale of max-
imum deposition will be some finite distance downwind from the crest, cre-
ating the tendency toward a convex summit. This tendency is sometimes
complicated by slumping that occurs along the slip face when deposition in
the crestal zone increases the lee slope beyond the angle of repose.
It is axiomatic that maintenance of the equilibrium shape requires for-
ward movement of the from the rear slope must
entire feature, because erosion
be volumetrically balanced by deposition on the lee slope. There seems to be
general agreement that dunes retain their original form as they advance and
Table 8.3 Term nology for basic types of dune forms.
Name Used in Ground

Form Number of Study of Form Slip Face,
Slip Faces and Internal Structure
Circular or
mound
None 3 Dome
elliptical
Crescent in
1 Barchan
plan view
Row of connected
crescents in 1 Barchanoid ridge
plan view
Asymmetrical
1 Transverse ridge
ridge
Circular rim
1 or more Blowout
of depression
U shape in
1 or more Parabolic
plan view
Symmetrical
2 Linear (seif)
ridge
Asymmetrical
2 Reversing
ridge
Central peak
with 3 or 3 or more Star
more arms
Modified from McKee 1979. table 1. p 10

internal structures may show embryo barchan type with one slip face
°Dunes controlled by vegetation
that the shape components, especially the height, influence the rate of forward
migration (Finkel 1959; Long and Sharp 1964; McKee 1966; Inman et al.
1966; Norris 1966; Hastenrath 1967). In fact, it has been repeatedly dem-
onstrated that the horizontal displacement (c) that occurs in any time interval
can be defined by the equation
e = Q/yH,
where H
is the height of the dune and y is the specific weight of the sand. The
absolute rates of forward migration are variable because they depend on local
controlling factors, but movement measured in tens of meters per year seems
to be common. This equation shows a reasonable correlation with some actual
measurements (Finkel 1 959), but the value of Q is subject to so many external
constraints (Chepil 1959; Svasek and Terwindt 1974) that to expect precise
predictions of advance rates may be too much to ask.
The cross-sectional characteristics of dunes tend to be somewhat change-

able because the complexity of wind motions and the variability of surface
conditions interfere with their ideal development. An even more complex matter
is the attempt by students of desert landforms to categorize dunes according
to their plan view, a property known as the dune pattern. Pattern classifications
are almost as diverse and bewildering as the forms they claim to classify. Like
other features, dunes have been grouped on the basis of shape, genesis, wind
types involved, surface conditions, and the like, but each attempt somehow
fails to account adequately for nature's incredible diversity. For our purposes,
it seems best to employ a modified form of the classification devised by the

U.S. Geological Survey (table 8.3; fig. 8.13).
E
o
Q
- —
E ^
; r
«-
o
__ o
: -
° c
- r
_t o —
n
w o n
J.
«—
2§ O TJ
*~
.is
CO u. o £ 5 a
Hack (1941) suggested that in the region he studied, three basic dune
forms which he called transverse, parabolic and longitudinal. Transverse
exist,
dunes are usually free of vegetation; they may be the most probable dune pat-
tern if winds are unidirectional over a limited supply of sand and the sand is
free to migrate. Actually, the mechanics involved in maintaining this dune
form may be controlled by the wind velocity (and its associated turbulence)
needed to transport the largest particles in the sand (Lancaster 1982). The
normal transverse dune is a crescent-shaped feature, called a barchan dune
(fig. 8.14), in which tapering edges or horns of the crescent point downwind.
Transverse dunes may also stand as simple ridges oriented perpendicular to

the wind, or they may form a sinuous asymmetrical ridge, called a barchanoid
ridge, which is composed of connected crescents (fig. 8.13).
Parabolic dunes differ from transverse dunes only in that they are U or V
shaped, and their horns or tapered extremities point upwind. When the lateral
edges of a transverse ridge become anchored by vegetation, the wind removes
sand from the central zone and deposits it on the leeward slope. This, of course,
allows the middle segment to advance relative to the edges and develops the
characteristic parabolic shape. In addition, the dune surrounds a scoop-shaped
hollow, called a blowout, from which the sand in the dune was derived, indi-
cating that the pattern is partly erosional in origin.
Longitudinal forms are called linear dunes. They exist as narrow ridges
that extend parallel to the forming wind. They are usually wider and steeper
at the upwind end, gradually tapering downwind until they merge with the
desert surface. In the Navajo country, linear dunes are separated from one
another by sand-free flats up to 100 m wide. Both the flats and the dune flanks
may be vegetated, leaving only the sand of the ridge tops bare of vegetal cover
and susceptible to wind transport. Locally the dunes may form in the wind-
shadow behind an obstruction, or they may spring forth as wind-rift dunes
(Melton 1940; AGI Glossary 1972) where blowouts provide sand that is ex-
tended downwind as a linear ridge. A special variety of linear dune is called
a seif dune (Bagnold [1941] considered it to be one of only two true dune
forms, the other being the barchan). Seifs are elongate, sharp-crested ridges
that often consist of a succession of oppositely oriented curved slip faces that
impress a sinuous or chainlike appearance on the dune crest. In many cases
seifs attain spectacular length (up to 300 km) and are really draa in terms of
size. Size may be regarded as a distinguishing property of seifs, but the term
has been applied to much smaller features.
In addition to the basic patterns described above, other types have been
reported including dome-shaped, reversing, and star (McKee 1966, 1979),
and coppice dunes (Melton 1940) that are fixed by clumps of vegetation. It is
important to understand that all dune types may be present side-by-side in a
single region with the same prevailing wind, indicating that other controls are
significant in pattern development. Thus, variables of multiple wind direc-
tions,topography, size and abundance of sand, and vegetation may be so in-
constant that complex patterns are the rule, and the simple patterns of our
classification the exceptions. It seems reasonable, therefore, to use the sim-
plest ideal forms — barchans and —
linear types as models to demonstrate how
Chapter 8 324
Figure 8.14.
(A)Barchan dune in Sherman
County. Ore September 1899
,
(Photo by G K. Gilbert, U S
Geological Survey)
(B)Asymmetric wind ripples on
compound barchan dune
side of
Dawson County. Mont
September 1928 (Photo by C E
Erdmann. US Geological Survey)
(C) Dune forms in the Tularosa
Basin. Otero County, NM Wind
direction from lower right
is
Dunes are crenulated transverse

in lower right of photo Dune
forms progressively change to

individual barchans and finally
become U-shaped in the
downward direction
Figure 8.15.
Stages of seif development from
barchan under influence of
secondary wind direction: P =
/ / primary wind: S = secondary

wind
controlling factors may influence dune patterns. This choice is predicated on

the generally accepted premise that one forms perpendicular to the wind and
the other parallel to However, dunes oriented at oblique angles to prevailing
it.
winds also are found every region where dunes are reported. Thus, even
in
though linear and transverse dunes are most easily related to wind action, they
may be less important in understanding processes than the explanations of the
oblique forms.
Bagnold (1941) explained barchans as forms that develop in an area where
sporadic sand patches exist within a desert pavement. assumed (Bagnold
It is
1941, p. 222) that certain conditions must be met in order to mold the bar-
chanoid form: (1) a constant, unidirectional wind; (2) a rate of sand supply
that is symmetrically distributed on either side of the longitudinal axis; and
(3) a slip face that is completely sheltered so that all sand crossing the crestal
zone is trapped on the lee slope. Under these constraints, sand movement will
be fastest near the lateral edges where the sand patch thins to meet the surface
of the adjoining desert pavement. The greatest vertical growth of the sand
body occurs in the middle segment, decreasing gradually to the edges. Thus,
the lateral extremities will advance more rapidly than the center, and the cres-
centic form will develop as the equilibrium state is attained. As Cooke and
Warren (1973) point out, however, Bagnold's hypothesis does not explain why
barchans seem to have a regularly repeated width, nor does it account for the
tremendous diversity in the length and direction of the barchan horns. Those
authors suggest that divergence from the ideal form is a function of complex
secondary flow patterns in the wind and of differential sand supplies. In fact,
as suggested above, ideal barchans probably develop only where there is a
limited supply of sand and unidirectional winds. Abundant sands lead to linear
dunes rather than individual barchans.
Parallel ridges of sand, either dunes or bigger draa-sized features, are
prevalent in all large sand deserts. Often the linear ridges bifurcate in the
upwind direction, forming a Y or tuning-fork junction. Melton (1940) inter-
preted this phenomenon as the result of the wind excavating blowouts in deep
sand. As such, the dunes are partly erosional, a hypothesis accepted by Folk
(1971a, 1971b) for the linear dunes in northern Australia, but rejected by
Cooke and Warren (1973).
The origin of linear dunes and draa has traditionally revolved around their
relationship to the prevailing wind direction. Bagnold (1941) suggested that
seifs develop when barchans are modified by crosswinds so that one of the
horns is dramatically extended, as shown in figure 8.15. According to this idea,
linear dunes are not aligned parallel to the prevailing wind, but their trend
represents the resultant direction of more than one wind. The secondary wind
influence may arise from storm winds, diurnal reversals, or seasonal changes
Chapter 8 326
Barchanoid
Transverse ridge Linear Linear Star
Dune type A B C D E
Area Walvisbaai, South- Bahrain, Arabia Badanah, Arabia Timimoun, Algeria Ghudamis. Libya
West Africa
DP 518 540 528 245 658

RDP 448 435 327 46 46
R'DP/DP 0.86 0.81 0.62 0.19 0.07
Figure 8.16.
Five commonly occurring in wind direction. The multiwind hypothesis has received support from workers
relationships ofmodes on sand investigatingdune formation in a variety of physical environments (Cooper
roses and related dune patterns.
(A) Narrow unimodal. (B) Wide
1958; McKee and Tibbitts 1964; McKee 1966; Lancaster 1980).
unimodal. (C) Acute bimodal. The possibility of a resultant wind being generated by the influence of
(D) Obtuse bimodal (a special crosscurrents on the prevailing wind has never been questioned, but consid-
example in which the modes are
almost exactly opposed).
erable argument remains as to how much and what type of geomorphic work
(E) Complex D.P. (drift potential) it can accomplish. Substantial evidence has been reported to support the con-
and RDP. (resultant drift
clusion that seif chains are aligned with the resultant of several prominent
potential) in vector units. Arrows
indicate resultant drift direction
wind directions (Cooper 1958; McKee and Tibbitts 1964; Brookfield 1970;
(R D D.) (Modified from Fryberger Warren 1970). However, other data argue against a major genetic role for
1979) resultant winds. For example, seifs are often oriented in more than one direc-
tion within the same region, and some are distinctly oblique to the resultant
wind direction (Brookfield 1970; Warren 1971).
It seems clear from the above that the relationship between wind direction
and dune pattern is more complex than we might think. Nonetheless, eolian
features are generally so wind dominated that landform characteristics can
be utilized to make interpretations concerning wind direction and velocity.
Where features are large, Landsat imagery can be employed to make such
interpretations, and normal aerial photography is more suited to document
wind properties from smaller features (for general discussions see McKee 1979;
El-Baz and Maxwell 1982; Marrs and Kolm 1982).
A number of studies have produced rather interesting insights as to how
some of the complexity might be resolved. Fryberger (1979) and Fryberger
and Ahlbrandt (1979), for example, have evaluated wind data by estimating
the transporting capability of the wind and plotting those estimates in sand
roses (fig. 8.16). Sand roses are essentially circular histograms in which the
lengths of arms are proportional to the amount of sand that can be transported
toward the center of the circle along a given direction. For convenience, sev-
eral terms were employed in these studies to indicate energy and directional
properties. Drift potential (D.P.) is a measure of the total annual transporting

capability of the wind at any locality. Resultant drift potential (R.D.P.) mea-
sures the net transport capability after considering all wind directions at the
locality.
The significance of this technique demonstrated

in figure 8.16. Note
is
that the type and complexity of the dune changes

wind regime becomes
as the
more complex. Simple dune forms develop when the R.D.P./R.D. ratio is high
and the winds are strongly unidirectional. As the winds become multidirec-
tional, the R.D.P./R.D. factor decreases and more complex dune forms result.
The star dune, probably the most complex, forms where significant winds come
from three or more directions.
In summary, the complete understanding of depositional processes and
features is veiled in the complexity of a multitude of interacting geomorphic
systems. Features usually occur as ripples and dunes, but megascaled forms
called draa are found in the great sandy ergs of North Africa. Each feature
of the hierarchy occurs in patterns that are mainly a function of the wind
properties. The major dune patterns are probably oriented parallel and per-
pendicular to the prevailing wind direction, but enough forms trend obliquely
to suggest the possibility of a more complicated relationship involving form
adjustment to multidirectional winds. In addition, vagaries of grain size and
sorting, vegetation, topography,and sand supply introduce additional com-
plications,and geomorphologists are only beginning to understand the influ-
ence of environmental changes on the genesis of dunes (Twidale 1972). Some
dunes undoubtedly formed under wind and sand conditions that no longer exist,
and these relict forms confuse the issue even more because their geomorphic
characteristics may not be in equilibrium with the modern wind and sediment
conditions.
Fine-Grained Deposits
Our examinationof wind action in geomorphology logically concludes with a
look at sediment that is not normally moved near to or in contact with the
surface. Most silt and clay transported by the wind is carried in suspension
and comes to rest as blanketlike deposits of loess (fig. 8.17). Loess is usually
characterized as homogeneous unstratified silt, up to 100 m thick, which is
highly porous and has the capacity to maintain vertical or nearly vertical slopes
(Lohnes and Handy 1968). It covers all surfaces regardless of their topo-
graphic position, capping drainage divides as well as valley bottoms. Most loess
is moderately well sorted, with nearly 50 percent of the deposit consisting of
silt grains between 0.01 and 0.05 mm
in diameter (Pesci 1968). It usually
contains significant amounts of clay (5 to 30 percent) and 5 to 10 percent sand.

The mineral composition is fairly consistent, with quartz being dominant and
feldspars, carbonates, heavy minerals, and clay minerals present in smaller
amounts. Each of the minor constituents varies in percentage according to

local controls. For example, the amount of calcium carbonate in loess tends
to be higher in dry regions, while clay mineral content increases with greater
humidity.
Chapter 8 328
The hypothesis that loess originates as wind-blown dust stems from at

least a century of observations. The conditions that facilitate its wind deri-
vation are an abundant supply of loose, fine-grained sediment, moderate to
strong prevailing winds, and a surface free from a continuous vegetal cover.
Deserts obviously meet these requisites (Pewe 1981), but curiously several
continents with vast deserts (Africa, Australia) have almost no loess deposits,
and some regions notably lacking in desert conditions have experienced major
loess deposition. Chief among these is the periglacial environment (Pewe and
Figure 8.17. Journaux 1983), especially where glacial meltwater has spread vast outwash
Loess deposit on east side of
Mississippi River valley near
debris in the path of the prevailing winds (Pewe 1955). There silt is winnowed
Chester, III.
from the outwash before the surface can be fixed by vegetation. In the mid-
continental United States, for example, outwash is a logical source for many
whose distribution shows a marked affinity
loess sheets, to the outwash bodies
occupying large valleys of the region. Presumably the loess can grow to its
considerable thickness because the silt supply is continuously replenished
during annual or even diurnal flooding by the proglacial rivers.
In either hot or cold loess sources, deflation, saltating grains, or other pro-
cesses entrain the silt particles into the higher velocity zones, and with proper
turbulence and lift, the grains can rise to elevations of several kilometers. When
conditions are right, the amount of dust lifted into the atmosphere can be
enormous. In a 1935 dust storm over the interior plains, about 5 million tons
of dust was estimated to be in suspension over a 78 km 2 area near Wichita,
Kansas, and at least 300 tons/km 2 of dust was deposited in one day of the
same storm near Lincoln, Nebraska (Lugn 1962). The topographic effect of
such loess deposition is different from that of most constructional geomorphic
processes; it molds no important landforms but tends to form plains by filling
in depressions and thereby smoothing out preexisting relief. Dust is carried as
particulate matter suspended in the air, and its transport distance depends
mainly on the constancy of the wind velocity, both vertical and horizontal, and
the settling velocity of the grains. Loess constituents can be transported for
hundreds of kilometers and deposited over vast areas (Van Heuklon 1977).
Although the shape of grains included in loess deposits and the markings
on their surfaces suggest wind transport (Millette and Higbee 1958; Krinsley
and Donahue 1968), wind-blown loess is difficult to distinguish from silts de-
posited in other ways. In fact, much controversy has arisen in the past because
sediments bearing the characteristics of loess have been deposited by processes
such as mass wasting, wash, weathering, and fluvial action. This has produced
what some authors refer to as the "loess problem." Obviously some confusion
results when reworked by other transporting processes, and some of
loess is
the clay and CaC0 3 comes from weathering of the original mass. In addition,
"loess" often thickens towards topographic lows, indicating that downslope
movement has been prevalent after the material's original deposition. None-
theless, the relationship between widespread blanket deposits and the pre-
sumed prevailing winds often so clear that an eolian origin for the original
is
concentration of the material cannot be denied. In fact, arguments for a wind

origin of the loess in the midcontinental United States are quite strong
(Swineford and Frye 1951; Lugn 1962, 1968).
Some of the problem can be resolved if we restrict the term "loess" to
those deposits that have the characteristics described and are definitely of wind-
blown origin. Proof of an eolian genesis may be circumstantial in some workers'
minds, but the evidence of field relationships and deposit properties is often
more than convincing. In the western Great Plains, for example, the thickest
loess (up to 70 m) occurs immediately downwind from the Sand Hills Region
of Nebraska, which contains the largest accumulation of sand dunes in North
America. In Illinois, as the map in figure 8.18 shows, the loess is thickest on
the bluffs overlooking the Mississippi and Illinois valleys and thins gradually
eastward into Indiana. The usual decrease in loess thickness downcurrent is
Chapter 8 330
Figure 8.18.
Approximate thickness of loess on
level, uneroded topography in
Illinois.
~2 > 300" of loess
200-300" of loess
120-200" of loess
60-120" of loess
I 36-60" of loess
| <36 of loess
Major alluvium and

m outwash loess sources
so systematic it can often be expressed mathematically. For example. Hutton

(1947) showed that loess in Iowa southeast of the Missouri River thins ac-
cording to the equation
Y= 1250.5 - 528.5 log X,
where Kis thickness in inches and A" distance in miles. Different equations fit
other situations (G. D. Smith 1942; Frazee et al. 1970), but all demonstrate
a regular, systematic decrease in thickness with increasing distance from the
source. Additionally, as might be expected, the particle size also decreases
North latitude
41°
Holocene
stage
Q.
o
o Valderan
5
D O substage
Twocreekan
o
n
CO
o Woodfordian Wedron
c o
— o substage formation
(0 o
DC o"
Farmdalian
substage
Figure 8.19.
downwind, again with a demonstrable regularity (Ruhe 1969). It is not safe Stratigraphic relationships
to assume, however, that these systematic relationships are the result of one between loess deposits and
glacial deposits in the Late
strongly prevailing wind. Indeed, Handy (1976) reminds us that significant
Wisconsinan of Illinois. (After Frye
loess deposition in the upper midwest of the United States occurred on both and Willman 1970)
sides of the river valley source areas. This indicates that variable winds are
responsible for loess deposition and the gradual changes in its character.
Although historical geomorphology is not the prime concern of this book,
perhaps the most interesting aspect of loess is its value in deciphering Qua-
ternary stratigraphy. Individual loess deposits are widespread geographically
and blanket diverse topography, so that the deposits are ideal for stratigraphic
analyses. As mentioned, loess deposition is commonly related to periods of
glaciation. Where several loesses are superimposed, each deposit having been
formed during a specific glacial episode, the sequence provides a useful strati-
graphic framework. The upper and lower boundaries of each unit within the
sequence can usually be identified because each loess has distinctive properties
such as soil profiles (often buried), textural or mineralogical charactristics, or
fossils. These criteria may be used for regional correlation of the deposits and
therefore as bases for interpreting Quaternary history.
The use of loess in regional correlation requires caution and careful field
study because loess properties, like properties of any rock substance, change
laterally. In addition, in glaciated areas, single loess units may divide at the
position of the former glacial margin so that loess associated with the same
glacial event lies both beneath and above the till of that glaciation (fig. 8.19).
Beyond the glacial margin the two loess bodies are combined into one, indi-
cating that loess was deposited continuously during the complete cycle of gla-
cial advance and retreat. The Peoria Loess in Illinois, for example, was
deposited throughout the entire Woodfordian substage and correlates with both
the Richmond Loess and the Morton Loess, which were deposited before and
after the Woodfordian till (fig. 8.19).
Chapter 8 332
The fossil assemblage contained in loess also provides useful information

about the environmental conditions at the time of deposition. The predomi-
nant fossils in loess are land snails. They commonly are abundant when the
loess is thick, but are rare in thin loess because leaching destroys the carbonate
shells. Snail ecology indicates that most loess in the North American mid-
continent originated when the climate was slightly cooler and moister than
today. However, considering the diverse environments in which modern loess
originates, it is probably safe to say that loess by itself has little climatic sig-
nificance. Climatic indicators are substances such as plant fragments or mol-
lusks included in the wind-blown debris at the time of its deposition or shortly
thereafter. As shells and wood can be dated by l4 C, loess stratigraphy is often
definable in terms of absolute time. This is especially important in the stra-
tigraphy of the Late Wisconsinan. Older deposits may be beyond the limits
of validity for the radiocarbon method.
Summary Wind is an effective geomorphic agent in regions with sparse vegetation and
an abundant supply of unconsolidated sediment. Although the most discern-
ible evidence of wind action is found in deserts, it must be emphasized that
eolian processes function in any locality having strong winds and the proper
conditions of vegetation and sediment. The amount of geomorphic work ac-
tually accomplished depends also on the properties of the wind, especially its
velocity and turbulence.

Particle entrainment occurs when the wind reaches a critical velocity called
the fluid threshold. For any given sediment size, the value of the threshold
velocity depends on a number of variables such as particle shape, sorting, soil
moisture, and surface roughness. Once in motion, however, particles may strike
stationary grains and cause their entrainment at wind velocities well below
the fluid threshold (impact threshold). Wind-transported sediment moves in
suspension, by saltation, or by surface creep. The largest portion of the load

is carried within 2 m of the surface.
Geomorphic features associated with wind action are both depositional
and erosional, ranging in size from microscopic to those measured in kilo-
meters. Erosional features develop primarily from the abrasive action of wind-
blown sand. The most prominent depositional features occur in a hierarchy of
bed forms, including ripples, dunes, and draa, that are produced from sand-
sized debris. The geometric shape and spacing of bed forms probably reflect
an equilibrium condition between the wind and the characteristics of the sand.
The system mechanics, however, is so complex that a precise quantitative
expression of the equilibrium relationship is not yet possible. Fine-grained sed-
iment occasionally accumulates in dune form, but most wind-blown silt and
clay are deposited in sheets of loess that tend to smooth out topography rather
than collect into pronounced constructional features.
Ahlbrandt. T. S., and Fryberger, S. G. 1982. Eolian deposits. In Sandstone

depositional environments, edited by P. Scholle and D. Spearing,
pp. 1 1-48. Tulsa: American Association of Petroleum Geologists.
Bagnold, R. A. 1941. The physics of blown sand and desert dunes. London:
Methuen and Co.
Chepil, W. S., and Woodruff, N. P. 1963. The physics of wind erosion and
its control. Advances in Agron. 15:21 1-302.
Cooke, R. U., and Warren, A. 1973. Geomorphology in deserts. London:
Batford Ltd.
Marrs, R. W. and Kolm, K. E., eds. 1982. Interpretation of windflow
characteristics from eolian landforms. Geol. Soc. America Spec. Paper
192.
McKee, E. D., ed. 1979. A study of global sand seas. U.S. Geol. Survey
Prof. Paper 1052.
Sharp, R. P. 1964. Wind-driven sand in Coachella Valley, California. Geol.
Soc. America Bull. 75:785-804.
Wilson, I. G. 1972, Aeolian bedforms-their development and origins.
Sediment ology 19:173-210.
"^~
Glaciers and Glacial Mechanics
9
I. Introduction 2. Slip
II. Glacial Origins and Types 3. Variations with Time
III. The Mass Balance V. Ice Structures
IV. The Movement of Glaciers A. Stratification
A. Internal Motion B. Secondary Features
B. Sliding 1. Foliation
C. Velocity and Flow 2. Crevasses
D. Complications of the Simple VI. Summary
Model VII. Suggested Readings
1. Extending and
Compressive Flow
335
Chapter 9 336
Introduction Some of the most spectacular landscapes in the world are the results of the
erosional and depositional action of glaciers, and every textbook of physical
geology and geomorphology includes numerous photos and descriptions of these
remarkable features. Nonetheless, to be true to the theme of this book, such
features, regardless of their unique beauty, will be considered only because
they manifest the processes that form them. It would also be tempting to stress
the stratigraphic relationship between different glacial deposits and the effects
exerted by glaciation on climate,life, and the physical setting. Obviously these
subjects are of paramount interest and significance to geomorphologists be-

cause much of the modern landscape is closely associated with the physical
events that occurred during the alternating glacial and interglacial stages of
the Quaternary Era. While recognizing the importance of these topics, I re-
alize pragmatically that an attempt to treat them in a general way is destined
to fail. Therefore, our goal in this chapter will simply be to understand gla-
ciers: how they form and how they move. Glacial erosion and deposition and
the landforms resulting from these processes will be examined in the chapter
following.
The study of glacial mechanics is an intimate part of the science of gla-
ciology. Although this field is in its infancy, many of the techniques developed
in recent years to study glaciers are extremely sophisticated and involve geo-
physics, remote sensing, and computer analyses well beyond the scope of an
introductory discussion. Excellent reviews detail the basic concepts of the dis-
cipline (Embleton and King 1968, 1975a; Sharp 1960; Paterson 1969, 1981;
Shumskii, 1964; Sugden and John 1976; Colbeck 1980), but those desiring
greater depth and discussions of more recent advances in the field should refer
to current periodicals, especially the Journal of Glaciology, which contains
articles pertinent to all aspects of our discussion.
Glacial Origins A glacier is a body of moving ice that has been formed on land by compaction
and Types and recrystallization of snow. Assuming this simple definition suffices, it is
obvious that two critical requirements must be met before a mass of ice can
be called a glacier. First, the ice must be moving, either internally or as a
sliding block, and second, the mass must be due to the accumulation and meta-
morphism of snow. Areas in which the winter snow is entirely lost to summer
melting and other forms of dissipation cannot bring forth glaciers, even if the
amount of snowfall is enormous. Where a portion of the snowfall does survive
the summer melt, it is buried by the next winter's accumulation; with con-
tinued annual increments, the snow pack grows in size, changes its constituent
properties, and finally, with enough mass, begins to move. In any region a
snowline or ihefirn line, above which some
specific elevation exists called the
snow remains on the ground perennially and permits the formation of a gla-
cier. In polar climates, snowlines are usually near or at sea level, and they
gradually increase in elevation in climatic zones with higher annual temper-

atures. However, temperature is not the only determinant of snowlines since
337 Glaciers and Glacial Mechanics
Table 9.1 Increasing density of snow during transition to ice
Materials Density (gm/cc)
New snow 005-007

Firn 0.4-0.8
Glacier ice 0.85-0.9
glaciers form at lower elevations at the equator than in the desert zones near
30° north. This somewhat ironic situation arises because the annual precip-
itation in the two regions is vastly different.
When snow accretes over a period of years, changes in the properties of
the particles mark the transition of a snow pack into true glacier ice. Newly
fallen snow having a very low density (usually 0.05-0.07 gm/cc) and a del-
icate hexagonal crystal structure is transformed into glacier ice through a se-
ries of complex but recognizable stages (fig. 9.1). In the initial phase, points
more spherical
of the crystalline flakes are preferentially melted, resulting in a
particle shape and tigher packing due At the same
to settling of the grains. Initial form of snow
(A)
time, water produced by melting percolates to the base of the snow pack where
it refreezes in the pore spaces (Woo and Heron 1981). As table 9.1 shows, this
tends to decrease porosity and dramatically increase the density. The time
needed for this initial change varies depending on the climate and the pressure
added by continuous accumulation of snow; where temperature remains near
freezing or where partial melting occurs, the transition from fluffy snow to
coarse granular snow may take hours or days (Embleton and King 1968). On
the other hand, extremely frigid conditions retard the process such that years After two weeks
may be required. (B)
In temperate regions,
colating water and is

snow lying on the surface for a complete year be-
comes granular and usually increases in density to about 0.55 gm/cc through
the rounding and settling processes just mentioned. The material, now known
as firn, is much denser than the original snow but is still permeable to per-
not yet true glacier ice (table 9.1).
cation beyond this point, and the mechanics of further transformations is
The rate of densifi-
*
After seven
(C)
weeks
different. The time needed to convert firn to ice is also variable, depending
once again on temperature and the rate of increased load on individual grains. After eight weeks
Where accumulation is rapid and water is plentiful, the transition from firn (D)
to ice probably takes place in less than 50 years, but in drier, colder regions
the process may take several hundred years. Figure 9.1.

The mechanics involved in transforming firn to ice produces enlarged grains Changes in the shape of snow
crystal in the transition to firn
(in some cases up to 10 cm long) by recrystallization (fig. 9.2). When densities
are low and porosities high, most of the stress exerted by the accumulating
load is in the form of vertical compression. With increased density, however,
the stress pattern is hydrostatic, and crystal growth occurs in any or all di-
rections. Melted water may occupy air spaces lower in the firn pack, and as
mentioned above, regions that produce free water will make the transition
from firn to ice much more rapidly. Gradually, pore space is eliminated by
crystal growth or by freezing of the downward permeating meltwater until

the density is approximately 0.8-0.85 gm/cc. The only air remaining is trapped
Chapter 9 338
as bubbles within the crystal. For all practical purposes, the creation of ice is
complete at this stage, even though the bubbles continue to be slowly expelled
by compaction, and the density may increase to about 0.9 gm/cc.
The rate of density increase and crystal growth is closely related to the
temperature of firn. Firn in temperate regions will transform into ice faster
and at a shallower depth than it will in polar regions. This fact is probably
best demonstrated by the depth-density relationship in glaciers from the two
regions (Sharp 1951; Behrendt 1965). In the Antarctic, a 0.85 gm/cc density
is not achieved until at least 80 m of firn is accumulated. This is in drastic
contrast to the 13 m depth needed to change snow to ice on the temperate
Upper Seward Glacier in Alaska. Since accumulation rates are greater in gla-
ciers of temperate regions, it follows that the greater depth to true ice in the
Antarctic ice sheet represents the monumentally greater time needed to create
glaciers in dry, cold climates.
Glacier ice is not preserved intact after its creation but is susceptible to
further changes with continued increase in stress. For example,
it seems clear
from experimental work that the random orientation in polycrystalline ice

cannot be maintained at higher stresses. The process known as recrystalli-
zation occurs when deformation of the polycrystalline mass exceeds several
percent. This suggests that changes in the orientation of the c axis (known as
the fabric) should occur in a vertical plane of a glacier. Such changes in fabric
with depth have been noted in many glaciers (Gow and Williamson 1976; Rus-
sell-Head and Budd 1979; Hooke and Hudleston 1980, 1981). In most cases,
the fabric changes from being a weakly oriented c axis in fine-grained, near-
Figure 9.2. surface ice to coarser ice with a broad, single maximum fabric at some lower
Increase in gram size caused by
recrystallization process level. With increasing depth, the fabric usually develops multiple orientation
(A) Original texture of water- directions. Finally, in some, but not all, glaciers the fabric returns to a single
soaked snow (B) Texture after
maximum orientation at the glacier base (Hooke and Hudleston 1980).
application of stress
The depth at which fabric transition occurs is variable and depends on a
complex interaction of density, impurities, temperature, etc. Importantly, the
c-axis orientation may result from and be related to the overall scheme of
glacier movement because the orientation of basal crystal planes is probably
parallel to the direction of shear (Steinemann 1954, 1958; Rigsby 1960).
Glaciers have been classified on the basis of many salient properties, but
as with any physical phenomenon, the best classification depends on the prime
purpose of the groupings. The classification used here is no more astute than
others; it simply serves our needs better by relating more closely to the pro-
cesses of glaciation. Ahlmann (1948) suggested several bases for the classi-
fication of glaciers as morphological, dynamic, and thermal. Morphological
classifications, which are based on glacier size and the environment of its
growth, are most commonly employed, and the different categories, such as
valley glaciers and ice sheets, are undoubtedly familiar. Although Ahlmann
(1948) recognizes many morphological subdivisions, Flint (1971) suggests that
glaciers can be placed in three broad categories {cirque glaciers, valley gla-
and ice sheets) and two intermediate types {piedmont glaciers and
ciers,
mountain ice sheets). In this abbreviated form the classification is quite useful,
especially in a descriptive sense.
r
x
Figure 9.3.
The La Perouse Glacier in Alaska
A very active glacier that spreads
onto lowland when it emerges
from a narrow valley Crevasses
are distinct, and annual dark
bands (ogives) are clearly visible
For our process orientation, however, the other two classifications are
probably more useful, and we will at times utilize parts of each. The dynamic
classification is based on the observed activities of glaciers and consists of three
main groups: active, passive, and dead glaciers. Each type is closely related
to the balance between losses and gains of ice and probably also depends some-
what on thermal properties. Active glaciers, like the one in figure 9.3, are char-
acterized by continuous movement of ice from their accumulation zones to
their edges. The movement may occur in response to normal snow accumu-
lation, or it can be generated by avalanches or ice falls that provide the im-
pulse for the forward motion. As you might expect, a is passive when
glacier
its movement is minimal. Dead ice has no discernible
internal movement.
The thermal classification is based primarily on the temperature of the
ice. It is well known that glacial behavior is directly influenced by ice tem-
perature, and therefore the thermal classification should lend itself to process
Chapter 9 340
analysis. Two types of glaciers exist in this classification: temperate glaciers

and polar or cold glaciers. In temperate glaciers the ice throughout the entire
mass is at its pressure-melting point, although the upper 10 m may freeze in
the winter. Meltwater seems to be present in abundant amounts within or be-
neath the ice mass and, in contact with the underlying rock, often facilitates
slippage of the ice over the bed. This causes velocity and erosive action to be
generally greater in temperate glaciers than in other types. Near the snout,
meltwater may emerge as basal streams, or it may be temporarily dammed
within the ice; both situations promote extensive fluvial removal of debris from
the terminus of the glacier.
Polar glaciers were subdivided by Ahlmann into subpolar and high-polar
types. In subpolar glaciers the accumulation zone is characterized by a thin
layer of firn, perhaps 20 m thick, which contains some water in the summer
if temperatures are warm enough to melt the surface ice. The surface of a
high-polar glacier remains below freezing at all times, resulting in a com-
pletely water-free ice mass and a thick firn zone extending to at least 75 m
before true ice is encountered. The absence of meltwater within polar glaciers
is an extremely significant difference between polar and temperate glaciers,
and its geomorphic importance cannot be overemphasized. It requires that ice
at the base of polar glaciers be below its pressure-melting point and, for all
practical purposes, be solidly frozen to the underlying bedrock. Since slippage
cannot occur over the bed, ice movement is totally internal, and the glacier's
erosive action is greatly diminished.
The suggestion that temperate ice is at its pressure-melting point is per-
haps unfortunate because pressure is only one of several factors that deter-
mine the temperature within a Temperature variations can be caused
glacier.
by downward transfer of upward transfer of geothermal heat
surficial heat or
that is released at the glacier base (Sugden and John 1976; Hooke 1977). Heat
is also produced inside the ice by friction when the glacier moves. Thus, it
should be anticipated that a temperate glacier may have patches of frozen ice
at its base (Robin 1976; Goodman et al. 1979) and that portions of polar ice
sheets may have free water. In addition, the effect of climatic change may be
gradually superimposed on the other types of heat variation and may not per-
meate through an entire glacier for centuries. For example, it seems likely that
the West Antarctic Ice Sheet is still responding to a climatic change initiated
at the end of the Wisconsin stage (Whillans 1978; Thomas 1979). Thus, most
"temperate glaciers" have a temperature distribution that disagrees with the
pressure distribution predicted by the thickness of the ice (Harrison 1975).
even though the temperature generally increases with depth (Hooke et al.
1980). In some cases the temperature increase with depth may be linear (fig.
9.4); however this is not always true, and the gradient of temperature increase
may vary considerably from glacier to glacier or within a single ice mass (Hooke
et al. 1980). In sum, the simple model of pressure-melting may be an ap-
proximation, but it is not a precise truth.
Regardless of the difficulties discussed above, it is important to under-
stand that the condition of basal ice with regard to melting temperature is
perhaps the primary determinant of a glacier's ability to do geomorphic work.

-0.16 Figure 9.4.

Temperature-depth relationship on
a temperate glacier
-0.12
B -0.08
ID
Q.
£
-0.04
• October 1971
x September 1972
Ml 40
I I
80
I
120
I I I
160
fill 200 240
Depth (meters)
Glaciers with basal ice at the melting temperature tend to move faster, erode
more, and carry greater loads than polar glaciers. Therefore, we can expect
erosion and deposition to be more pronounced when they arise from temperate
ice. In addition, many differences between modern and Pleistocene glacial ac-
tivity may be due to variations in the thermal characteristics of the ice at the
different times.
The mechanical behavior of glaciers and the geomorphic work they accom- The Mass Balance
plish are intimately related to their mass balance, sometimes called the glacial
budget. The mass balance is essentially an accounting or budgeting of the gains
and losses of snow that occur on a glacier during a specific time interval. The
water equivalent of ice and snow added to a glacier during the period in ques-
tion is called accumulation and may result from a variety of processes in-
cluding snowfall, rain and other water that freezes on the surface, and
avalanches. Processes that remove snow or ice, collectively known as ablation,
commonly include melting, evaporation, wind erosion, sublimation (Beaty
1975a), or the breaking off of large blocks into bodies of standing water, a
process called calving (discussed in Holdsworth 1973). Losses by melting within
or beneath the glacier are usually minor compared with surface volumes and
are therefore neglected in budget studies.
The time interval used in most balance analyses is the budget year (or
balance year), which is ordinarily taken as the time between two successive
stages when ablation has attained its maximum yearly value. Usually these
values are achieved at the end of the summer season, but the two ablation
maxima may not occur on the same day, and so the budget year is not nec-
essarily 365 days long.
Chapter 9 342
Accumulation On a glacier surface two values of accumulation and ablation can be de-
(+) termined: (1) a gross annual accumulation or ablation, representing the total
volume of water equivalent added to or lost from the glacier during the budget
Firn line year; and (2) a net annual accumulation or ablation, representing the differ-
Equilibrium line
ence between the gross values and indicating whether there was an actual gain
Ablation
or loss of mass during the year. The latter value is simply the algebraic sum
(-)
of accumulation and ablation, and for the budget year at any single mea-
surement locality is called the net specific budget. If net specific budgets are
determined for a network of points distributed over the entire glacier surface,
their values can be integrated into the mass balance. Any glacier may have a
positive mass balance, meaning that more accumulation has occurred during
Figure 9.5.
Zones of accumulation and the year than ablation, a negative mass balance, indicating an excess of abla-
ablation on a glacier as
tion, or a mass balance of zero if the volumes of accumulation and of ablation
determined by budget analyses
have been precisely the same.
It should be obvious that even though both accumulation and ablation
occur on all parts of the glacier surface, the higher elevations of the glacier
will usually receive more accumulation and experience less ablation than the
lower reaches of the glacier, where the opposite is true. Thus, large areas with
either positive or negative net specific budgets can be identified on most gla-
ciers, as diagrammed in figure 9.5. The two areas are separated by the equi-
librium line, along which the annual volumes of accumulation and of ablation
are equal. The equilibrium line should not be confused with the firn line or
snowline because the two may not occur at the same place. On many glaciers,
the firn line is clearly marked as a contact between snow above the line and
dense blue ice below the line. On others, some of the dense ice downglacier
from the snowline may have been formed by refreezing of meltwater and, as
such, represents a net accumulation of mass
This superimposed ice
(fig. 9.5).
is found enough cases to warrant the distinction between the firn line and
in
the equilibrium line, and it sometimes creates a rather complex transitional

zone between the accumulation area and the ablation area (Miiller 1962).
Nevertheless, equilibrium lines and firn lines are usually close enough to ap-
proximate one another.
What does all this have to do with glacial mechanics and the resulting
processes of erosion and deposition? If glaciers are viewed as open systems,
both the mass balance and the absolute total amounts of accumulation and
ablation are important factors in the character of glacial movement. When
the net budgets are perfectly balanced (total mass balance equals zero), no
expansion or shrinkage of the glacier occurs, and the glacial extremities re-
main stationary. This equilibrium condition, however, is seldom maintained

for a long period of time, and so the glacial front and sides usually fluctuate
constantly. Glaciers with a positive mass balance actively advance and char-
acteristically maintain a relatively steep or vertical front. In contrast, an overall
negative budget induces recession of the glacial front, and the snout will be
gently sloping and often partially buried by debris released from the ice. The
mass balance, therefore, is closely related to the position and type of morainal
system constructed by the ice.
The absolute amount of snow and ice relative to the area of a glacier (gross 50
accumulation and ablation) directly affects its internal activity. Large gross-
(A) Temperate,
accumulation values on small glaciers promote very rapid flow from the ac- 40
highly active glacier
cumulation area to the ablation area; in most cases temperate glaciers are
likely to have large accumulation and ablation values and, consequently, high
flow velocities. In contrast, polar glaciers, which tend to be rather passive,
usually have small values of accumulation and ablation and low internal flow
(B) Polar glacier
velocities. It is very important to recognize that glaciers with different gross with low activity
values can have the same mass balance and that, in fact, if they are in an
equilibrium condition, no advance or recession of the glacier front is occurring
even though the internal transfer of ice may be enormous. Figure 9.6 is a hy- Winter Summer
potheticaldiagram demonstrating a polar glacier with low activity and a highly
-Âccumulation —Ablation
active temperate glacier, both of which have a mass balance equal to zero.
The snouts and lateral edges are stationary in both cases, but the temperate Figure 9.6.
glacier, having a large value for accumulation and ablation, is moving at a Annual budgets on two glaciers in
which the net mass budget of
higher rate internally. This dynamic activity causes pronounced erosion and
each equals zero. Active,
rapid transportation of debris through the system. With the proper budgets, temperate glacier (A) has greater
large moraines may result at the terminal and The polar
lateral boundaries. total amounts of ablation and
accumulation than the low-activity
glacier depicted in figure 9.6 will have little if any internal motion of its ice,
polar glacier (B). Even though
and as a result, will probably not form any significant depositional features. both are in equilibrium and their
Every glacier's depositional and erosional character, therefore, is fundamen- does
fronts are stationary, (A)
more work because must it
tally determined by the characteristics of the snow and ice added or lost from
transfer more ice during the year
its surface. from the accumulation zone to the
snout
Internal Motion The Movement

Several hundred years ago, through direct observations, Alpine residents re- of Glaciers
alized that glaciers move, and measurements of the rate of glacier flow were
made back as the early eighteenth century. It is now generally accepted
as far
that glaciers move by two mutually independent processes: (1) internal de-
formation of the ice, called creep, and (2) sliding of the glacier along its base
and sides.
Thefirst real attempt to explain the physical dynamics of englacial ice
motion seems to have been in the work of J. D. Forbes, who in 1843 suggested
that glaciers respond to stress much like a plastic substance. By the beginning
of the twentieth century, numerous models based on a plastic flow mechanism
had been developed, although different ideas were also being proposed. They
included flow caused by differential shearing along numerous, closely spaced
planes (Phillip 1920; Chamberlain 1928); and the widely accepted notion that
ice behaves as a viscous liquid that deforms in linear proportion to stress (La-
gally 1934). Each of these concepts suggests that glacier flow manifests a re-
lationship between ice deformation (strain) and the stress (force/unit area)
that produces it. Stress generated at any point can be separated into two parts,
hydrostatic pressure and shear stress. Hydrostatic pressure, which is related
to the weight of the overlying ice, is exerted equally in all directions. Shear
Chapter 9 344
Figure 9.7.
The relationship between stress
and deformation (A) Newtonian
viscous fluid Relationship Is
normally plotted as a function of
stress and the rate of strain
(B) Perfect plastic deformation.
The material experiences no
deformation until stress is
increased to the value of the yield
stress. The material will then
continue to deform as long as the
stress is applied. Normally plotted
as a function of stress and strain.
(C) Ice creep according to Glen's
flow law.
Strain rate
stress, however, is a function of the overlying weight and the slope of the gla-
cier surface. Therefore, shear stress, which causes glacier motion, can be cal-
culated as t = pgh sin a, where r is shear stress, p is ice density, g is acceleration
of gravity, h is glacier thickness and a is the slope on the glacier surface. This
relationship implies that the slope at the base of the ice is the same as that on
the ice surface. The relationship also defines the basic flow laws of ice. For
the purposes of demonstration let us consider the two simplest possibilities:
(1) that ice behaves as a viscous fluid and (2) that ice deforms as a perfectly-
plastic substance.
If ice behaves as a Newtonian viscous material, and we assume a constant
viscosity, the application of stress should result in a linear relationship between
the stress value and the strain rate (fig. 9.7). In addition, deformation will
begin in the ice as soon as stress is applied and will maintain the linear pro-
portionality regardless of the changes occurring in the stress. In contrast, as
the diagram shows, a plastic substance shows no immediate response to stress
but is capable of supporting a certain amount of stress without sustaining any
deformation. Thus, at low stresses the strain of a plastic will be zero. As stress
is increased, however, it eventually attains a value, called the yield stress, where
the ice will experience limitless and permanent deformation. An entire glacier
would behave plastically only if the shear stress along its base were equal to
the yield stress (Kamb 1964).
It is now known that neither of the simple cases discussed above prevails
in the mechanics of ice motion, although in many instances glacial properties
predicted according to pure plasticity closely approximate the observed char-
acteristics (Nye 1952b), and therefore, glacial movement has been referred
to as pseudoplastic (Meier 1960, Johnson 1970). Laboratory studies since the
late 1940s have shown that single ice crystals under stress deform as soon as
stress is applied. Under any given stress the strain will increase rapidly at first
(Glen 1952, 1955), but within a short time (tens of hours) it approaches an
almost steady value that plots as a nearly straight line on a graph (fig. 9.7).
This continuous deformation with no increased stress is the creep process that
allows glacial ice to flow steadily under its own weight.
The rate of strain during the creep process is related to varying stress
values by the following equation derived by Glen (1952, 1955) and now com-
monly referred to as the power flow law:
e = kr",
where t is is stress, and k and n are constants. The

the strain rate (dt/dr), r
values of «, determined by a number of investigators, seem to vary from ap-
proximately 2 to 4 for individual crystals. In polycrystalline ice they range
from 1.9 to 4.5, with the mean value being close to 3. In any case, the n values
associated with the power flow law are significantly greater than the value Time-
1,
required for a linear viscous flow. The power flow law indicates that minor
changes can produce a major response Figure 9.8.
in stress in the strain rate. For example,
The effect of temperature on the
If n = 4, doubling of t increases the strain rate 16 times. creep deformation curve. All
It is significant to note that glacier ice is always polycrystalline. The flow curves represent deformation
law, although based on single crystal deformation, under a stress of 6 bars. (After
still seems to predict the
Glen 1955. Used with permission
response of glaciers (Thomas et al. 1980), although minor modifications of of the Royal Society of London)
the equation may be needed (Meier 1960; Colbeck and Evans 1973). The value
of k in glacier flow might be reduced by interference of adjacent grains and
recrystallization. In addition, as shown in the diagram in figure 9.8, cold ice
deforms less readily mainly because the constant k is also

than temperate ice,
dependent on temperature. In Glen's experiments the value of k decreased by

two orders of magnitude (0.17-0.0017) when the temperature was lowered
from 0°C to — 13°C. In fact, Paterson (1969) suggests that the strain pro-
duced in ice at — 22 °C by any stress is only 10 percent of its value when the
ice is at 0°C.
Thus, ice clearly does not behave like a viscous fluid, although it may
approximate a viscous response under low stress when the strain rate is still
in its transient phase. At stress <
1 bar in temperate ice, for example, n values
as low as 1.3 have been measured (Colbeck and Evans 1973). However, when
the nearly steady strain is attained or the ice is under high stress, the defor-
mation becomes more plastic. Although this condition approximates plas-

ticity, there is no distinct yield stress associated with the creep process,
indicating that ice is not a perfectly plastic material.

Much research has been done to determine why ice under high stress takes
on a nearly plastic behavior. The answer probably rests in the mechanics of

strain. Ice deforms by recrystallization and by slip within the grains, a process
involving a dislocation of planes of atoms inside the crystal (see Glen 1958;
Weertman and Weertman 1964). The dislocation process works most effi-
ciently through gliding on the basal plane of the ice crystal. In fact, it was
suggested years ago that presumed plastic flow was closely associated with ice
crystals lying in a preferred orientation (Perutz 1940). It follows that a con-

sistent orientation of grains in polycrystalline ice mass (ice fabric) would pro-
duce strain exerted in the same direction (Russell-Head and Budd 1979; Duval
1981).
Chapter 9 346
In sum, the power flow law derived in Glen's laboratory seems to fit the
actual observed motions of glaciers. As mentioned earlier, it is too much to
expect that polycrystalline ice will respond in precise agreement with the flow
law, and modifications of the equation are generally required as the situation
demands. In addition, where stress systems are complex because of irregular
valley floors, cross-sectional shapes, or nonisothermal ice, the flow law will
have to be generalized or modified (Nye 1957, 1965). Attempts to go beyond
the general approach and replace the simple power law with a more sophis-
ticated mathematical model become extremely difficult (Hutter 1982). Thus,
in spite of inherent difficulties, Glen's power flow law seems to provide an ex-
cellent first approximation between theoretical and observed flow data and,
with proper modification, probably best describes the internal mechanics of
glaciers.
Sliding
The difficulty in applying the power flow law directly to glaciers is in part due
to the fact that glaciers also move by sliding over the underlying bedrock, this
motion being in addition to the creep operating within the ice mass. Sliding
processes operate at the contact between the underlying bedrock and the base
of the glacier, or within the lower ice layers. The processes themselves are
poorly understood because direct observation of their action requires tun-
neling through the ice mass to the bedrock floor, an endeavor seldom tried and
more rarely accomplished. Nonetheless, it is generally accepted that ice tem-
perature and the character of the bedrock-ice interface are the prime factors
that determine how sliding actually works.
Perhaps the most intuitive and acceptable process of sliding is the phe-
nomenon of glacial slippage over a water layer that rests between the under-
lying rock surface and the base of the glacier. Where basal ice is near the
pressure-melting point, a thin film of water, probably only millimeters thick,
can exist at the bedrock-ice interface. This water not only lubricates the un-
derlying surface, but if the subglacial water is under enough hydrostatic pres-
sure, it may partly offset the weight of the overlying ice. This, of course,
increases the movement produced by slipping of the glacier on the water layer
(Weertman 1964).
The critical unresolved questions about slippage are whether water can
exist as a continuous sheet beneath a glacier and, if not, how much of the ice-
rock contact must be occupied by water to produce an influence on sliding.
We know that water-filled cavities exist along the basal contact (Haefeli and
Brentani 1955; J. E. Fisher 1963; Savage and Paterson 1963; Vivian and Bou-
quet 1973), and although they are usually isolated pockets, they might pos-
sibly merge by enlargement or coalesce as they migrate along with the moving
ice. Lliboutry (1964, 1968) developed a sliding model based on the hypothesis
that ice will separate from its bedrock floor downstream from a bedrock ob-
struction, and that the cavities thus formed will fill with subglacial water under
pressure. Assuming an irregular bedrock topography, water-filled cavities may
nearly submerge the obstacles, thereby reducing the area of contact between
ice and rock. The reduction of overall friction thus produced will allow a drastic
increase in sliding velocity.
K -
Although the above model has been questioned, the presence of pressur-
ized water at the base of a glacier has been demonstrated
(Hodge 1976). Such
water may play a fundamental role in the mechanics of sliding (Bindschadler
1983) and, in fact, may cause uplift of a glacier surface during the beginning
of a melt season (Iken et al. 1983). Clearly, pressurized subglacial water can
be a significant factor mechanics. It cannot be stated, however, that
in slippage
such water is always part of those mechanics because the generation of effec-
tive hydrostatic pressure requires that the underlying material be imperme-
able. Where underlying bedrock is pervasively fractured or the subsurface
composed of unconsolidated sediment, it is probable that free water will drain
into the underground system. Such downward release of water should lessen
the possibility of creating enough hydrostatic pressure in the remaining in-
terstitial water to facilitate slippage. It is also possible that subglacial water
may drain in well-defined channels or in a sediment layer between the bedrock
and ice bottom (Engelhardt et al. 1978). In either case, downslope water
drainage will lower the hydrostatic pressure and reduce sliding velocity
(Chadbourne et al. 1975; Engelhardt et al. 1978).
The premise that lubrication of the bedrock surface, aided by hydrostatic

pressure, will initiate slippage is probably easy to accept, but the question
arises as to how ice can move over an irregular surface by this mechanism
alone. A glacier slipping
up a slope on a film of water stretches credibility and
suggests that other processes must be involved in the sliding phenomenon. Two
such processes that have been recognized are regelation and enhanced creep.
The process of regelation involves the melting and refreezing of ice due to
fluctuating pressure conditions and usually results in a texturally distinct layer
of ice, only a few centimeters thick, that rests in contact with the bedrock floor
(Kamb and LaChapelle 1964).
The mechanics of regelation was first revealed by Bottomley (1872), who
demonstrated that a thin wire under tension could be passed through a block
of ice without splitting it apart. The pressure exerted by the wire melts the
ice beneath and the water thus released flows in a thin layer to the upper
it,
surface of the wire where it refreezes. The temperature of the ice is lower
beneath the wire than above it because the pressure depresses the melting
point. The heat of fusion released by refreezing above the wire is transferred
through the wire to provide the heat needed for the pressure-melting along
the leading edge, and so the speed at which the wire moves through the ice
block is partly dependent on the rate at which the wire conducts heat. Other
objects such as cubes, spheres, and discs have been forced through ice (Kamb
and LaChapelle 1964; Barnes and Tabor 1966; Townsend and Vickery 1967;
Morris 1976), and wires of various size and composition have been utilized to
study the regelation process (Nunn and Rowell 1967; Drake and Shreve 1973).
Each experiment has reinforced the correctness of Bottomley's original inter-
pretation of the process, although simple regelation theory seems to predict

much greater velocities than those actually measured (Morris 1976). Thus, it
seems that any object can be passed through ice without severing the mass
and without changing the properties of the ice except along the path of trans-
port. Along that path, the ice develops a new texture, similar in all aspsects
to that observed in the thin basal layers of glaciers.
Chapter 9 348
Direction of
Figure 9.9.
glacier movement
The process of regelation as it
functions beneath a glacier Melting

Melting occurs on upstream flank Bedrock Refreezing
of obstruction where pressure is obstruction
greatest Refreezing occurs
downstream of obstruction where
pressure is least,although some
cavitation may form
In glaciers the process of regelation, shown in figure 9.9, allows ice to

circumvent minor irregularities of the bedrock surface over which the glacier
is moving. The basal ice melts at the upstream edge of the bedrock knobs
where pressure is the greatest. The meltwater thus released flows around the
obstacle and refreezes in the region of least pressure, which is along the down-
stream edge of the barrier. The mechanism seems to be quite effective when
the obstacles are small but becomes less important when protuberances be-
come larger (Weertman 1957, 1964). This, of course, begs the question as to
what magnitude of bedrock irregularity will render the regelation process in-
operative and how thick the layer of regelating ice can be.
The problem is complicated by the fact that ice near the glacier base also
may deform according to the flow laws explained earlier. In that case, ice at
the upglacier edge of the obstacle, where pressure is greatest, will have a higher
strain rate (enhanced creep) than that at the downstream edge. Larger ob-
stacles will augment the stress and cause higher flow velocities; i.e., the ve-
locity will be directly proportional to the size of the bedrock knob. Because of
these complications, Weertman (1957, 1964) suggested that sliding over an
irregular bed is produced by components of the two processes. One (regela-
tion) is the predominent control on sliding when the barriers are small, and
the other (enhanced creep) when the obstacles are large. It follows that some
intermediate size, called the controlling obstacle size, determines which pro-
cess will function. In the subglacial bedrock topography, ice moving over or
around protuberances smaller than the controlling obstacle size will do so
mainly by regelation; where surface bulges are larger, creep will probably
dominate.
Finally, we now know that other factors have a direct influence on glacial
sliding. The most important of these is the type and character of the substrate
material. Boulton (1979) points out that the shear strength of this material
in an unconsolidated state is
S = C + (P - />„.) tan </>,
where 5 is shear strength, P is pressure exerted by the overlying ice, PK is pore

pressure, and <p is the angle of internal friction. You should recognize this
equation as being the same as the strength factor used when we examined
K
Figure 9.10.
^âtionzone- Hypothetical diagram showing
why the velocity of glacier flow
reaches a maximum at the
equilibrium line If no change in
cross-section area (width X

thickness) is allowed, increasing
discharge down-ice requires the
increase in velocity shown by
arrows.
Q- A + B + C
Q=A+B+C+D
slope stability in chapter 4. Investigators now realize that unconsolidated sed-

iment beneath the ice may be highly deformable, especially when its strength
islowered by having a high porewater pressure. Deformation of the substrate
may actually move the overlying ice with the mobilized material. For example,
Boulton (1979) estimates that in some cases as much as 90 percent of basal
ice movement is caused by deformation of unlithified substrate. The mobili-
zation of subglacial debris is an important factor in the development of glacial
features that originate beneath the ice, and we will return to this phenomenon
in the next chapter.
In addition to unconsolidated substrate, the sliding velocities of temperate
glaciers moving over carbonate rocks are affected by chemical reactions within
the subglacial meltwater (Hallet 1976a, 1976b). This is especially true in the
case of regelation because solutes in the water become concentrated on the
lee side of obstacles. This lowers the freezing temperature there and inhibits
heat transport away from those zones, a factor needed for the process of re-
gelation sliding to work efficiently.
Velocity and Flow

We can now ask whether measured glacial velocities agree at all with the me-
chanics of ice motion discussed above. For the purpose of developing a theo-
retical model, let us first assume that an active glacier remains unchanged in
its total and local dimensions over an entire budget year; i.e., the total accu-
mulation equals the total ablation, and all cross sections of the glacier are
equal in area and remain constant in area during the year. To maintain its
area, each successive downglacier cross section must transport from its lower
boundary exactly the amount of ice and snow delivered to its upstream
boundary. As figure 9.10 shows, in the accumulation zone, the cross section
at the highest elevation has only a small area of accumulation above it and so
must discharge only volume of ice, equivalent to the snow accumulated
a small
in that restricted surface area. Each section farther down-ice, however, must
Chapter 9 350
transfer a progressively larger volume of ice, since it moves not only the ice-
equivalent of accumulation on its surface but also the cumulative volumes of
all the higher sections. Without a change in cross-sectional area, the velocity
of flow must increase to a maximum at the equilibrium line. This follows be-
cause discharge is increasing to that level and because glacier discharge, like
river discharge, is equal to the area times the velocity. In the ablation area,
we can similarly expect a gradual decrease in velocity from the equilibrium
line to the terminus.
The model assumes that glaciers strive for and maintain some type of
Figure 9.11.
equilibrium and that the ice movement obeys the power flow law in some form.
Velocity distribution in longitudinal Mathematical treatment of even this simple glacial model is quite complicated
section ot glacier Surface velocity unless we further assume, as did Nye (1952a), that the motion is two-dimen-
( Vs ) is sum total of flow rates at
every level within the ice Strain
sional, plastic, and laminar such that the lines of flow are parallel to the bed
rate due to shear stress is highest and surface at all places (fig. 9.1 1). Accepting these conditions, the shear stress
at base of ice and is shown as the at the base of a glacier, measured along the central longitudinal axis and per-
basal velocity (Vb ).
pendicular to the surface, is given as
rh = pgh sin a,
where r b is the shear stress at the glacier base, p the ice density, g the accel-
eration of gravity, h the ice thickness, and a the slope of the surface.
By assimilating Glen's flow law into the above equation, the velocity at
any depth along the central axis (xy axis) can be estimated by assuming that
shear stress is proportional to depth, and that the strain rate directly relates
to that stress. Theoretically, then, the internal velocity profile of a glacier in
a longitudinal section should show a decreasing rate of flow from the surface
to the bedrock Even though the strain rate is greatest where shear stress
floor.
is the highest (at maximum thickness), the velocity increases from the base
to the surface because each internal layer not only moves in response to the
shear stress generated at that level but is also being carried on top of, and at
the speed of, the adjacent lower layer. Thus, the surface velocity (v s ) is the
sum total of strain rates for all the layers within the ice mass (fig. 9.1 1). The
equation also shows us that internal velocity is proportional to the product of
surface slope and ice thickness. The product seems to be fairly consistent,
meaning that where the ice is thin the surface gradient will be steep, and vice
versa.
Using the same basic approach, Nye (1957, 1965) predicted the velocity
distribution for an entire cross section, using different cross-sectional shapes
as models, and showed that velocity should decrease from the center of the
ice to the lateralboundaries both at the surface and at depth (fig. 9.12). Thus,
the simplest ideal model of a glacier, assuming continuity of discharge and
shape, and laminar flow, shows velocity decreasing with depth and distance
from the central axis. In the downstream direction, velocity should increase
toward the equilibrium line and decrease away from it.
The velocity of ice at a glacier surface is obtained by marking the position
of stakes driven into the ice relative to some nearby fixed point. This can be
done by normal surveying techniques or by photography repeated at some
specified time interval. Actual measurements of surface velocities substantiate
^s
Figure 9.12.
Velocity distribution in glacier
flowing a channel with a
in
parabolic shape. Diagram

represents one-half of the channel
cross section.Numbers represent
velocity expressed in any
common velocity units.
in a general way the theoretical predictions (Meier 1960; Meier et al. 1974),
although few measurements have been made in accumulation zones. In most
temperate valley glaciers, surface velocities range from 10 to 200 m a year,
but vary locally above or below these values. Outlet glaciers and ice streams
associated with ice sheets may also attain similar and, in some cases, even
higher velocities (Embleton and King, 1968; Flint 1971). For example, the ice
lobes around the periphery of the Pleistocene ice sheets of North America also
probably advanced at rates between 10 and 200 m a year (Ruhe 1975, p. 192).
As predicted, velocities tend to increase from the glacier head to a maximum
near the equilibrium line (Meier and Tangborn 1965) and to decrease from
there to the snout (Meier 1960; Meier et al. 1974). The transverse velocities
are usually greatest along the central axis, decreasing to the lateral margins
(Raymond 1971). In both patterns, however, enough variations occur to rec-
ognize that the flow model is infinitely more complex than our original as-
sumptions allow. In fact, even though flow may show a relatively simple
relationship to ice thickness and surface slope as predicted, in detail it may
not do so on a local scale; other factors such as variations in ice temperature,
basal sliding, and subglacial water pressure (Raymond 1971) may be in-
volved.
Measurements of englacial velocity require that a borehole be drilled
through the ice and cased to prevent its closure. The differential movement
with depth can then be calculated from an inclinometer, which measures the
angle between the axis of the borehole and the vertical. Almost every borehole
that has penetrated to a glacier floor or to great depth shows a velocity profile
similar to that predicted by theory (fig. 9.13). Velocity does not change sig-
nificantly with depth in the upper zones of most glaciers, although Meier (1960)
has demonstrated that some differential movement can occur even in the upper
meters of a glacier. In the lower half of most glaciers, the velocity decreases
more rapidly with depth.
Complications of the Simple Model

Real glaciers seem to possess the general traits that were predicted in idealized
flow, at least in the distribution of velocity. It is too much to expect, however,
that glacier characteristics will be predictable in detail. The initial assump-

tions are invalid on a local scale, and in fact, velocity itself is controlled by
several factors that are external to the system defined by the flow laws; we
know also that thickness and surface slope are free to change. We will briefly
examine some of the phenomena that may cloud the relationship between pre-
dicted and observed ice motion.
Chapter 9 352
Figure 9.13.
Internal velocity of two glaciers.
Horizontal displacement of
boreholes indicates the velocity.
(A) Saskatchewan Glacier (Meier
1960); (B) Malispma Glacier
(Sharp 1953)
-10
- 20
o
CD
"O
-30
-40
289.75
305.0
0.2 0.4 0.6 0.8 1.0 1.2 2 4 6 (feet)
0.5 1 1.5 (meters)
Horizontal deformation, (feet)
August 6, 1952, to August 5. 1954 Horizontal distance
Horizontal exaggeration 25X
(A) (B)
Extending and Compressive Flow Returning to our ideal model, we see that
allowing no change in cross-sectional area requires that the elevation of each
point along the surface of a balanced glacier be maintained. If this is to happen,
then the flow cannot consist of laminar horizons moving parallel to the surface
because the ice must move slightly downward in the accumulation zone and
slightly upward in the ablation zone. (Accumulation tends to elevate the sur-
face and ablation to lower it unless these responses are counterbalanced by
motion of the ice.) Thus, the pattern of flow should be as shown in figure 9.14
and not as laminar sheets.
Actually ice will tend to thicken in some places and to thin in others, pro- Zone of
accumulation
cesses called compressive and extending flow. Nye (1952a) first investigated
(+)
these flow types and suggested that rates of accumulation and ablation, as well Equilibrium
as changes in the slope of the underlying bedrock, will determine which type line
will prevail. Compressive flow results in a decrease of velocity and occurs where Zone of
ablation
the underlying rock surface is concave-up or where there is a consistent loss
of surficial ice. In contrast, extending flow has increased downglacier velocity
and exists where ice is added at the surface or the glacier bed is convex. At a
constant bedrock slope, therefore, accumulation zones should be character-
ized by extending flow and ablation zones by compressive flow. Within each
of these zones, however, topographic irregularities on the underlying surface
may produce local changes in the flow type. Figure 9.14.
Longitudinal flow lines in
This analysis generally fits the velocity distribution that we surmised in hypothetical glacier
our balanced glacier and observed in real glaciers. In ablation zones, with the
pervasive compressive flow, the velocity should decrease downglacier because
the ice is being compressed (Nye 1952a), and in accumulation zones, with
extending flow, the velocity should increase downglacier. We must ask, how-
ever, whether it fits the direction of flow predicted for a balanced glacier. Nye
(1952a) showed that extending and compressive flow should each generate a
pattern of potential slip planes that follow the orientation of maximum shear
stresses (fig. 9.15). The family of planes are such that their resolution at any
point will give the two directions of maximum shear; thus, they will be per-
pendicular and parallel to the bed of the glacier and form 45° angles with the Extending flow
surface of the ice. The potential slip planes show that in zones of extending (A)
flow (accumulation zones), the predominant downglacier slip direction will be

downward at the surface; while in compressive zones, such as ablation areas,
the low-angle downglacier slip will be upward.
This seems to coincide with the predicted mode of flow in our ideal glacier
(fig. 9.14), but in real glaciers the pattern is much more complicated. An ir-
regular bedrock profile will cause reversals of flow type in the ice (fig. 9.16),
and where measurements have been made for an entire glacier (Meier and
Compressive flow
Tangborn 1965), the flow pattern (fig. 9.17) is not the simple one of our ideal
(B)
case. Furthermore, more than one velocity maximum may occur, and the equi-
librium line does not necessarily mark a zone of maximum flow velocity. In
Figure 9.15.
fact, on the South Cascade Glacier (Washington), the equilibrium line is usu-
Potential slip planes under
ally near a zone of lower surface velocity, but the thickness is so great that
(A)extending and
the total discharge through the section is still very high. (B)compressive flow. The
preferred downglacier slip paths
will follow A- A' and P-P'.
Slip A second complication arises because theoretical estimates of velocity
do not consider the added component of basal sliding, the magnitude of which
is represented by the displacement of boreholes at the bed. Although few di-
rect measurements are available, they show that sliding may account for less
than 10 percent of the surface velocity in some glaciers (Engelhardt et al.
1978) and as much as 90 percent in others (McCall 1952; Vivian 1980). Even
worse, sliding velocities may be extremely variable within different portions
of the same glacier (Savage and Paterson 1963). Although, as discussed ear-
lier, the mechanics are not understood, it is clear that the influence of basal
Chapter 9 354
Figure 9.16.
Longitudinal section of
hypothetical glacier showing
irregular bedrock profile and
preferred slip planes within the
ice. Zones of extending flow and
compressive flow indicated by E
and C
Figure 9.17.
Longitudinal section showing 2000
average calculated bedrock
profile and surface velocity <5
vectors on the South Cascade a>
Glacier, Wash. ~E 1800

1961 Profile
o
1958 Profile
as
>
Lake
m 1600
1000 2000 3000 4000

(meters)
sliding must complicate the simple relationship between flow laws and ve-
locity, at least intemperate glaciers. Slip also can occur along the lateral mar-
gins of glaciers, and abnormally high velocities are sometimes encountered at
the glacier sides (Meier et al. 1974).
It is now apparent that our basic model of ice sliding over a bedrock sur-
faceis simplistic. As discussed earlier, a large proportion of forward glacial
movement may come from deformation of underlying sediments rather than

sliding or englacial creep. This may be especially true in deformable sediment
with low permeability. In those cases, water builds within the sediment rather
than in the subglacial zone and aids in the substrate deformation (Boulton
and Jones 1 979). In addition, Engelhardt and his colleagues (1978) have iden-
tified a thin (10 cm) layer of debris between the underlying bedrock surface
and the base of the Blue Glacier in Washington. This layer, which they called
the active subsole drift, is involved in decreasing the sliding velocity because
its surface is rough, its movement involves friction between the included grains,
and water enters and flows within the layer, thereby reducing hydrostatic pres-
sure at the glacier base.
Variations with Time The third factor complicating the relationship be-
tween flow theory and observed flow rates is that velocity is not constant in
any given On
most glaciers it varies significantly with time; to add to
glacier.
the confusion, the interval over which velocity variations occur is different
from one glacier to another and even in specific locations on the same glacier.
For example, many investigators have noted sudden, jerky motions that in-
crease surface velocity for hours, days, or weeks (Meier 1960; Glen and Lewis
1961; Goldthwait 1973; Jacobel 1982). These spastic movements, which may
involve a change in velocity of as muchhundred percent, seem to
as several
be a surface-ice phenomenon (Goldthwait 1973) and tend to be restricted in
extent. They are normally explained by local controlling factors such as weather
conditions, fault slips, or a sudden release of ice that has been retarded in its
flow by some obstruction.
Seasonal or yearly variations in flow also are common. Velocities in the
ablation zone usually increase at or near the end of the summer. These fluc-
tuations most logically reflect accelerated basal and side slip, which is facil-
itatedby the abundance of free water at that time (Meier 1 960; Elliston 1 963;
Paterson 1964; Hodge 1974; Vivian 1980; Anderson et al. 1982). Seasonal
velocities may deviate by as much as 80 percent from the mean annual values.
They are not restricted to temperate glaciers but also affect the movement of
subpolar glaciers (Friese-Greene and Pert 1965). Yearly variations probably
represent internal adjustment to accumulation rates, but they also can be in-
fluenced by basal meltwater lubrication (Meier et al. 1974).
Short-term variations in velocity are usually recognized because the span
of most investigations is long enough to demonstrate their presence. However,
imagine the sampling program needed to observe and document long-term

fluctuations of velocity that are produced by changes in climate and the as-
sociated glacial regime. Any positive change in a budget should increase the
discharge of flow, and according to theory, the mechanical response to the in-
creased accumulation will probably involve the propagation of a kinematic
wave moving down the glacier two to five times faster than the actual particles
of ice. The wave will reach the glacier snout long before the new ice formed
from the accumulation could possibly be transported that far. Without
flux in
going into detail, it appears that the wave motion will have little effect on ice
that is extending; but in the ablation zone, where compressive flow dominates
and velocity decreases down-ice, the oncoming wave accentuates the compres-
sion and the glacier becomes very unstable there. As the wave approaches any
point in the ablation zone, the ice will thicken and the surface will rise dra-
matically. As the wave passes, it will quickly thin and subside to its former
level. On some glaciers the ice surface may rise and fall more than 100 m in
the sequence of events.

Chapter 9 356
Velocity measurements on some glaciers do not fit the predictions based

on kinematic wave theory (Lliboutry and Reynaud 1981). Nonetheless, if ki-
nematic waves generally represent the mechanism by which glaciers respond
to changes in mass balance (Nye 1960), it becomes important to recognize
that the time needed for the wave transmission varies from glacier to glacier.
For example, the time needed to reestablish a steady state (i.e., the response
time) in single, temperate glaciers varies from 3 to 30 years. The significance
of this fact can be seen in a hypothetical case. Suppose that during the years
1954-1958, abnormally high amounts of snow accumulated on two adjacent
valley glaciers.. In one with a short response time, a kinematic wave rapidly
traveled the length of the glacier, and the snout advanced dramatically in 1960.
The second glacier, having a longer response time, showed no visible effects
of the accumulation by 1960; the kinematic wave did not reach the terminal
zone until 1975, when the terminus suddenly advanced. The two glaciers were
completely out of phase, although both advanced in response to the same event.
In addition to problems of mechanics, measurements made on a portion of the
ice experiencing the wave action are not documenting the flow of ice but rather
the movement of waves.
where response times are much longer, the difficulty of re-
In ice sheets,
lating flow to a change in regime is magnified. For example, Whillans (1978)
suggests that the modern thinning and accelerated flow in the central portion
of the West Antarctica Ice Sheet was initiated by warming that commenced
10,000 years ago. Thinning in the interior of ice sheets requires transfer of ice
to the margins, thereby thickening the marginal ice and promoting its forward
movement. Thus, it is possible that the rapid advances often noted during the
final phases of Pleistocene glaciations may have been in response to warming
rather than cooling climates.

Kinematic wave velocities may attain catastrophic values, reaching
hundreds of meters a day (Meier and Johnson 1962; A. E. Harrison 1964);
they have often been cited as the mechanics governing surging glaciers. How-
ever, in many cases glaciers surge even though there is no discernible net ac-
cumulation of mass to the glacier or any other external stimuli for the
movement (Post 1960, 1969; Meier and Post 1969). It is probably best to sep-
arate surging from kinematic wave transfer even though the two movements
have many of the same characteristics (Palmer 1972).
A surging glacier, then, is one in which sudden, brief, large-scale ice dis-
placements periodically occur. The moves 10 to 100 times faster than its
ice
flow rate in the quiescent periods between surges resulting in the fact that
surging glaciers entrain significantly more debris than normal glaciers (Clap-
perton 1975). The surge periodicity seems to range from 15 to more than 100
years (Post 1969) and probably results from unique conditions that create a
Some surges have no distinct periodicity
cyclic instability within the glacier.
and may be generated by outside events such as earthquakes or episodes of
abnormally high geothermal heat.
Glacial surges are now recognized as fairly common phenomena, with 204
surging glaciers identified in western North America alone by 1969 (Post
1969). In fact, Budd (1975) has suggested that they may represent a com-
pletely separate type of glacier. However, they have no distinct size, shape, or
Pi
and they can contain temperate or subpolar ice. Their only unifying
activity,
seem to be that surges are initiated in the ablation zone slightly

characteristics
down-ice from the equilibrium line, and a long and pronounced stagnation of
iceoccurs in the terminal zone in the interval between surges (Post 1960, 1966).
During the surge the glacier surface is broken chaotically, and medial mo-
raines and ice bands are intensely contorted.
Apparently, for some unknown reason, a reservoir of ice collects in the
upper portion of the ablation zone, where ice is still active, while the terminal
area is physically lowered by ablation. Eventually a critical disequilibrium
condition is reached, and the response is a sudden burst of glacial movement
into, and often overriding, the stagnated terminal zone. The key to under-
standing cyclic surging thus seems to lie in two factors: (1) how thickening of
ice, with a concomitant increase in basal stress, can occur in one region of the
glacier while stagnation occurs in another; and (2) what triggers the sudden
release of the ice reservoir.
Several models have been proposed to explain the internal controls of pe-
riodically surging glaciers. Budd (1975) suggests that certain glaciers with
the proper sliding velocity gradually develop a lubricating factor that in turn
lowers the basal stress enough to induce a sudden movement. Surging glaciers,
in his model, do not have sufficient mass flux to continue this rapid flow, and
so they periodically oscillate from the slow flow of an ordinary glacier to sudden
bursts of activity.
Another interesting model (Robin and Weertman 1973) proposes that
surges end when the bed shear stress reaches a low value, a conclusion pos-
tulated first by Meier and Post (1969). During the quiet period following the
surge, the glacier develops zones of vastly different basal shear stress. In the
stagnating terminal area, basal shear will be relatively low. Further up-ice,
however, the stress will progressively increase with time along with the thick-
ening of the ice. Between the two regions is an area called the trigger zone,
where the gradient of basal shear stress is continuously increasing and even-
tually reaches a critical, unstable condition. Robin and Weertman also suggest
that a distinct pressure distribution, related to the basal stress is promoted in
the water at the bed. The water-pressure gradient is inversely related to the
shear-stress gradient, and water is dammed in the trigger zone. The next surge
begins when the trapped water lowers frictional resistance so much that the
sliding velocity drastically increases.
One objection to this model, clearly recognized by Robin and Weertman,
is that the basal water to be dammed must be prevented from draining down-
glacier through interconnected channels. This objection might be partially
overcome if the glacier snout is frozen to the bedrock, a possibility suggested
for subpolar surging glaciers (Jarvis and Clarke 1975). Even in temperate
glaciers, patches of cold ice may (Robin 1976).
exist at the glacier floor
To summarize, it has been shown theoretically that surges can be trig-
gered by lowering basal shear stress over a period of time (Campbell and Ras-
mussen 969); presumably, the lubricating effect of meltwater can accomplish
1
the necessary decrease in stress. If that is true, as the proposed models suggest,
the key to surging lies in the mechanics of basal sliding.
Chapter 9 358
tee Structures Glaciers usually display a variety of structures that develop during growth of
the glacial mass —which we can call primary structures — as well as secondary
structures that develop in response to glacial movement. Primary structures
are indicative of accumulation and ablation characteristics during the gla-
cier's history, and secondary structures relate closely to the glacier's mode of
flow and stress field.
Stratification
Primary structures in glaciers appear as discernible layers or bands within the
ice. The layering results from processes that reflect an annual cycle of snow-
accumulation and ablation above the firn line. During the winter a thick pile
of new snow is added to the glacier and, with time, proceeds through the phases
of metamorphism that culminate in a layer of clear white ice. In the ablation
season, however, the upper portion of the winter accumulation is subjected to
melting and refreezing and develops a texture different from that forming lower
in the snow pack. In addition, sediment and organic debris may collect in the
partially ablated upper zone, making it slightly darker in color. The alter-
nating white and "dirty" layers give the ice a stratified appearance and allow
Figure 9.18.
Saskatchewan Glacier — outcrop glaciologists to estimate the annual growth in the glacier thickness. The layers
pattern of stratification in
tend to be tilted and deformed as the glacier moves. At the surface of the
Castleguard sector. View
ablation zone, the bands look like truncated beds of plunging folds (fig. 9.18)
upglacier from cliff on south
margin below Castleguard Pass with the layers dipping into the glacier in an up-ice direction.
Splaying and en echelon
crevasses are also visible.
Province of Alberta, Canada.
^M
Secondary Features
Foliation A secondary type of layering, called foliation, is produced by shear

during ice motion; it is sometimes difficult to distinguish from primary strat-
ification because both types may display similar grain size or textures. Fo-
liation usually appears as alternating bands of clear blue ice and white bubble-
rich ice. The layers, which dip at all angles, are most prevalent near the ice
margins and commonly offset or wrinkle the primary stratification (fig. 9.19).
The origin of foliation is poorly understood. Some evidence suggests that it
originateswhere shear stress is greatest, but Meier (1960) argues that it is

neither formed nor preserved at great depth, where shear stress would pre-
sumably be at a maximum.
Crevasses Crevasses are cracks in the ice surface that range in size from
miniature fractures to gaps several meters wide. The fractures are important
in that they provide avenues for surface meltwater to penetrate the interior
of the glacier, although the openings are rarely deeper than 30 m. It is gen-
erally assumed that crevasses develop perpendicular to the direction of max-
imum elongation of the ice. On some glaciers, however, the crack directions
Figure 9.19.
do not precisely coincide with the measured surface strain rates (Meier 1960).
Nonetheless, crevasse types (fig. 9.20) do reflect a local tensional stress en-
Saskatchewan Glacier — gently
dipping stratification, wrinkled and
vironment. Splay crevasses or radial crevasses form near the flow centerline intersected by nearly vertical
foliationExposed on east wall of
a crevasse, 4 5 km below firn limit
in midglacier Province of Alberta,
Canada
Chapter 9 360
under compressive flow where spreading exerts a component of lateral exten-

sion (fig. 9.3). In contrast, near the ice margins the shear stress parallels the
valley walls, and crevasses develop diagonally to those sides. Here the cre-
vasses are either chevron or en echelon types.
Transverse crevasses (fig. 9.21) develop under extending flow where the
ice extends in a longitudinal direction. In temperate glaciers, these cracks occur
most commonly where the glacier cascades over convex irregular-
in ice falls
ities in the underlying bedrock slope. In the summer, large ice crystals may-
develop in the transverse cracks when water trapped there repeatedly melts
and refreezes. Sediment may also accumulate in these openings. At the base
of the ice fall the crevasses are closed by compression, and a band of dirt-
stained ice forms. During the winter, however, ice descending the fall recon-
stitutes into clear, bubbly ice (King and Lewis 1961). The annual downvalley
Figure 9.20.
Types of crevasses in valley
glaciers: (A) marginal or
chevron —
(1)old rotated
crevasses, (2) newly formed
crevasses; (B) transverse;
(C) splaying; (D) radial splaying
Arrows show flow direction.
Figure 9.21.
Transverse crevasses developed
at ice fall on Little Yanert Glacier,
Alaska Range.
flow, therefore, tends to produce a distinct series of alternating white and dark
bands, called ogives, that are prominent at the foot of ice falls (fig. 9.3). They
are different from primary layering in that they do not dip as strata into the
glacier but are purely a surface phenomenon.
In this chapter we examined the origin and movement of glaciers, as well as Summary
the factors that might explain the properties associated with the various gla-
cier types. Glaciers can be classified on the basis of their morphology or geo-
graphic position, their activity, or their thermal characteristics. Any glacier
develops when snow accumulated over a period of years is transformed into
ice by compaction, recrystallization, and melting and refreezing. These pro-
cesses progressively increase the density of snow and firn as the space within
the original mass is removed by pressure, crystal growth, and orientation of
the grains. When the ice reaches a critical thickness, it is capable of move-
ment.
The amount and type of geomorphic work a glacier can accomplish de-
pends on mass balance and on the total volume of ice added and lost during
its
a period of time. The mass balance, or budget, is the net difference between
accumulation and ablation during the time period in question. A positive mass
balance indicates that more ice has been added to the glacier than has been
lost; and the transfer of ice from the accumulation zone to the ablation zone
causes the glacier front to advance. A negative budget results in a retreat of

the front because the amount of ice transferred from the accumulation zone
to the ablation zone is not sufficient to replenish the volume of ice lost by
ablation. The volumes of accumulation and of ablation determine the
total
level Two glaciers may have identical mass balances, but
of glacial activity.
the one having the higher total volumes of addition and loss will be more active
and do more geomorphic work.
Glaciers move by internal deformation of the ice and by sliding along the
bedrock floor at the base of the glacier. The internal movement occurs as a
type of creep that is mechanically different from viscous flow or pure plastic
flow. In most glaciers the deformation caused by flow can be estimated by an
equation called the power flow law or some modification of it. The mechanics
of sliding is poorly understood but probably consists of the combined effects
of regelation, lubrication by water under pressure, and enhanced creep.
Velocity of flow relates well with modified versions of the power flow law,
but many characteristics of glaciers complicate a direct relationship. For ex-
ample, ice does not flow in parallel laminar sheets but thickens and thins along
the length of the glacier. This develops the compressive and extending flow
that causes ice to move downward in the accumulation zone and surfaceward
in the ablation zone. In addition, flow velocity varies with time. The phenom-
enon of surging is difficult to fit into an all-inclusive flow model.
Primary and secondary structures in glaciers are related to the processes
of accumulation and ablation and to the stress field generated during the ice
movement. The major structural features are stratification, foliation, cre-
vasses, and ogives.

Chapter 9 362
Embleton, C, and King, C. A. M. 1968. Glacial and periglacial

geomorphology. Edinburgh: Edward Arnold Ltd.
Engelhardt, H. F.; Harrison, W. D.; and Kamb, B. 1978. Basal sliding and
conditions at the glacier bed as revealed by bore-hole photography.
Jour. Glaciol. 20:469-508.
Glen, J. W. 1955. The creep of polycrystalline ice. Proc. Royal Soc.
London, ser. A: 228:519-38.
Hutter, K. 1982. Glacier flow. Am. Scientist 70:26-34.
Meier, M. F. 1960. Mode of flow of Saskatchewan glacier, Alberta, Canada.
U.S. Geol. Survey Prof. Paper 351.
Meier, M. F., and Post, A. S. 1969. What are glacier surges? Can. J. Earth
Sci. 6:807-17.
Nye, J. F. 1965. The flow of a glacier in a channel of rectangular, elliptic,
or parabolic cross-section. Jour. Glaciol. 5:661-90.
Sugden, D. E., and John, B. S. 1976. Glaciers and landscape. London:
Edward Arnold Ltd.
Glacial Erosion, Deposition,
and Landforms
10
I. Introduction 1. Moraines
II. Erosional Processes and Features 2. Stratified Marginal
A. Minor Subglacial Features Features
B. Cirques a. Kames and Kettles
1. Cirque Glaciers b. Eskers
2. Headwall Erosion D. Interior Ice-Contact
C. Glacial Troughs Features
III. Deposits and Depositional 1. Ground Moraine
Features 2. Fluted Surfaces and
A. Drift Types Drumlins
1. Nonstratified Drift E. Proglacial Features
2. Stratified Drift IV. Summary
B. The Depositional V. Suggested Readings
Framework
C. Marginal Ice-Contact
Features
363
Chapter 10 364
Introduction We are now ready to consider the geomorphic significance of the glacier me-
chanics we examined in chapter 9. It would be ideal if we could directly cor-
relate geomorphic features with specific ice processes, but we seem to be far
from such sophistication. The reasons for this involve several main factors:
( 1 ) Our conceptual models of glacier mechanics are greatly oversimplified and
based on theory rather than extensive observation. (2) Much erosion and de-
position take place at the base of glaciers where it is rarely feasible to docu-
ment the actualprocess. (3) We are just beginning to understand the feedback
mechanics that function between the basal ice and the geologic framework.
For example, we know that glaciers have the ability to erode because we can
see the results of that erosion after the ice disappears. What we don't under-
stand is precisely how the erosive process itself is modified by remolding of
the subglacial framework. Does removal of certain obstructions by erosion
decrease the subsurface roughness enough to facilitate rapid sliding velocities,
as suggested by Kamb (1970), and might this in turn accelerate further ero-
sion?
Until we can resolve these deficiencies in our understanding, the process-
feature relationship will be arrived at by traveling a one-way street in the
wrong direction — we are interpreting the process by the character of its re-
sults. Nonetheless, glaciers have left us a variety of deposits and myriad fea-
tures that demand our attention. It is not completely satisfying to reconstruct
their origin without a wealth of solid observations, but this type of deductive
reasoning has been a part of geology for a long time and is not necessarily
incorrect. Furthermore, investigations of features and deposits can provide
tangible clues about their origins that become invaluable in deciding how and
where to study glaciers in the future. It seems inevitable, however, that new
techniques for investigating glacier mechanics will demand continuous mod-
ification and testing of our present interpretations of features and deposits.
Erosional Processes Minor Subglacial Features

and Features Glacial erosion is accomplished primarily by two processes, a scraping action
called abrasion and a dislodgement or lifting action called quarrying or
plucking. Since ice is not a hard mineral ( .5 on Moh's scale at 0°C), it cannot
1
abrade most solid rock material unless it utilizes as grinding tools the frag-
ments of rock carried in its load. Therefore, the efficiency of abrasion and the
features it produces depend on the character and concentration of the debris
being dragged along the base of the ice and, of course, on the properties of
the bedrock being overridden. Abrasion is also influenced in a complex way
by the nature of the subglacial topography and the velocity and direction of
ice flow. The rate of abrasion, estimated by various methods, usually ranges
from 0.06 to 5 mm a year, but it may be considerably higher under thick, high-
velocity glaciers. Boulton (1974) reports that the abrasion of a marble plate
inserted beneath the Glaciere d' Argentiere in France proceeded at a rate up
to 36 mm a year where the ice was 100 m thick and moving at 250 m a year
365 Glacial Erosion, Deposition, and Land forms
Figure 10.1.
Theoretical abrasion rates plotted
against effective normal pressure
for different ice velocities. The
rate increases zone A and
in
decreases in zone B as pressure

increases. At any given velocity, a
pressure increase greater than
that at the higher X-axis intercept
causes deposition of lodgement
till, k, c, and p are constants
depending on relative hardness,

debris concentration, and
penetration hardness of the ice.
20 30 40 50
Effective normal pressure (bars)
The apparent dependence of abrasion on ice thickness does not agree with
theoretical analyses (McCall 1960; Hallet 1979). The reason for this is that
the yield stress for ice (2 kg/cm 2 at 0°C) suggests that thickness greater than
22 m should cause ice to flow around and under rock fragments without in-
creasing its ability to abrade the underlying bedrock surface. We know, how-
ever, that basal cavities and extensive contact between loose particles and the
bed do exist beneath thick glaciers, indicating that the ice can maintain some
strength. Therefore, Boulton (1974) suggested that at any given ice velocity,
assuming a constant concentration of basal debris, the abrasion rate will in-
crease with increasing normal pressure until it attains a peak value, as plotted
in figure 10.1. Further increase in normal stress will cause the rate to decrease
rapidly to zero, where abrasion ceases and any debris in transport will be de-
posited beneath the ice. The form of the curves shown in figure 10.1 have been
verified by direct measurements (Boulton 1979).
Small erosional features produced by abrasion are usually in the form of
linear scratches or crescentic marks that show the relationship between the
size and composition of the abrading particles and the resistance of the un-
derlying bed. Very fine particles in sufficient abundance produce a smoothly
polished surface composed of microscopic scratches; as the grain size of the
load increases, the scratches become larger in a transitional sequence from
polish to striations to grooves or furrows.
Striations like thoseshown in figure 10.2 have been noted in every gla-
ciated region of the world.They are best preserved on fine-grained rocks that
have not been deeply weathered and on smooth bedding planes that dip gently
away from the direction of ice movement. Striations and larger linear features
also can form in unconsolidated material, such as till or loess, if it is highly
compacted (Westgate 1968). Striae are only millimeters deep and are most
likely eroded by sand grains or by jutting edges of larger particles carried in
Chapter 10 366
Figure 10.2.
Striationson outcrop of limestone
in south central New York State.
the basal ice. Striations tend to be continuous for only relatively short dis-
tances, probably because the sliding clast itself produces a carpet of plowed
debris at its forward edge. When sufficient debris accumulates, the scratching
particle will ride over the material, thereby interrupting its contact with the
solid bedrock until it reaches a fresh surface downstream from the debris cover
(Boulton 1974). This affect, however, can be eliminated if circulating sub-
glacial water removes the fines (Vivian 1970). It is also known that the entire
scratching phenomenon will become ineffective unless particles within the ice
mass continuously move downward to replace the original abrasive grains.
Those grains become smoothed during the abrasive action, and failure to re-
plenish the basal ice with new particles having sharp edges and corners will
cause abrasion to end (Boulton 1979).
367 Glacial Erosion, Deposition, and Landforms
Increasing grain size in the load produces grooves and furrows larger than
Normally grooves are up to 1-2 m deep and 50-1 10 m long, but
striations.
under the proper controlling factors they may achieve giant proportions. For
example, H. T. U. Smith (1948) describes grooves in the Mackenzie River
valley ofCanada that are 30 m deep, 100 m wide, and several kilometers long.
These are not the product of a single boulder but possibly represent gouging
by a pocket of boulders solidly frozen together (Embleton and King 1975a).
There may, in fact, be a limit to the size of boulder that can act in the grooving
process, because large fragments, having too much surface contact with the
underlying rock, will force the ice to flow over and around the boulder instead
of carrying it as part of the basal load.
In addition to scratches of all types, a group of small features, generally
referred to as friction cracks and/or chattermarks, are formed by chipping
or grinding of the underlying rock surface (Harris 1943). Most workers be-
lieve these features result when ice flow is temporarily retarded in its forward
motion and then suddenly released. This produces a jerky flow component
commonly referred to as a slip-stick movement. The various cracks and marks
are usually lunate in form, 10-12 cm long and 10-25 mm
deep, and perpen-
dicular to the direction of ice flow as determined by other criteria. Chatter-
marks are sometimes present on surfaces of minerals within glacial deposits
(Folk 1975; Gravenor 1982). These are indicative of the grinding action within
a glacier, even though chemical etching may produce similar features.
The features of abrasion are thought to reflect the direction of ice move-
ment. Remember, however, that other processes such as mudflows or snow-
slides can form striations, and even floating ice blocks can cause them in
nonresistant materials (Dionne 1974). Ice will diverge and converge over an
irregular bedrock topography and basal ice may be moving in different direc-
tions than surficial ice (Engelhardt et al. 1978). Thus, minor abrasion features
have somewhat limited value as indicators unless a large number of mea-
surements are obtained and treated for statistical significance.
Quarrying differs from abrasion in that the functional success of the pro-
cess depends less on the type of load being transported than on the properties
of the underlying rock. In fact, fractures must exist in the bedrock if plucking
is to operate at all. Intense shattering of rock in preparation for plucking prob-

ably requires some form of pressure release (Lewis 1954; Glen and Lewis 1961),
crushing (Boulton 1974), or cyclic freezing and thawing. These processes,
which weaken the internal cohesion of the bedrock, may occur subglacially
(Sugden and John 1976; Anderson et al. 1982) or in association with perigla-
cial conditions prior to the arrival of the glacier.
In detail, plucking has two basic requirements: first, the ice must exert a
shear force on the loosened particles, and second, this force must exceed the
resistance caused by friction when the particle is dragged over the residual
bedrock (Boulton 1974). It is not enough simply to dislodge the particle from
its original position, but the driving force must overcome any frictional resis-
tance generated in the system. Boulton suggests that where ice flows over un-
consolidated sediments, plucking is a relatively straightforward process related
totally to the shear stress.
Chapter 10 368
In tightly lithified materials, the mechanics are more complex. The pro-
cess is directly influenced by periodic opening and closing of cavities beneath
the ice, which are caused by obstructions and fluctuations of ice thickness or
velocity. When cavities are open, free water may freeze to the fragmented
particles or within the shattered mass. If the cavity is subsequently closed, the
new ice into
glacier incorporates this its basal layer; its forward motion will
pluck some rock fragments away from the surface. Thus, plastic flow may also
be involvedin the encasement of shattered material and its subsequent re-
moval (Boulton 1974). It is important to recognize that frictional resistance

to plucking increases rapidly as the cavity is closed because it is directly pro-
portional to the normal pressure. As Boulton suggests, plucking is probably
most effective when the normal pressure is sufficient to incorporate the loose
particles into the ice but not great enough to inhibit their forward movement
by increasing the frictional resistance. Exactly when this condition prevails
will vary from glacier to glacier, and other factors may complicate the dy-
namics. For example, rock fragments projecting downward from the ice may
dislodge particles from the rock floor even when the interface is slightly open.
These do not necessarily become encased in the ice but are shoved spasmod-
ically along the bedrock surface until they encounter the downglacier end of
the cavity.
The evidence of plucking action in glaciated landscapes is usually found
in erosional features somewhat larger than those produced by abrasion, but
commonly the two erosive processes are closely associated in the same land-
forms. Where abrasion is dominant, the landscape may be indented with
smoothly curved elongate surfaces whose long axes are subparallel to the di-
rection of ice flow. Some of these surfaces are distinctly higher at one end.
and they taper laterally and longitudinally until they blend into the sur-
rounding ground level, producing a unique teardrop or raindrop shape which
Flint (1971) describes as a "whaleback form." Some whaleback forms may
be related to streamlined depositional features, such as drumlins, in that their
shape represents the minimum resistance to flowing ice. The composition of
streamlined forms seems to be of little consequence, however, as they can be
entirely bedrock, entirely sediment, or any combination of the two. Whale-
backs, therefore, may be merely a transitional form in a range of streamlined
features from pure bedrock to pure drift (Flint 1971).
When plucking is a significant factor, whalebacks develop a pronounced
asymmetry, having a gently sloping upstream surface and a steep rock face
on the down-ice side of the feature. Such a form, commonly called roche mou-
tonnee, is the result of abrasion on the upstream slope and intense quarrying
at the position of the steep, downstream face. Its development is due to irreg-
ular spacings of fractures within the bedrock. Where joints are widely spaced,
abrasion is the dominant process, while closely spaced jointing facilities
plucking and more rapid erosion (fig. 10.3). Flint (1971) objects to the term
"roche moutonnee" because of its wide misuse and suggests that stoss and lee
topography better describes the forms developed by the combination of abra-
sion and plucking.
Figure 10.3.
Relationship between joint
spacing and roche moutonnee
development in Yosemite Valley
(After Matthes 1930)
\/\ /\
2500
Cirque lip Figure 10.4.
and moraines Headwall Long profile through Blea Water
2000 ~ come, a cirque.
Blea Water
1500 —
I
Cirque floor
-1000
3000 4000 5000 6000
(Feet)
Subglacial erosion is much more complex than our simple analysis of

abrasion and plucking suggests. Many subglacial processes are involved in
fracturing «f the underlying bedrock (Gray 1982), and these directly affect
the intensity of plucking and the character of ice-scoured topography (Gordon
1981). Flow of basal ice around obstacles often results in extreme velocities
known as ice streaming, which produces large, elongate erosional forms
(Boulton 1979; Goldthwait 1979). Water may also be involved in subglacial
erosion. For example, in carbonate regions subglacial meltwater may differ-
bedrock surface (Hallet 1976a, 1976b). In addition, rapid
entially dissolve the
drainage of meltwater through subglacial channels can physically erode the
bedrock (Dahl 1965), as can the fluid mobilization of subglacial till (Gjessing
1967; Gray 1982).
Cirques
The striking landscapes found in the uplands of glaciated mountains have been
sculptured primarily by the erosive action of ice contained in cirques. The
term "cirque" was first used in the early 1 800s to describe the collecting basins
for valley glaciers in the Pyrenees, and been given a
locally the feature has
variety of names including cwm, and botn. A cirque by any name
corrie, kar,
is still a deep erosional recess with steep and shattered walls that is usually
located at the head of a mountain valley. It is normally semicircular in plan

view, often being described as an amphitheater, and it is floored by a distinct
rock basin where the surface has been smoothed by abrasion. As figure 10.4
shows, the bowl-shaped rock basins commonly contain lakes, called tarns, that
are dammed by a convex-up rock lip that stands as a threshold boundary
in
between the cirque floor and the downstream part of the valley. The cirque lip
is often capped by small moraines that contribute to the damming effect. The
rock basins can be of spectacular dimensions. For example, the rock basin floor
of Blea Water corrie in England (shown in the figure) is 96 m (316 ft) lower
than the rock lip (Lewis 1960).
Chapter 10 370
Cirques range in size from shallow depressions to monstrous cavities that

are kilometers wide and several thousand meters high along the rear wall.
Their dimensions and their geomorphic form depend not only on the rocks into
which they are cut, being larger and more perfectly developed in igneous or
high-rank metamorphic rocks, but also on the rock structures (Olyphant 1981),
the preglacial relief, and the time span of the formative glaciation. Most ma-
turely developed cirques seem to possess a reasonably consistent geometry when
their length to height ratios are compared, indicating that cirques probably
attain some equilibrium form related to the processes of their formation.
Cirques are often preferentially oriented according to the direction of solar
radiation and the prevailing winds (Graf 1 976), and their elevation is probably
(but not necessarily) related to the snowline at the time of their formation
(Porter 1977; Trenhaile 1977). Thus, although most cirques originate in the
headward reaches of stream valleys, any hollow, regardless of its origin, that
stands at the proper elevation and has the ideal orientation may progressively
accumulate snow and finally become a maturely developed cirque like those
in the photograph in figure 10.5.
The significance of cirque processes in the development of Alpine scenery

is that cirque expansion by continued erosion gradually eliminates the pre-
glacial upland surface. As a number of cirques grow headwardly and laterally,
they progressively consume much of the intervening upland region and leave
Figure 10.5.
as its only vestiges spectacular horns and aretes, the features so indicative of
Cirques in the Alaska Range near
Mt McKinley, Alaska. Knife-edged mountain glaciation (figs. 10.5, 10.6). With prolonged headward erosion, ad-
ridges between cirques are jacent cirques may merge, forming col depressions in the knife-edged aretes.
aretes
r-i >*>
Figure 10.6.
Schwan Glacier in the Chugach
Mountains of Alaska Note dark
lobes on glacier surface where
rockslide avalanches have moved
onto the ice Medial moraines are
displayed as long linear dark
"iri aretes are
Chapter 10 372
Firn bank
Firn bank
rqueglacier
Protalus
rampart
Incipient
I
i.
•**•
cirque
(B)
Figure 10.7.
Stages of cirque development. The erosive work in cirques lends itself to the suggestion that Alpine topog-
(A) Nivation beneath firn bank.
raphy can be classified according to stages of its development, and such hy-
(B) Nivation cirque. (C) Cirque
with fully developed cirque potheses are not uncommon (see Embleton and King 1975a). Our purpose
glacier. here, however, is to understand the processes that create the Alpine topog-
raphy, not the sequence of its development. Therefore, we must examine the
origin of cirques and the erosive mechanics that functions within their bound-
aries. Of utmost importance are the processes that scour the basin in the cirque
floorand those that cause the recession of the cirque walls.
Cirques result from two separate groups of processes: (1) mechanical
weathering and mass wasting, and (2) erosion by cirque glaciers. The devel-
opment of a cirque, diagrammed in stages in figure 10.7, begins in a patch of
firn that fills a small depression and stands near the regional snowline. In the
ablation season, meltwater released during the day percolates into fractures
of the bedrock beneath the firn bank and refreezes there at night. The repeated
pressure associated with freezing and thawing presumably wedges out parti-
cles of rock that are then moved slowly downslope by creep and by water flowing
at the base of the firn. The combined processes are commonly referred to as
nivation (Matthes 1900). Actually, nivation refers to a set of geomorphic pro-
cesses, including chemical weathering (Thorn 1976), each of which may func-
tion more effectively under different controlling factors (Thorn and Hall 1980).
Nonetheless, the shape of the original depression is gradually deepened and
widened, and eventually it approaches a semicircular form that can logically
be called a nivation cirque. The nivation hypothesis also contends that with
continued accumulation, the firn changes into true glacier ice, and erosion by
cirque glaciers rather than nivation becomes the dominant process in further
development of the cirque. Exactly when this transition occurs is not clear,
but it is virtually impossible for rock basins cut into cirque floors to be formed
by physical nivation processes alone since they cannot carry particles upslope
to the cirque lip.
Cirque Glaciers Observations made in tunnels excavated into cirque glaciers

indicate that such glaciers move by a process known as rotational sliding, in
which ice slides over the arcuate bedrock floor, rotating at the same time around
The ice exposed in the tunnels displays recognizable yearly
a horizontal axis.
accumulation layers that are separated by marked ablation surfaces, giving
the entire glacier a banded stratigraphy (Grove 1960). Some of the ablation
Flow lines
Figure 10.8.
Ablation surfaces as observed at random positions Long section through a cirque
,.
•—•- Ablation surfaces as computed at 10-year intervals glacier in Norway, showing
' « Debris patch at 10-year intervals

ablation surfaces and debris
patches at 10-year intervals. Note
10yr rotation of ablation surfaces in the
100 down-ice direction.
50
00 yr
250
Meters
zones are laden with debris that fell onto the ice surface near the cirque head-
wall. As figure 10.8 shows, these layers originally dip downglacier at the angle
of the ice surface, and with time each layer is incorporated into the ice mass
as more snow accumulates in the headwall area. As the ice moves, however,
the layers are reoriented so that (as in the figure) near the equilibrium line
they are almost horizontal; while close to the terminus they dip steeply up-
glacier. The deformation of the layers, combined with changes in the position
of stakes in the tunnels and on the surface (McCall 1952, 1960), makes it
clear that flow lines are moving downward near the headwall, parallel to the
surface at the and upward near the terminus.
firn line,
Rotational sliding is an appealing mechanism to explain the scouring of

the bowl-shaped depression in the cirque floor, for this process should be ca-
pable of carrying the products of abrasion or frost wedging upslope and over
the cirque lip. We should emphasize, however, that not all cirque glaciers ex-
hibit rotational sliding as the dominant flow mechanism. The rate of cirque
erosion and growth seems to vary with the geologic and climatic setting. In
polar or subpolar regions, erosion rates have been estimated as 8-76 mm/
1000 years (Anderson 1978), and plucking seems to be the dominant erosional
process. In temperate settings where abrasion is more important, the rate has
been estimated as 95-165 mm/ 1000 years but could have been much greater
during full glacial conditions (Reheis 1975). Actually, the rate of cirque growth
produced by erosion is difficult to ascertain because the time span over which
erosion acted is not precisely known, and the influence of preglacial topog-
raphy is not usually considered (Olyphant 1981b). Regardless, because the
length-to-depth ratio of cirques in widespread locations is normally 2:1, head-
wall retreat probably occurs at a faster rate than deepening of the cirque floor
(Gordon 1977; Olyphant 1981a).
Chapter 10 374
Headwall Erosion The creation of Alpine topography requires removal of

large amounts of bedrock from the head and side walls rather than the mil-
limeter-by-millimeter abrasion occurring on the cirque floor. The surfaces of
the cirque walls are assumed to be prepared for erosion by severe frost action
associated with climates in high elevations. In addition to the shattering caused
by freezing and thawing, joints often develop parallel to the wall faces when
pressure is released by the removal of outer layers of rock (Glen and Lewis
1961). These dilatation joints aid in the fracturing process by providing av-
enues for percolating water and by isolating rock material into discrete units.
The actual processes of frost shattering are not so simple as they first
appear and, likemany processes, seem to be accepted more on faith than on

solid evidence. The phenomenon of frost shattering will be discussed in greater
detail in chapter 1 1. At this juncture, however, you should be aware that lab-
oratory experiments (Battle 1960) suggest that maximum shattering by frost
action occurs only when the temperature falls rapidly to between — 5°C and
— 10°C. A slow decrease in temperature will have little if any effect. In ad-
dition, freezing of water in a crack must proceed from the top of the opening
downward to the bottom. An ice plug must form first at the top of the fracture
in order to produce the closed system that will allow pressure from ice growth
to exceed the tensile strength of most rocks.
Since W. D. Johnson descended into a deep bergschrund in 1904, many
workers have investigated the role played by bergschrunds mechanics in the
of headwall retreat. A bergschrund is a crevasselike opening near the headwall

that separates actively moving ice of the glacier from nonactive ice frozen to
the headwall. Johnson (1904) suggested that surface meltwater gained access
to the base of the headwall by percolating down the bergschund and thus pro-
duced extensive frost shattering when water that permeated into rock frac-
tures was alternately frozen and melted.
Although the bergschrund hypothesis is attractive, it suffers because ac-
tual measurements of temperature fluctuations in bergschrunds (Battle 1960)
are not severe enough to produce shattering. In addition, some cirque glaciers
do not have bergschrunds, and in others the gap does not always penetrate
deeply enough to intersect the rock of the headwall.
In light of the above, we must entertain the possibility that we have mis-
interpreted the driving mechanics. S. E. White (1976a) suggests that hydra-
tion may exert almost as much pressure as that produced by ice expansion at
— 22° C. As minerals adsorb water, expansion and contraction of this non-
freeable water are produced by temperature fluctuations in the range of
freezing and thawing. The shattering seen at the base of the headwall might
then take place beneath a glacier without the stringent conditions necessary
to generate forces by ice growth. The hydration-shattering proposal is thought
provoking and deserves careful investigation. The process not only can explain
shattering where temperature fluctuations are minimal, but it also dispels the
problems associated with the bergschrund hypothesis. In spite of these, ex-
perimental evidence suggests that hydration by itself may not be as effective
as frost action in promoting disintegration (Fahey 1 983), and Washburn ( 1 980)
reminds us that most large products of disintegration are found in periglacial

environments, the point being that hydration shattering should operate in any
climatic zone as long as water is available. Clearly, more work needs to be
done before the relationship between frost action and hydration shattering is
completely understood.
Glacial Troughs
Glacial erosion is not limited to the cirque environment, for ice passing over
the cirque lip can also remold the preglacial valley topography into a char-
acteristic glaciated form. The ability of ice to remove rock protuberances tends
to produce valleys with steep, nearly vertical, sides and relatively wide, flat
bottoms (fig. 10.9). The transformation of a V-shaped river valley into a U-

shaped glacial valley has been explained mechanically by A. Johnson (1970).
Assuming ice to be pseudoplastic and using Nye's (1965) analysis of ice be-
havior in a triangular form, Johnson suggests that dead regions (areas of no
Figure 10.9.
(A) U-shaped cross-profile of
glaciated valley and a hanging
valley —
view up Yosemite Valley
from vicinity of Artist Point El
Capitan at left, the Cathedral
Rocks and the Bridalveil Falls at
right. Yosemite National Park,
Mariposa County, Cal.
(B) Longitudinal section of glacial
valley in Yosemite National Park
illustrating staircase profile AA is
preglacial valley floor; BB is
present valley floor. Broken lines

represent intermediate stages of
development (After Matthes
1930)
^J\
-*-c-
Chapter 10 376
Figure 10.10.
Possible sequence of events flow or shear stress) should exist in a glacier that invades a mountain stream
leading from V-shaped mountain
canyon to U-shaped glacial valley.
valley. The dead regions should be present in the ice at the top of the valley
(A) V-shaped mountain canyon. sides and along the valley bottom (fig. 10.10). Erosion would be negligible in
(B) V-shaped canyon visited by these zones but appreciable along other parts of the valley sides where shear
glacier. (C) Glacier erodes sides of
stress and velocity would be high. As differential erosion causes the sides to
canyon (D) "Dead" regions
disappear and entire side of bow outward, the dead regions are progressively removed because the shear
canyon is rasped by rock-studded stress and velocity distribution changes with the approach to a more parabolic
glacier. (E) Glacier disappears,
cross-profile (fig. 10.10). Eventually, erosion affects all parts of the valley sides,
leaving U-shaped valley (From
Johnson 1970, Physical Processes resulting in the characteristic U-shaped glacial valley (fig. 10.10).
in Geology With permission of A parabolic cross-profile aids glacial movement because it probably exerts
Freeman, Cooper and Co.)
the minimum resistance to glacier flow (Flint 1971). The precise width-depth
dimensions, however, probably depend on the intensity of the glacial dynamics
(Graf 1970) and the properties of the geologic framework. The formation of
the cross-sectional shape occurs by both lateral and vertical erosion of the
preexisting valley. Whether the parabolic form is derived predominantly by
widening or predominantly by deepening depends on the properties of the rocks
and the ice, as well as the extent of preglacial weathering and the amount of
load available within the ice to be utilized as cutting tools. In any case, the
erosion leads to the truncation of rock spurs jutting into the valley and the
formation of hanging valleys. These occur particularly where trunk valleys
carry more ice and are more extensively eroded than their tributaries.
Glaciated troughs are also characterized by uniquely irregular longitu-
dinal profiles that essentially represent a series of interconnected basins and
steps; the diagram in figure 10.9 shows this typical profile in Yosemite Na-
tional Park. The basins may contain lakes, a series of which are sometimes
called paternoster lakes. Immediately downvalley from the basins holding the
lakes are rock bars or steps that commonly show the effects of intense abra-
sional smoothing on their relatively fiat surfaces. At the distal end of the step,
a sharp break in slope occurs where plucking has produced a steep scarp that
connects the rock bar to the next lower basin or step.
The staircase profile has intrigued geologists for years and has been the
subject of considerable speculation. Hypotheses for its origin include
(1) variation of rock structures, especially spacing of joints, causing differ-
ential erosion, (2) preparation of weak zones in the rock by preglacial weath-
ering, (3) irregularities in the preexisting valley topography, and (4) increased
erosive power at the confluence of tributary valleys with the main valley (see
Bakker 1965). It is possible that rock bars producing a steplike profile require
no special conditions for their formation other than the normal slip-plane ori-
entation noted in glacier flow (Nye and Martin 1967). Theoretically, rock bars
can form where there is no obviously harder rock, and conversely, zones of
resistant rocks do not necessarily evolve into rock bars under glacial erosion.
The abrasion prevalent on the treadlike steps and the quarrying at the
position of the steep scarps led Lewis (1947) to note that each step resembles
a large roche moutonnee and so probably has a similar origin. He concluded
that the mechanics probably worked best beneath thin glaciers where melt-
water could easily penetrate to the glacier floor and shatter the rock by re-
freezing. Although freeze-thaw mechanics might be a reasonable accomplice
in the formation of staircase profiles, we cannot ignore the fact that devel-
opment of the associated basins probably required some differential ice move-
ment, capable of flowing uphill and carrying load in the process. Such a
requirement does not encourage the belief that the ice was thin; in fact, it is
common knowledge that many valley glaciers have thickened to the point of
overtopping their divides and spreading into adjacent valleys. In addition, as
discussed earlier, the freeze-thaw component in rock shattering is somewhat
suspect because we do not completely understand the thermal conditions at
the base of glaciers. What we are left with, then, is the problem that while
plucking is necessary to form the scarps, the rock shattering required for this
process to function is probably not a result of frost action. Furthermore, some
scarps form where no regional joint system is present to facilitate the shat-
tering process.
Boulton (1974) suggests an alternative hypothesis of shattering that de-
serves our consideration becauseit not only resolves the above problem, but
also relates directly to glacier mechanics. When a glacier moves across a hor-
izontal bedrock surface, the effective normal pressure bepgh — wp where
will
p,g,h and w_ are, respectively, density, gravity, ice thickness and water pres-
sure at the bedrock interface. The dashed line in figure 10.1 1 represents the
normal pressure on the horizontal bed and estimated by p gh (where p,
is p). t
=
AP in the figure represents the increase or decrease of normal pressure pro-
duced by a bedrock obstruction; the total normal pressure at any point be-
coming pjgh +
AP. Therefore, if the bed is irregular the normal pressure will
fluctuate so that it will be higher than average on the upglacier side of a bed-
rock obstruction and lower than average on the lee side. Boulton shows that
Chapter 10 378
Figure 10.11.
Schematic view of normal
pressure distribution at glacier
bed as ice flows over a bedrock
obstruction IP represents the
increase or decrease of normal
pressure produced by a bedrock
obstruction The normal pressure
where no obstruction exists is
represented by the dashed
horizontal line and has a value
equal to p.gh
the shear stresses included in the bedrock will be greatest where the normal
pressure is lowest, i.e., down-ice from the crest of the obstruction. The exact
positionand absolute magnitude of the maximum shear stress depend on
whether cavitation occurs downglacier from the bedrock knob. Furthermore,
Boulton points out that the shear strength of the rocks will be least in the
downglacier position where the shear stress is greatest, creating the ideal sit-
uation for rock failure at that locale. Assuming this analysis is correct, the
inevitable conclusion is that with the proper bedrock configuration, even hard
unfractured rocks may be crushed beneath the ice, and the rocks already jointed
can be intensely shattered. Preglacial irregularities in the long profile, struc-
tural weaknesses, and lithologic variations all will be accentuated by this pro-
cess because shattering will prepare those zones for the plucking action that
produces the scarps in the staircase profiles.
A special type of glacial trough exists mainly in high-latitude coastal re-

gions that are underlain by resistant rocks, so that the general land surface
stands at considerable elevation above the nearby ocean. These troughs are
called fiords, and they differfrom other types only because they are partially
submerged by the ocean. The inundated bottoms of fiords have the same va-
riety of topographic elements, both erosional and depositional, that exist in a
normal continental glaciated valley (Holtedahl 1967). Their history may in-
clude components of both glacial and fluvial processes, and they may be partly
controlled by tectonic and lithologic factors. For these reasons, it is unwise to
make sweeping generalizations about their origin.
Perhaps the most salient property of fiords is that part of their develop-
ment took place when the icewas physically beneath the ocean (Crary 1966).
Flint ( 1 97 1 ) reminds us that a glacier 000 m thick with a density of 0.9 will
1
remain in contact with its bed and be fully capable of erosion at water depths
up to 900 m. Even at greater depths, when the snout begins to float, high
topographic irregularities of the valley might still be eroded (Crary 1966).
The water depths over fiords, several hundred meters in many and greater
than 1000 m in some, are well beyond that which can be attributed to a post-
glacial rise in sea level (see Flint 1971), adding credence to the suggestion
that much fiord erosion was accomplished in a submarine environment.
Before glaciers were recognized as viable geomorphic agents, deposits con- Deposits and
taining boulders that obviously came from a distant source were called drift. Depositional Features
This term arose because elimination of known processes led to the belief that
the anomalous boulders reached their site of deposition by riding on top of
floating ice. After glaciers were recognized as the transporting vehicles, the
term "drift," or glacial drift, was retained and expanded to include all deposits
associated with glaciation. It is estimated that drift covers 8 percent of the
Earth's surface above sea level and almost 25 percent of the North American
continent. The thickness of this cover varies greatly. In the United States, for
example, only a thin layer of drift (< 20m) covers upland areas in most of
New England, although drift may be several hundred meters thick in buried
valleys. The drift in the central United States is generally from 10 to 60 m
thick, but once again these are average values. In some places the drift is merely
a thin mantle on top of bedrock, and in other regions, such as parts of Mich-
igan, it exceeds 200 m in thickness. The exact volume of drift deposited de-
pends on the time span of glacier activity, but with high velocities and loads,
as much as 30 m of drift can be accumulated in less than 1 years (Flint 1971).
Through the years geologists have been intrigued with the amazing va-
riety of glacial drift and the complex interrelationships that exist between the
different types. This complexity arises because ( 1 ) drift may be deposited from
mediums that contain vastly different amounts of water; (2) deposition occurs
beneath, within, or on top of ice, at the glacier margins, in bodies of standing
water, or in fluvial settings far from the glacier, the debris being transported
there by streams rising in the ice mass
the depositional sites and
itself; (3)
environments and the drift change with time because glaciers

composition all
themselves are not constant in their properties or fixed in their position; and
(4) the glacier may be active or stagnant. Because of these complicating real-
ities, any discussion of glacial drift and depositional features is difficult to or-
ganize in a way that is entirely satisfactory. In this approach, we will first
briefly examine the varieties of glacial drift and then examine features ac-
cording to the depositional environment in which they originate. An attempt
will be made
to relate the morphology of features and their sedimentary prop-
dynamics of the system. Throughout this discussion it is important
erties to the
to recognize the distinction between the sedimentological character of drift
and the morphology of features that result from its deposition. Morphological
terms such as moraines and kames are not to be interpreted as implying a
particular drift type. Many features with similar morphology are composed
of a number of drift varieties, especially where the environment of deposition
is subject to repeated change.
Drift Types
Over the years glacial deposits have normally been divided into two categories
based on their sedimentary characteristics. Particularly important are the
presence or absence of layers and the degree of sorting in the deposit. In our
discussion we will separate drift into stratified and nonstratified types.
Chapter 10 380
Nonstratified Drift Sediment originating directly from glacial ice charac-

teristically has no discernible stratification. The material is usually called till
and typically is a nonstratified mass of unsorted debris that contains angular

particles composed of a wide variety of rock types. In addition, the term "till"
usually connotes material that has been transported and deposited by the ice
itself, a process often indicated by striations or microscopic fractures on the
grains (Krinsley and Donahue 1968). Many examples justify this description
of till; absence of layering and poor sorting (especially an almost universal
bimodal size distribution) seem to be the most reliably consistent properties.
The bimodality observed in most tills probably is due to the differences in
grains produced by abrasion and those derived by plucking. In an excellent
review, Goldthwait (1971, p. 4) points out that till is probably more variable
than any other sediment that is described by a single name.
Any of the identifying criteria in our definition may be missing at a par-
ticular till locality as a result of varying transporting and depositing me-
chanics and of heterogeneity in the rocks over which the ice has passed. For
example, many clasts in till have a subangular pentagonal or triangular shape,
but these forms may be significantly altered by rounding during their trans-
portation. This is especially true in till that was transported subglacially. In
that environment, parent debris rounded by attrition of edges and corners
is
(Boulton 1978), although the degree of rounding is partly dependent on rock

type (Holmes 1960; Vagners 1966), the original clast shape (Drake 1968.
1974), and the distance of transport. In addition, glaciers that override older
Figure 10.12.
stream deposits may incorporate in their load boulders that have already been
Late Wisconsinan (Pinedale) till in
Rock Creek valley, Beartooth rounded, resulting in a till that is notably less angular than one would expect.
Mountains. Mont. Note rounded Figure 10.12 shows till that contains rounded boulders.
boulders in till.
at. *•
*
Overall particle size tends to be reduced by attrition in subglacially trans-

ported till (Mills 1977; Boulton 1978). Each particular source rock tends to
produce a characteristic texture depending on how easily its clasts can be
crushed and how far they have been transported (Mills 1977; Dreimanis and
Vagners 1971). Thus, the overall angularity and texture of till depend on the
lithologic heterogeneity of the source rocks as well as the distance of transport
from the location of their outcrop.
The unpredictability of till increases still more because the material is
carried in different parts of the glacier where the ice is characterized by di-
verse mechanics. In ice sheets, debris is carried as englacial load in layers
1-100 m above the base of the ice, as indicated by the fact that the surface
is free of any load except near the margin (Goldthwait 1971). In valley gla-
ciers, on the other hand, considerable debris may be shed onto the surface
from the valley sides and be transported as supraglacial load. Each transport
subenvironment produces till with different characteristics. Supraglacial till
has a texture dominated by coarse, angular clasts because the particles are
not crushed during transport, and fines tend to be washed away as the surface
ablates. Subglacial till is more compact and contains a higher percentage of
fine-grained sediment. Englacial load can be deposited directly on the bedrock
floor as subglacial matter, and a considerable thickness of the glacier will re-
main above it at the moment of deposition. It can also rise within the ice along
shear planes or along the normal upward flow lines in the terminal zone. This
material finally emerges as supraglacial debris when ablation of the surface
ice releases the contained particles. In this case the till will be emplaced on
the bedrock surface as the ice beneath it dissipates, and no ice will remain
above the till at the time of its final deposition. Till laid down in the subglacial
environment under pressure of the overlying ice is usually referred to as basal
till or lodgement till. In contrast, ablation till occurs where the debris is con-
centrated at or near the surface and is gradually let down onto the bedrock
as the ice disappears.
Ablation till forms as one of two types, flow-till (Hartshorn 1958; Boulton
1968) and melt-out till (Boulton 1970b). When englacial debris is released at
the surface, the till tends to be highly mobile and likely to move downslope as
a quasi-mudflow, by creep, or as a semiplastic slide. However, when the thick-
ness of the surface cover reaches about 3 cm, the ablation process is severely
retarded, and if m, ablation ceases. Further
the surface layer exceeds 1 or 2
increase in the thickness of the supraglacial cover can occur only if more de-
bris moves to the site from higher levels of the ice surface, or if the ice melts
beneath the layer already concentrated at the surface. Boulton (1971, 1972b)
asserts that supraglacial till may have two components: an upper zone accu-
mulated from debris transported from a higher source (flow-till), and a lower
zone accumulating in situ and never actually exposed at the surface (melt-out
till). Melt-out tills rarely exceed 3 m in thickness, but the total supraglacial
layer can be much greater if flow-till continues to accumulate on the surface.
Little is known about the rate of accumulation of ablation tills, but Mickelson
(1973) has calculated that the accumulation rate of basal melt-out till of the
Burroughs Glacier in Alaska ranges from 0.5 to 28 cm a year.
Chapter 10 382
Stratified Drift The second major category of glacial drift is distinguished

by the fact that the sediment was transported by moving water before its final
deposition, thereby acquiring a degree of stratification not normally seen in

tills. Such drift is often referred to as fluvioglacial because running water is
involved in its origin, even though the water may not always be confined in
discrete channels. Fluvioglacial deposits are also distinguished from till in that
they are usually sorted and the clasts contained mass are more rounded.
in the
However, the demarcation between some types of fluvioglacial deposits and

thoroughly washed ablation tills is a matter of degree rather than substance.
Exactly where the line between the two is drawn becomes somewhat arbitrary.
Highly saturated flow-till, for example, might move in a nearly fluvial manner.
The layering and sorting in a fluvioglacial deposit depends on precisely
where it is formed with respect to the ice that provides the transporting melt-
water. Sorting is also partly a function of the energy possessed by the melt-
water, the distance of transport, and the continuity of the sorting process.
Because the discharge of meltwater is notoriously inconstant, varying dras-
tically with time of day, local climate, and the characteristics of the ice. sig-
nificant differences in the sedimentology of fluvioglacial deposits can be noted

over short distances. These are especially evident where deposits are formed
in contact with the ice and the free circulation of meltwater is restricted (see
Shaw 1972 for sedimentary characteristics). If debris is transported away from
the glacier terminus, the sedimentary characteristics tend to vary more reg-
ularly.
We should stress again that the prime requisite of stratified drift is trans-
port by water, much of which is released from melting ice. However, this puts
no constraints on the environment in which the sediment is deposited. Flu-

vioglacial debris can come to rest in stream channels, floodplains, lake or ocean
floors (Rust 1977), deltaic plains, in contact with ice, or in any other place
where sediment-laden running water loses its transporting energy. In some
cases sedimentologic properties in the deposit, such as the degree of roundness
and mean grain size, can help identify the depositional environment (King and
Buckley 1968), but normally there are too many variables in the system to
rely on these critiera alone (R. J. Price 1973). Detailed field study is almost
always necessary to reach a firm conclusion about environments of deposition.
Deposits that originate in contact with the ice often contain interbedded
bodies of ablation till, and their particles tend to be less well rounded and
sorted because of the limited distance of transport. The characteristics of such
a deposit often vary from the bottom of the sequence to the top because the
environment of deposition in contact with the ice repeatedly changes with time,
especially if the ice is stagnating. In any deposit, then, a particular layer may
be superseded vertically by one with different sedimentary properties that re-
flect a new depositional environment. The stratigraphy is complicated by the

fact that much of the drift is physically supported by ice during its deposition,
and when the ice dissipates, the removed and the sediment collapses.
support is
Such a process, as figure 10.13 shows, leads to flexures in some of the layers,
minor faults, and beds dipping at angles well beyond the angle of repose for
such material. Overall, the ice-contact setting at places produces an intercon-

nected maze of stratified and nonstratified drift in which every conceivable
process and environment is possible and probably has been present at some
time during the depositional history. Moreover, the manner in which the ice
ablates influences the resulting deposit. Drift produced while the ice is still
active may be quite different, especially from that derived from

in distribution,
a large mass of stagnant ice that is simply downwasting as it melts and is not
influenced by internal glacial movement.
Sediment deposited beyond the terminal margin of the ice is formed in
the proglacial environment and is often referred to as outwash. Outwash is
usually well sorted and normally consists of rounded sands and gravel repre-
senting bedload carried and deposited in stream channels. Silt and clay are
usually carried as suspended load and are commonly removed from the system
unless, as in the lower Mississippi River valley, the transport distance is so
great that some of the outwash is silty in texture. It is important to understand
that streams transporting outwash do not usually head at the glacier terminus
but begin on top of or within the ice, well upglacier from the margin. Pro-
glacial features and deposits often can be traced into and through the maze
of ice-contact deposits, increasing the complexity of the depositional sequence
developed near the ice margin.
Figure 10.13.
(A) Gravel pit cut in kame, south
The Depositional Framework New shows
York,
central
Before considering what geomorphic features might be produced during de- stratification in kame deposits.
position, we must first establish a realistic framework within which we can (B) Structures and deformation of
strata in ice-contact stratified drift
give some order to the subtle variations of depositional features associated
develop as ice melts and debris
with glaciation. This is done in table 10.1. From our brief look at the nature collapses and is lowered onto
bedrock floor.
Chapter 10 384
Table 10.1 The depositional framework and associated features.
Setting Features Type of Drift
Ice contact
Marginal End moraines Till and fluvioglacia
Kames and kame Fluvioglacial
terraces
Kettle holes Fluvioglacial
Eskers Fluvioglacial
Interior Medial and interlobate Till
moraines
Ground moraines Lodgement till
Fluted surfaces Lodgement till
Drumlins Lodgement till
Proglacial environment Sandar Fluvioglacial (outwash)

Kettled sandar 3 Fluvioglacial (outwash)
a May merge with marginal environment
of drift, it is obvious that both stratified and nonstratified deposits can be formed
in ice. They differ only in whether the ice alone was the pri-
contact with the
mary transporting and depositing agent or whether a stage of water transport
intervened between the release of particles from the ice and their final depo-
sition. As a result, the ice-contact environment must be considered as one of
the major settings of our depositional framework. The ice-contact environ-
ment can be subdivided geographically into marginal and interior zones de-
pending on where the drift was originally deposited (table 10.1).
Features formed in the ice-contact environment can be composed of either
stratified or nonstratified drift (till or fluvioglacial sediment) or a combination
of both types. The region beyond the terminal edge of the glacier is classified
as the proglacial environment; in contrast to the ice-contact setting, features
formed there are composed entirely of fluvioglacial sediments (outwash).
The distinction of marginal and interior zones in the ice-contact environ-
ment is problematical for several reasons. First, glacial margins migrate for-
ward and backward with time according to the mass budget. An active glacier
with a negative mass balance should have a terminal margin that is progres-
sively receding toward its source; marginal features will be formed in regions
that were interior when the ice was at its greatest extent, and deposits of the
two zones may be complexly intertwined. Second, the process operating near
the contact between the marginal and interior zones are dependent upon the
characteristics of the subglacial environment. For example, the outer 2 to 3
km of ice sheets are commonly frozen to the underlying surface even though
farther upglacier the basal ice may be temperate. Where marginal ice is frozen
to the bed and high porewater pressure exists, thrusting along preexisting planes
of weakness in the substrate can inject large blocks of subglacial material into
the ice (Moran et al. 1980). These thrusted blocks tend to concentrate where
ice margins rest on aquifers and on the upslope edges of upland areas. Pro-
ceeding upglacier toward the interior zone, the ice is free to slide because the
ice-bed contact is unfrozen. Thrust blocks in the transitional area between the
Figure 10.14.
marginal and interior zone are smaller and smoothed over by the sliding ice. Map view of the depositional
framework after ice has
In the true interior zone, no thrusting will occur, and the subglacial terrain is
disappeared Part of the
characterized by streamlined forms (Moran et al. 1980). Clearly, the position Whitewater, Wis. quadrangle
of the frozen ice/thawed zone contact may migrate with time and changes in (U.S.G.S. 15') Marginal zone
contains terminal moraine and ice-
the subglacial environment. Thus, in any given glaciated area, the inner por-
contact stratified features, and
tion of the marginal zone may be transitional into the interior zone rather than interiorzone is characterized by
marked by a well-defined contact between the two. ground moraine Proglacial zone is
a large outwash plain or sandur

Regardless of the problems associated with precise boundary locations,
the depositional framework fits our purposes because the suite of features found
in each zone is a direct reflection of the genetic processes involved. The utility
of the classification shown in the map in figure 10.14. In this region the
is
terminal moraine, marking the ice-contact marginal zone, is characterized by

a maze of small hills and depressions (kames and kettles) that distinguish de-
posits formed near the boundary between active and stagnating ice. North of
Chapter 10 386
Table 10.2 Moraine types.
End Moraines Moraines produced at front or sides of an actively flowing

3
glacier
Terminal moraines Mark the farthest advance of an important glacial episode
Lateral moraines Deposited at or near the side margin of a mountain glacier.
Recessional Formed at glacier front during temporary halt or readvance of ice

moraines in a period of general recession
Ground Moraine Gently rolling surface formed of debris released from beneath the
ice.
Interior and
Minor Varieties
Washboard Small, parallel ridges oriented transverse to direction of ice
moraines movement Also called moraine ridges or cross-valley
moraines.
Interlobate Formed at junction of two ice lobes
moraines
Medial moraines Elongate ridge developed at junction of two coalescing valley
glaciers.
Rogen Moraines Large sequence of ridges transverse to ice flow Formed in the
interior zone.
^Lateral moraines may be excluded by some geomorphologists because end moraines are commonly considered only as
topographic features developed at the front of a glacier
the moraine the relatively low, flat area is underlain by material that was de-
posited beneath the active glacier (ground moraine); it represents the ice-
contact interior zone. South of the moraine, the proglacial zone is marked as
a plain composed of debris (outwash) transported and deposited by meltwater

streams heading within or on top of the ice.
Marginal Ice-Contact Features
Moraines The term "moraine" originated several hundred years ago as a

local name for ridges of debris found at the edges of glaciers in the French
Alps. Since then many definitions have appeared, but for our purposes we can
think of a moraine as a depositional feature whose form is independent of the
subjacent topography and that is constructed by the accumulation of drift,
most of which is ice-deposited (Flint 1971). A precise morphological definition

is not possible since, as table 10.2 shows, moraines take many different forms
and have a variety of dimensions. The term "moraine" is not synonymous with
"till," as many have suggested, but in reality refers to a suite of topographic
forms on which the only restriction is that they must be composed of drift
(and even that rule has been violated in some cases).
The most spectacular moraines develop at or near the edges of active gla-
ciers and so are designated as end moraines. The end moraine constructed at
the downstream edge of the ice at the farthest point of advance is called a
terminal moraine. In valley glacier systems, it merges imperceptibly into lat-
eral moraines on both sides of the valley because ablating ice deposits debris
along the glacier's lateral extremities as well as at its terminus (fig. 10.15).
Figure 10.15.
(A) Terminal moraine of Pinedale
glaciation, East Rosebud valley
near Roscoe, Mont Low ridge
above road in left center of photo
is lateral moraine merging with
terminal zone. (B) Recently

formed lateral moraine in Rocky
Mountains of Canada.
(C) Terminal moraine on southeast
end of Devils Lake, Wis. Moraine
dams old valley of the Wisconsin
River
Chapter 10 388
Figure 10.16.
Diagram showing retreating
nrough
margin of Barnes ice cap Material
moves up inclined shear planes
and accumulates as till cover on — P-* v u
J2 JJJag e
Ice-cored
moraine
core of ice Alluvial I
fan End moraine in place
Ice sheets form terminal moraines that tend to be long (often hundreds of
kilometers), linear, topographic highs marking the forward boundary of the
ice, but lateral moraines are absent because there is no side to the ice sheet.
Where ice sheet movement was end moraines called
distinctly lobate, however,
inter lobate moraines (Flint 1971) may develop along the junction of two lobes.
Ideally, end moraines assume a rather narrow, ridgelike shape, but ac-
tually the form and size depend directly on the amount of glacial load, the
mass budget, and the volume of meltwater circulating in the system. Tem-
perate valley glaciers tend to build higher and more massive terminal mo-
raines (some reaching 300 m in height) because these glaciers have higher
flow velocities, loads, and total budgets. Ice sheets seem to generate moraines
that are less dramatic in size, seldom exceeding 50 m in height. Where mo-
raine ridges are transverse to the flow direction, it is difficult to distinguish
between those that develop at the glacier front (end moraines) and others that
are formed subglacially. This difficulty arises because ridges in both geograph-
ical positions result from similar processes. R. J. Price (1973) distinguishes
these processes as dumping, squeezing, and pushing.
The mechanics of the dumping process were first described in detail by
Goldthwait (1951), who observed that debris in the Barnes ice cap is con-
tained in a series of shear planes near the terminus of the ice (fig. 10.16). The
sheared zone extends only approximately 150 m from the snout, and the at-
titude of the planes suggest they cannot penetrate the ice to depths greater
than 75 m. This indicates that, if the debris is brought surfaceward from the
basal ice layers, the process occurs only in the outer 500 m of the glacier.
Figure 10.16 shows that as the ice fringe thins by ablation, dirt contained
in the shear zones is released, creating a film of sediment on the ice surface
that Goldthwait refers to as "black ice." Some of the debris moves downslope,
as described earlier, where it blankets the lower ice up to a thickness of a
meter, thereby retarding the ablation process. Black ice with only a thin dirt
cover ablates more rapidly than the white ice upglacier from the sheared zone,
and a trough is produced parallel to the ice front that separates the active
black ice zone from ice-cored debris at the front edge of the glacier. The final
hummocky surface, so characteristic of moraines, results from debris sliding
off the ice-cored ridge, debris collapsing as the underlying ice melts, and from
fluvioglacial action.
The shear-plane hypothesis is attractive when the outer fringe of the ice
is stagnant. It is possible, however, that active ice can also move basal debris
surfaceward along normal flow lines that consistently intersect the ice surface
near the terminus (Boulton 1967). Regardless of the precise mechanics by
which load topography in a morainal
rises to a glacier surface, the evolution of
belt proceeds complex manner depending on (1) the character of the ice,
in a
(2) the spacing of the debris bands as they emerge at the surface, and (3) the
thickness of the supraglacial cover. These factors produce differential ablation
rates and widespread zones of drift that are held up by cores of solid ice.
In contrast to moraines formed by the dumping process, moraine ridges
are also developed by squeezing drift originally deposited beneath the ice.
Ridges generated by squeezing are usually smaller than dumped moraines,
normally standing less than 10 m high, although they can be higher where the
subglacial till is highly saturated. Theoretically, the process involves the re-
sponse of water-soaked lodgement till to pressure exerted by the weight of the
overlying glacier. The till will move from under the ice and emerge along the
ice front, or it will be squeezed into zones of low pressure within the ice, rep-
resented by crevasse openings (R. J. Price 1970; Mickelson and Berkson 1974).
In either case the ridges seem to be typified by (1) steeper distal slopes than
proximal slopes, (2) a plan-view outline consisting of linked arcuate segments
that parallel the ice front, and (3) a till fabric with pebbles oriented perpen-
dicular (or nearly so) to the ridge crests, though occasionally oriented parallel
to the trend of the ridge (Mickelson and Berkson 1974).
R. J. Price (1970) examined a series of small morainal ridges in Iceland
that formed during the last 70 years as the ice retreated from the main, and
larger, terminal moraine. The ridges apparently rise quickly (in one area 15
ridges formed in 16 years), prompting Price to suggest that they develop an-
nuallywhen summer meltwater descends to the glacier base and saturates the
till (a proposal made earlier by Andrews and Smithson 1966).
subglacial
Squeezed ridges also seem to originate beneath bodies of standing water
(Mickelson and Berkson 1974); since saturation of the debris would be rather
complete in such an environment, this adds credence to the proposed mech-
anism. Hydrologic processes, however, may be involved in the origin of sub-
lacustrine or submarine ridge development (Barnett and Holdsworth 1974).
A third mechanism capable of forming
a moraine ridge is the collision of
advancing which deform them into a ridgelike feature.
ice with older deposits,
Such end moraines, called push moraines, are only partly composed of sedi-
ment carried by the ice and may in fact include blocks of older rocks of a
nonglacial origin (Kaye 1964b; Andrews 1980). Push moraines usually have
a steep distal slope, and they often display a sharp break in slope at the contact
between the moraine and the proglacial deposits (Embleton and King 1975a).
They are more distinctive, however, in their internal structure. Here individual
layers, sheared away from the ground lying in front of the ice, are intensely
deformed into large (sometimes overturned) folds and faults of all kinds (Mills
and Wells 1974). Thrusting, with plates as thick as 30 m, is not uncommon
and may occur as imbricate slices or even as an underthrust phenomenon (Kaye
1 964a, 964b). The exact style of internal deformation depends on the rigidity
1
Chapter 10 390
of the bedrock being shoved by the ice front or on how tightly unconsolidated
material is held together by freezing around the ice perimeter.
Most, if not all, of the processes that function at the glacier front can also
create moraine ridges subglacially (Sugden and John 1976). For example, large
transverse ridges, called Rogen moraines or ribbed moraines, are 10-30 m
high, > 1 km and tend to develop as a series of separate hills spaced
long,
100-300 m apart. They probably form by shearing or thrusting behind the
glacier front, usually in broad depressions of older till sheets or in local bed-
rock valleys. Significantly, the ridges are often in direct assocation with flutes
or drumlins, and therefore, it is reasonable to assume that they all have a
common origin. This suggests that the ridges reflect an interaction of basal
debris, porewater pressure, and ice temperature and that the Rogen system
develops where transverse variations in stress exist at the glacier bed (Sugden
and John 1976).
The mechanics associated with Rogen moraines are remarkably similar
to the thrusting phenomena proposed by Moran and his colleagues (1980). As
discussed earlier, the thrusted blocks seem to occur along the inner part of the
marginal zone and often assume a morainic topography. This topography may
be subdued, however, where temperate ice slides over the blocks.
The pushing and squeezing phenomena may also relate to smaller vari-
eties of subglacial moraines. Within the marginal zone the locus of deposition
shifts periodically, giving rise to ridges, mounds, and depressions of varying
size. The ridges in the sequence, usually small and composed of till, commonly
parallel the orientation of the ice front. They have been given many names,
such as cross-valley moraines (Andrews 1963; Andrews and Smithson 1966)
and washboard moraines or moraine ridges (R. J. Price 1970), and several
theories have been suggested for their origin (Elson 1968). The ridges are
usually segmented or interwoven with many deposits of stratified drift, giving
the entire topography a chaotic expression rather than a regular undulation
of ridge crests and intervening sags.
Moraines are not always characterized by distinct ridges that have de-
veloped transverse to the direction of ice flow. Instead, moraines may exist as
a belt of interspersed mounds and depressions that are merged into a chaotic
topography completely devoid of linear ridges. These moraines, commonly
called disintegration moraines (Gravenor and Kupsch 1959), have local relief
up to 70 m and develop from supraglacial drift in the lower part of the ablation
zone. When the ice in glacier margins stagnates, it often breaks into isolated
blocks of wasting ice covered by ablation till (fig. 10.17). Thus, although the
depositional environment is ice-contact marginal, the morainic topography
gradually develops as the glacier surface downwastes over dissipating ice cores.
This is different from ridges formed from active ice that oscillates back and
forth or from those created by pushing, squeezing, or thrusting. In stagnant
marginal zones, flow till, melt-out till, and fluvioglacial deposits can all coexist
on the wasting surfaces and, because the ice cores melt at different rates, may
Figure 10.17.
s Ice-cored moraine, Yanert Glacier,
Alaska
Chapter 10 392
become and mixed within the chaotic surficial expression. For ex-
shifted
ample, slow melting allows till to flow into and fill hollows and trenches in-
terspersed between the ice-cored mounds. When the ice finally disappears, the
topography is inverted so that the former depressions, now filled with consid-
erable thickness of drift, stand as circular or irregular hills (fig. 10.18).
Stratified Marginal Features As pointed out earlier, many glaciers are char-
acterized by marginal zones of thin, stagnating ice.Within these zones a suite
of genetically related ice-contact features is developed, composed predomi-
nantly of stratified drift and morphologically distinct from the moraines just
described. These features form by deposition of drift (1) where water flows
through openings in and beneath the ice, or in ice-surface channels, (2) in
spaces between the ice and the bedrock of the valley sides, and (3) where dis-
seminated sediment is passively concentrated by melting of the encasing ice.
Many of the channelways for the flowing water originate when stagnating
ice breaks into individual segments along planes of structural weakness in the
ice,and so the features are genetically related to many of the ice-disintegra-
tionforms described by Gravenor and Kupsch (1959), Stalker (1960), Clayton
(1967), Parizek (1969), and many others. Our discussion, however, is re-
stricted to those features that consist primarily of fluvioglacial drift, even
Figure 10.18. though ablation till may be a minor ingredient. Their morphology is entirely
Hummocky topography in terminal
moraine East Rosebud valley at constructional, although they can be affected by slumping as the ice walls that
front of Beartooth Mountains near supported the drift during its deposition are melted away.
Roscoe. Mont.
Karnes and Kettles Karnes are moundlike hills of layered sand and gravel
that vary in size from minor swells to conical protuberances standing up to
50 m high and extending 400 m along their base. Kame material can accu-
mulate at the ice-substratum interface, and also in cavities located within
stagnating ice or on its surface. Englacial or supraglacial debris can be low-
ered onto the ground surface as the ice dissipates (Cook 1946; Holmes 1947).
The moundlike shape results only if the accumulate was originally isolated
and nonlinear (fig. 10.13), because sediment deposited in linear openings is
ultimately converted to ridges rather than mounds. In addition, some kames
form when debris collects as fans or small deltas built against the ice or out-
ward from the ice, with the apex resting at the stagnant margin. In either case,
melting of the supporting ice allows the drift to collapse into a kame, a process
evidenced by slumped strata within the deposit.
Kames are only one of many forms with essentially the same origin. They
are transitional into eskers or minor ridge and circular forms that Parizek
(1969) calls ice-contact rings and ridges, or into other features whose names
utilize the term "kame" as a descriptive adjective, i.e., kame delta, kame mo-
raine, etc. Perhaps the most common of the latter type is a kame terrace.
Kame terraces originate from drift deposited in narrow lakes or stream chan-
nels between the valley side and the lateral edge of the stagnating ice. When
the supportive ice disappears, the inner edge of the deposit collapses into the
terrace scarp. Kame terraces differ from normal river terraces in that they are
restricted in their longitudinal extent and are usually narrow; the tread may
slope gently into the valley, and the surface may be dimpled by kettle holes
(McKenzie 1969).
Kettle holes are circular depressions that are formed in a variety of ways
(Fuller 1914), most commonly by the burial of isolated blocks of ice by strat-
ified drift. The gradual ablation of the ice leads to a gentle downward flexing
of the layers as they settle over the dissipating mass. Some kettles are almost
50 m
deep and up to 13 km in diameter (Flint 1971), but these giants are
exceptions to the normal kettle size of less than 8 m deep and 2 km wide.
Kettles usually form in association with kames and other related features, pro-
ducing an irregular surface described as "kame-and-kettle topography," but
similar surfaces can develop in the absence of either or both of these features.
Eskers The term "esker," evidently stemming from the Gaelic word for
"crooked" or "winding," has been applied to a wide variety of ridged ice-
contact features (fig. 10.19). Eskers range in shape from the single, narrow,
sinuous ridge that is the classic form to a complexly intertwined maze of
branching and joining ridges (Huddart and Lister 1981). Eskers seem to form
most commonly in stagnating margins of large ice sheets where the underlying
surface is broad and relatively flat. The ridges are not necessarily continuous
but may consist of crudely connected linear segments. In broad valleys the
eskers usually parallel the slope of the valley floor, but this is not always the
case. Some ridges, for example, ascend the valley sides, transect the divide,
and descend the flanks of the adjacent valley.
Chapter 10 394
Eskers show dramatic inconsistencies in dimension. They range from 2 m

to more than 200 m in height, from several meters to as much as 3 km in
width, and from tens of meters up to 500 km in length (if gaps are considered
in the total distance). In cross-profile, they usually have rather sharp crests
and steeply inclined sides (up to 30°), but broad eskers can maintain rather
flat upper surfaces or may be pitted by kettle development. R. J. Price (1973)
suggests that the height and width of eskers may be directly related to their
overall length, longer ridges being proportionately higher and wider than
shorter ones.
It seems certain that eskers result from sediment accumulation in a va-
riety of openings, such as (1) ice channel fillings (crevasse fillings, fillings be-
tween stagnant blocks, etc.), (2) tunnels beneath or within the ice, (3) su-
praglacial channels or even, in rare cases, (4) narrow longitudinal embay-

ments of the ice front (Cheel 1982). The subglacial origin is made possible by
meltwater that descends from the ice surface to the base through fractures
and holes in the ice. Along the subglacial surface a network of interconnected
tunnels passes water and sediment toward the terminus, and normal fluvial
deposition fills or partially fills the openings. Filling probably takes place during
the last stage of deglaciation when the ice is stagnant and thin, for basal tun-
nels could not remain open if the glacier were still moving or if the ice were
thick enough to cause pressure flow (Flint 1971). In addition to subglacial
channels, field studies and photographic analyses of esker formation by the
Casement Glacier in Alaska (Price 1966; Petrie and Price 1966) and the Brei-
damerkurjokull Glacier in Iceland (Price 1969) provide rather conclusive evi-
dence that englacial and supraglacial deposits can be transformed into eskers.
Figure 10.19.
Esker near Whitewater, Wis
'W
N
.
-*.
In those areas, some of the eskers are definitely ice-cored and were measurably
lowered in elevation during the period 1948-1963 by wastage of the buried
ice. Presumably the final melting of the ice will rest the esker on the original
subglacial floor.
Interior Ice-Contact Features

Behind the marginal zone in the interior portion of a glacier, the predominant
features are deposited at the base of the ice. There pressure from the overlying
ice either spreads till rather evenly across the ground surface or molds water-
soaked till already in that position into distinct morphologic shapes. Supra-
glacial drift is somewhat rare in the interior zone, but where two valley gla-
ciers join, the debris dragged along their lateral edges may coalesce into me-
dial moraines. Although these moraines appear as striking linear belts on the
surface of the ice (fig. 10.6), they are superficial in that the deposits are shallow.
Not all of these linear features represent former ice margins. Some may have
an interior origin (Small et al. 1979) and sediment sources that are both sur-
ficialand englacial (Eyles and Rogerson 1978). Medial moraines are rarely
preserved on the ground surface because they are let down in the middle por-
tions of valleys where meltwater streams are likely to destroy them. The dom-
inant forms are those deposited from beneath the ice such as ground moraines,
drumlins, and fluted surfaces.
Ground Moraine In contrast to end moraines, ground moraine is distin-
guished by its apparent lack of topographic expression. It is accepted as a

moraine despite its low relief and a complete absence of transverse ridges (Flint
1971) because its surface expression is independent of the topography it covers.
Ground moraine occupies much of the surface in North America and Europe
that was covered by major Pleistocene ice sheets. The moraines usually exist
as smoothly undulating plains, like that in figure 10.20 seldom exceeding
10 m in total relief; they range in size from small areas interspersed among
younger marginal features to regions covering thousands of square kilometers
behind the terminal moraine.
Although in most cases the primary building material of ground moraine
is lodgement till, it is a mistake to consider ground moraine and lodgement
till as synonymous. The deposits from which ground moraine is constructed
also include ablation till and interbeds of fluvioglacial deposits that originate
in the same glacial advance (R. P. Kirkby 1969; Boulton 1972a, 1972b) or in
more than one episode of glaciation.
Fluted Surfaces and Drumlins The monotony of ground moraine topography

is sometimes broken by ridges or elongate hills that manifest the mechanics
functioning at the base of glaciers, especially in temperate ice sheets. The most
common features developed in the subglacial environment, fluted surfaces and
drumlins, appear to be fashioned by moving ice; Flint (1971) refers to them
as streamlined molded forms. Others (Sugden and John 1976) consider them
as special types of moraine ridges that form parallel to the direction of ice
flow. Regardless of classification, it appears certain that the two forms are
Chapter 10 396
Figure 10.20.
Flat till and loess
plain in southern
Gently undulating
Illinois related by their common subglacial origin, their orientation parallel to the
topography on ground moraine
direction of ice flow, and the fact that mobilization of subglacial debris is in-
has been smoothed by influx of
younger loess.
volved in their origins.
Fluted surfaces consist of narrow, regularly spaced, parallel ridges. The
ridges are normally less than 5 m high and several hundred meters long, al-
though individual ridges may be considerably larger. Small ridges are usually
composed of till, and their origin may be related to pressure-squeezing of sat-
urated debris into longitudinal cavities at the base of the ice, a process similar
to that forming some of the minor moraine ridges discussed earlier. There is
no question that till is mobilized and arranged in ridges where cavities open
on the lee side of large boulders (Hoppe and Schytt 1953). In fact, Boulton
(1976) defines flutes in a genetic sense as being formed when deformable
subglacial material is intruded into ice tunnels on the lee side of boulders or
other rigid obstructions. This occurs because unloading in the lee of obstruc-
tions sets up a pressure gradient in the till that causes it to flow into the cavity.
Some larger ridges are composed of material other than till (Lemke 1958;
Gravenor and Meneley 1958), and the surfaces between adjacent ridges are
often noticeably grooved. These characteristics have led many workers to be-
lieve that fluted surfaces are both erosional and depositional in origin. Grav-
enor and Meneley (1958), for example, suggest that alternating high and low
pressure zones in the basal ice produce the unique fluted surface. In their model,
grooving occurs in the material beneath the high-pressure zones, and debris
eroded from there is moved not only downglacier but also upward into regions
of low pressure. Boulton (1976) rejects the notion that fluted surfaces require
periodic pressure distributions in the ice and suggests that the features rep-
resent postdepositional deformation of preexisting materials. As such, they
are neither erosional nor depositional.
Drumlins also are elongated parallel to the direction of ice flow, their long
axes deviating only slightly from the average trend of the glacier movement
(fig. 10.21). These distinctive forms have received attention from a large
number of geomorphologists during the last century, but as yet their origin
has not been fully explained (see Muller 1974 for an excellent review). Drum-
lins have been described as having a plan-view shape that is similar to a Iem-
niscate loop (Chorley 1959) or an ellipsoid (Reed et al. 1962) and in long
profile a form that Flint (1971) likens to an inverted bowl of a spoon (fig.
10.21). The exact shape, however, is probably variable enough that no par-
ticular model will fit all drumlins. In any case, it can be stated that some as-
surance that drumlins are higher and wider near their rear edges and that they
narrow and thin downstream until they merge imperceptibly with the sur-
rounding surface. Drumlins average in size from 1 to 2 km in length and from
400 to 600 m in width, and stand anywhere from 5 to 50 m high; individuals
can be smaller or larger. Their length-width ratio seems to be reasonably con-
sistent, ranging from 2 to 3.5 (Reedet al. 1962; Vernon 1966; Trenhaile 1971),
even though length and width may be controlled by different factors (Mills
1980).
Drumlins display a variable internal composition; many are fabricated
entirely from clay-rich till, but others have obvious cores of solid rock or
preexisting drift that may or may not be stratified. The spatial character and
distribution of where drumlins develop are critical in any hypothesis con-
cerning their origin. First, drumlins rarely exist as individuals but instead
cluster together in fields that are commonly wider than most morainal belts
(Gravenor 1953). The density of drumlins within a field seems to be inversely
related to their sizes; that is, very dense clusters are composed of relatively
small drumlins (Doornkamp and King 1971). Second, many show a
studies
strong probability that drumlins within any field are spaced in a nonrandom
manner; that is, spacing between neighboring individuals is somewhat regular
(Reed et al. 1962; Vernon 1966; Smalley and Unwin 1968; Trenhaile 1971).
Third, most glaciated areas have no drumlin development. Fourth, drumlin
fields seem to be located in zones close to but behind the terminal moraines
that mark the limit of a particular glaciation.
Chapter 10 398
Figure 10.21.
(A) A portion of the drumlm field
located near Weedsport, NY

From the northeast quarter of the
Weedsport, NY, quadrangle
(U S G.S. 15'). Contour interval 20
feet. (B) Drumlm shown in
longitudinal profile. Looking east,

four miles northwest of Kalispell,
Flathead County, Mont
(A)
(B)
The models used to explain drumlin genesis fall into two main groups:
(1) drumlins are erosional features developed when moving ice streamlines
preexisting drift or rock, or (2) drumlins are depositional features formed when
a moving glacier deposits till and molds the material as the ice continues its
forward motion. There is reasonable evidence to support both theories. For
example, when the internal composition of drumlins is stratified drift or bed-
rock that predates the ice advance which the drumlins were formed, it is
in
rather difficult to ignore erosional processes in the drumlin origin (Gravenor

1953; Kupsch 1955; Lemke 1958; Boulton 1979; Whittecar and Mickelson
1979). On the other hand, many drumlins have no cores of older material but
consist entirely of lodgement till of the same age as the surrounding drift.
They often display an internal concentric banding that suggests some pro-
cesses of gradual accretion (Hill 1971). Drumlins such as these can hardly be
doubted as being depositional.
we should probably accept a multiple origin for drum-
In light of all this,
lins, as Embleton and King (1975a) suggest, and not pay undue attention to
hypotheses that utilize one process to the exclusion of all others. Yet it is both-
ersome that drumlins have such unique distributional properties. It seems,
somehow, that any explanation of drumlin development, whether by erosion
or deposition, should have its roots in the characteristics of the glacier and the
subglacial debris.One attempt to integrate all drumlins into a single genetic
model (Smalley 1966; Smalley and Unwin 1968) is based on the fact that
many tills have dilatant characteristics. In simplest terms, dilatant materials
increase in volume (expand) under stress. As stress increases on till at rest,
the debris will resist deformation until the stress produces the loose packing
attained in the dilatant condition. When dilatancy prevails, deformation pro-
ceeds more easily and will continue even if stress is decreasing, until finally
the stress becomes too low to maintain the dilatant property. Conceivably, then,
dilatancy occurs only within certain well-defined upper and lower stress limits.
Smalley and Unwin (1968) suggest that the proper stress conditions for
dilatancy are related to the thickness of the ice mass that exerts pressure on
the underlying till. Where the glacier is thick, the till is highly mobile because
the weight of the ice exceeds the load limit needed to initiate dilatance, and
all the interior basal drift will be carried forward. Near the ice margin, how-
ever, the glacier thins to a point where the stress exerted on the basal load is
less than the minimum required to maintain dilatance, and the material will
revert to a more tightly packed texture, thereby becoming stable and resistant
to further deformation. In essence, this implies that the ratio between stress
exerted by ice thickness and strength of the subglacial material increases in
the upstream direction (total deformation upstream and none at the margin).
Drumlins, therefore, develop in the zone between the thin marginal ice and
the thick interior ice. There the loading stress is less than that required to
initiate dilatancy but greater than the pressure at which the till returns to a
stable form (fig. 10.22). Some material collapses into stable cores in this zone
because of its particle size distribution, but the till that remains dilatant is
accreted and streamlined over the cores.

Chapter 10 400
Boulton (1979) suggests that dilatant subglacial debris is not a prereq-

uisite for drumlin formation. In his model, the shear strength of this debris
increases upglacier because ice pressure is increasing over pore pressure (shear
strength is S=C + {P —P H ) tan 0; see chapter 9). Strength values could also
change by altering other factors (see Smalley 1981 for a review). At some
point a critical value of strength is reached where the drag force exerted by
ice upstream from that point is not enough to deform the subglacial till. Down-
stream, the ratio of ice pressure to pore pressure decreases until the strength
reaches another (lower) critical value, and all subglacial till downstream from
there will readily deform. Drumlins form between the critical strength levels
where some undeformed debris serve as the nuclei needed for drumlin growth.
Other than dilatance, the essential difference between the preceding two
models rests in whether the ratio stress/strength increases or decreases up-
glacier and how subglacial till behaves above and below the critical values of
stress and strength. They both, however, provide an intermediate zone in which
drumlin formation should occur. Therefore, each explains the common dis-
tribution pattern associated with the feature.
Proglacial Features
The large volume of water released from glaciers carries with it a tremendous
quantity of sediment that is deposited in a number of environments beyond
the margin of the ice. This debris usually accumulates in stream channels and
associated fioodplains that, because of their continuous lateral shifting, spread
the sediment into a large plain called a sandur (from Icelandic; plural sandar
or sandurs). Downstream from the sandar, the meltwater streams may empty
into bodies of standing water and construct deltas, beaches, and other geo-
morphic features from the fluvioglacial sediment. These forms do not differ
appreciably from their counterparts developed in normal fluvial, lacustrine,
or marine environments and so will not be discussed here. Sandar are some-
what analogous to alluvial fans, but they are unique in that the hydrology of
streams that form them is controlled by intense seasonal variations in the
melting of ice. In addition, because each glacial advance develops its own re-
lated sandur, the distribution and age of the alluvial surfaces are extremely
helpful in unraveling Pleistocene history. Thus, the origin of sandar and the
geomorphic criteria for recognizing their form in ancient settings deserve our
attention.
Figure 10.22.
Cross section at edge of ice sheet
showing most likely stress
conditions for the formation of
drumlins according to the dilatant
hypothesis (A) Stress level above
that required to initiate dilatance;
all material moving (B) Stress
decreasing but drift is mobile
(C) Stress below level needed to
maintain dilatance
End moraine Drumlin field
V. •»-
The study of sandar began in the nineteenth century in both Europe and
North America. They are now recognized as consisting of two primary types
(Krigstrom 1962). A valley sandur, which originates within well-defined val-
leys, is created by one main river and its anabranches; the entire system rarely
occupies the total valley bottom at any given time. In the United States, valley
sandar have been called valley trains and are usually associated with indi-
vidual mountain glaciers (fig. 10.23). The second type of sandur, called a plain
sandur, differs in that it develops with no lateral constraints but represents the
coalescence of many braided rivers that spread debris across wide areas in the
form of a massive plain. In North America, these are often referred to as out-
wash plains and are usually associated with large ice sheets.
Most sandar are composed of gravel, although the deposits may include
lenses of sand. A general decrease in particle size is sometimes, although not
always, apparent in the downstream direction (Fahnestock 1963; Church
1972). The deflation of very fine material on active sandar that are unvege-
tated helps to produce a coarse-grained surface and simultaneously produces
loess (see chapter 8).
In general, the long profiles of sandar are similar to river profiles, being
concave-up in form and expressible as a simple exponential function (Church
1972). The concavity, however, may not be perfect because of the presence of
linear segments similar to those characterizing alluvial fans. Cross-profiles tend
to be convex-up, but the exact form may be irregular or may slope continu-
ously in one direction. In addition, the cross-profile shape seems to depend on
Figure 10.23.
where it is measured in relation to the ice margin.
Valley sandur near Mt. McKinley,
Alaska
£S»
Chapter 10 402
Krigstrom (1962) has recognized on sandar three distinct zones relative

to the ice front, called the proximal, intermediate, and distal zones, each of
which has different surficial characteristics. The proximal zone, closest to the
ice, is usually transversed by only a few main rivers that flow in well-defined
entrenched channels. These rivers and their deposits may pass continuously
onto the ice mass itself, where the sometimes buries extensive areas of
drift
stagnating ice. As the ice subsequently melts, kettle holes form and the prox-
imal sandur takes on a rough, pitted configuration. These kettled sandar (or
pitted outwash plains) are difficult to place in our framework classification
because they form in the marginal ice-contact environment, but they are con-
tinuous with the surface of the proglacial sandur and develop with the same
original mechanics (described in Price 1973). In addition, on many sandar the
proximal surface stands well above the elevation of rivers emerging from the
ice. Several hypotheses have been suggested to explain why rivers in the prox-
imal zone are so entrenched in their own deposits: (1) the rivers may be re-
grading in response to hydrologic and load characteristics (Fahnestock 1969);

(2) the sediment is supraglacially deposited and left elevated as the ice front
recedes and rivers emerge at a continuously lowered level; or (3) modest in-
cision may be normal in the proximal zone, with the sandur surface simply
representing the high flow level. It is conceivable that each of these interpre-
tations is correct.
In the intermediate zone, the channels become wide and shallow and dis-
tinctly braided, and the entire depositional network shifts its position rapidly
from side to side. This active lateral migration leaves a maze of abandoned
channels with a relief of one or two meters impressed on the surface topog-
raphy of the plain. Commonly the main channel is aggraded to a higher level
than the smaller channels, facilitating rapid changes in the position of the
river.Downstream the system changes gradually into the distal zone, where
channels become so shallow that the rivers may merge into a single sheet of
water during high flow. The flow here commonly feeds deltaic growth when
the river enters a body of standing water. However, the sandur may extend
itselfdownstream by growing over the rear portion of the delta, while the del-
taic front simultaneously progrades (Church 1972).
Sandar originate from the combined effects of a large sediment supply
and the high floods associated with melting ice. Most of the abundant load is
derived from older drift, morainal deposits, and the continual delivery of new
debris to the ablation zone and its release from the ablating ice. The greatest
fluvioglacial work occurs near the ice margin where floods are produced by
summer melting or as jbkulhlaups (sudden release of lake water dammed
within the an Icelandic word pronounced "yokel-lawp"). These floods are
ice;
characterized by rapid and drastic increases in discharge (Church 1972; Waitt

1980).
The bulk of aggradation on a sandur takes place during high-flow events
as channel fills, sandur levee deposits, and overbank sedimentation. Overbank
deposits are more prevalent in the intermediate zone where channels are shal-
lower and interchannel reaches are covered more frequently by floods. Al-
though high flow does initiate pronounced channel scouring, the amount of
aggradation during the peak and waning stages of a flood simply obliterates
the scour channels. Thus, aggradation may be rather rapid. For example, Fah-
nestock (1963) measured a net elevation gain of 0.36 m in a two-year period
on the sandur produced by the Emmons Glacier (Washington).
In general, then, sandar can be considered as transport surfaces that ag-
grade during high flows but are probably eroded and changed in form when
discharge and load are at normal volumes. The seasonal variations in load and
discharge may also be accompanied by changes in the river pattern (Fahnes-
tock 1963). Therefore, the ultimate size and properties of a sandur are prob-
ably related to a quasi-equilibrium condition established by the balance
between meltwater volumes and the quantity and size of the sediment made
available for transportation. The surface will always be a montage of flood
sediments, but the exact topography will change incessantly.
In this chapter we examined the landforms developed by the process of glacial Summary
erosion and deposition. Erosional features range in size from minor embel-
lishments of exposed bedrock to major forms that dominate the landscape.
Minor features such as striations, grooves, roches moutonnees, and friction
cracks are a function of the subglacial mechanics that control abrasion and
plucking. Major erosional landforms develop in two environments. In moun-
tain uplands, the expansion of cirques results in the creation of features such
as aretes and horns that give glaciated mountains their characteristically
rugged appearance. Cirques are created by nivation and rotational sliding of
cirque glaciers. Cirques increase in size by erosional retreat of their walls fa-
cilitated by repeated freezing and thawing or possibly by hydration shattering.
Glaciated valleys are created by large-scale abrasion and plucking that cause
the dominant staircase longitudinal profile and the U-shaped cross-profile.
Deposits associated with glaciation, called drift, consist of either stratified
or nonstratified material. Stratification requires that some sediment be trans-
ported by meltwater after the debris is released from the ice. Nonstratified
drift is deposited directly by the ice. Certain types of depositional landforms
tend to accumulate in particular geographic positions with respect to the ice
front. In the marginal zone, moraines and stagnant ice features are most
common. Mobilization and/or immobilization of subglacial drift at stress/
strength thresholds determines where these depositional zones will occur.
Frozen bed conditions and high pore pressure near the margins of temperate
glaciersmay produce thrusting of subglacial material. In the interior zone
behind the ice margin, ground moraine, fluted surfaces, and drumlins are most
conspicuous. Downstream from the glacial margin, in the proglacial zone, all
drift is and usually accumulates in the form of large plains called
fluvioglacial
sandar. Processes operating in each of these depositional environments have
been discussed with regard to how they might generate the landforms devel-
oped in the specific regions.
Chapter 10 404
Boulton, G. S. 1974. Processes and patterns of glacial erosion. In Glacial

geomorphology, edited by D. R. Coates, pp. 48-87. Proc. 5th Ann.
Symposium, S.U.N.Y., Binghamton.
. 1979. Processes of glacier erosion on different substrata. Jour.
Glaciol. 23:15-36.
Church, M. 1972. Baffin Island sandurs: A study of arctic fluvial processes.
Canada Geol. Survey Bull. 216.
Embleton, C, and King, C. A. M. 1975. Glacial geomorphology. New
York: Halsted Press.
Flint, R. F. 1971. Glacial and Quaternary geology. New York: John Wiley
& Sons.
Goldthwait, R. P., ed. 1971. Till: A symposium. Columbus: Ohio State
Univ. Press.
Lewis, W. V., ed. 1960. Norwegian cirque glaciers. Royal Geogr. Soc. Res.,
Ser. 4.
Moran, S.; Clayton, L.; Hooke, R.; Fenton, M.; and Andriashek, L. 1980.
Glacier-bed landforms of the prairie region of North America. Jour.
Glaciol. 25:457-76.
Muller, E. H. 1974. Origins of drumlins. In Glacial geomorphology, edited
by D. R. Coates, pp. 187-204. Proc. 5th Ann. Symposium, S.U.N.Y,
Binghamton.
Price, R. J. 1973. Glacial and fluvioglacial landforms. New York: Hafner.
Sugden, D., and John, B. 1976. Glaciers and landscape. London: Edward
Arnold Ltd.
Processes
Periglacial
and Landforms
11
I. Introduction B. Patterned Ground
II. Permafrost 1. Classification
A. Definition and Thermal a. Polygons
Characteristics b. Circles
B. Distribution, Thickness, and c. Stripes
Origin 2. Origin
III. Periglacial Processes C. Landforms Associated with
A. Frost Action Mass Movement
l. Frost Wedging 1. Gelifluction Features
2. Frost Heaving and 2. Blockfields
Thrusting 3. Rock Glaciers
3. Frost Cracking V. Environmental and Engineering
B. Mass Movements Considerations
l . Frost Creep A. Building Foundations
2. Solifluction B. Roads and Airfields
(Gelifluction) C. Utilities — Water and
IV. Periglacial Landforms Sewage
A. Landforms Associated with D. Pipelines
Permafrost VI. Summary
1. Ice Wedges and Ice- VII. Suggested Readings
Wedge Polygons
2. Pingos
3. Thermokarst
405
Chapter 11 406
Introduction A group of processes and features called periglacial characterize regions having
extremely cold climates. The term "periglacial" was first used (Lozinski 1912)
to describe the processes operating and the features developed in zones ad-
jacent to ancient or modern ice sheets. Since then, the original connotation of
"near-glacial" has been expanded, and most workers now accept the term as
encompassing all nonglacial phenomena that function in cold climates, even
if glaciers are not present (Dylik 1964; Butzer 1964; Washburn 1973). Wash-
burn 1980 adopts the term "geocryology" to indicate the study of frozen ground
processes and phenomena. However, he suggests continued use of the term
"periglacial" as a descriptive adjective. The periglacial environment is diffi-
cult to define in terms of precise temperature and precipitation values, al-
though several attempts have been made to do so. For example. Peltier (1950)
suggested that average annual from — 15°C to — 1°C
temperatures
(5°F-30°F) and an average annual rainfall between 127 mm and 1.397 mm
would constitute a periglacial morphogenetic region. L. Wilson (1968. 1969)
suggests somewhat different physical limits of — 12°C to 2°C (10°F-35°F)
temperature, and 50 mm to 1,250 mm for the precipitation values.
Another question as to what constitutes a periglacial regime arises be-
cause some scientists believe that permanently frozen ground, or permafrost,
is an essential ingredient (if not a prerequisite) for periglacial conditions (Tri-
cart 1967; Pewe 1969). Certainly some periglacial features occur in close as-
sociation with frozen ground, and a complete understanding of periglacial
systems therefore requires some fundamental perception of permafrost. How-
ever, periglacial processes also function where no permafrost is known to exist.
Most students of periglacial phenomena, then, would not view permafrost as
a requirement of a periglacial system. The two points that all investigators
seem to agree on are that (1) nearby glaciers are not necessary for the pro-
cesses to function, and (2) the fundamental controlling factors of geocryology
are intense frost action and a ground surface that is free of snow cover for
part of the year. Many regions meet these requirements, ranging from polar
to subpolar lowlands to high-elevation mountains that may rise in any regional
climate including temperate or even tropical zones.
It is not new for humans to occupy the high latitude regions of the world;
Eskimos and other nomadic groups have wandered these areas for millenia.
These inhabitants, however, lived in rather simple social organizations, and
their life styles brought them into little conflict with their natural surround-
ings. In more recent times, the emplacement of defense installations in peri-
glacial regions, the discovery of oil in the Arctic Circle, and the increased
exploration for mineral wealth brought with them the realization that devel-
oping a highly technical society in these regions is an engineering and scientific
nightmare. For inhabitants of temperate zones, shattered highways and frozen

or broken water pipes are inconveniences of a rigorous winter and the sub-
sequent spring thaw. Multiply these problems by a thousand; add disruption
of services, construction difficulties, and complications of all the trappings as-
sociated with a modern civilization; and you will begin to understand the
meaning of life in a periglacial environment.
407 Periglacial Processes and Landforms
Governmental agencies of several countries, including the United States

and the Soviet Union conduct research related to the problems of cold cli-
mates, and symposia and reports in scientific journals increasingly treat such
topics. Nonetheless, detailed understanding of the processes involved is lacking,
and faced with an inevitable human migration into cold regions, it is imper-
ative that geomorphologists devote additional efforts to this discipline. We have
the basics to do so because the processes responsible for periglacial effects are
not greatly different from thosewe already know. In fact, frost action and
mass movements form the core of periglacial processes. The main difference
between periglacial and temperate phenomena seems to lie in the magnitude
of the processes and the manner in which they are combined in the systemic
operations. Thus, as L. W. Price (1972) stresses, if we are ever completely to
understand the periglacial environment and provide viable information for fu-
ture planning, we must cast aside our provincial, mid-latitude approaches to
the study of this system.
Our treatment of geocryology will be necessarily brief, examining only
the major processes and features. For greater detail, several excellent texts
and reviews are available: Embleton and King 1968, 1975b; Price 1972; Wash-
burn 1973, 1980; Tricart 1969; Pewe 1969; National Research Council of
Canada 1978; National Academy of Sciences 1983.
Definition and Thermal Characteristics Permafrost

Although permafrost is not a requirement for the functioning of periglacial
processes, the most troublesome engineering problems occur in regions un-
derlain by zones of perennially frozen ground, and many periglacial features
are related to abundant ground ice. Permafrost was originally defined in terms
of temperature only (Muller 1947) as being soil or rock that remained below
0°C continuously for more than two years. This definition has been accepted
in many subsequent discussions of the topic (Washburn 1980). By this defi-
nition water is not a necessary component of permafrost, and in fact, "dry"
permafrost has been recognized. In most cases, however, ice is so important
in the mechanics of periglacial processes that some workers (Stearns 1966)
believe moisturemust be present before any material can be considered as true
permafrost. Other investigators have placed the temperature criterion for
permafrost below 0°C (Ferrians 1965; Brown 1970). In a geomorphic sense
the presence of ice, regardless of the temperature value, seems to be critical.
The ice may exist as a cement between soil particles or as larger masses of
pure ice. Where distinct ice masses are prevalent, they usually occur as hor-
izontal lenses, but they may also fill cracks in the parent material and stand
as vertical veins or wedges (fig. 11.1).
Figure 11.2 shows that permafrost is usually a subsurface phenomenon

existing as a zone of permanently frozen ground that can extend downward
to incredible depths. The upper surface of the permafrost, the permafrost table,
is overlain by a thin layer (15 cm to 5 m thick) of material that freezes and
thaws on a seasonal basis. This uppermost layer, the active layer, is thickest
mi
rl Figure 11.1.
*| Ice wedge (groun
(ground ice) in
MBfeCj permafrost expos
exposed by placer
mining near Liven
Livengood, about 50
5 miles northwest o ot Fairbanks.
B Alaska Tolovana district, Yukon
J^l^^ Sept 194$
region, Sept, 1949
409 Periglacial Processes and Land forms
Figure 11.2.
a region underlain by
Profile in
Suprapermafrost layer (ground
permafrost. Taliks are zones of
above permafrost table unfrozen ground within or beneath
including taliks, if present) the permafrost or between the
permafrost table and the base
of the active layer (After
Ferrians et al. 1969)
Permafrost table (upper
surface of permafrost)
Talik-
in the subarctic region (Price 1972), becoming shallower toward both the pole
and the mid-latitudes. In any area, however, significant variations in active
layer thickness can occur in short distances depending on environmental con-
ditions (Owens and Harper 1977), and it is usually thicker in sands and gravels
than in more fine-grained soil materials. The mechanics in the active layer is
similar in most respects to the normal seasonal freezing and thawing in tem-
perate climate zones. In permafrost regions, however, water released by thawing
cannot percolate into the solidly frozen substrate, and this tends to accentuate
the effects of frost action and mass movements.
Even within the permafrost itself, the temperature fluctuates with the sea-
sons.Temperature variations, however, become less radical with depth until
some level is reached, generally at 20 to 30 m deep, where the temperature
never changes (fig. 1 1.3). This level is known as the zero annual amplitude.
The thermal operations in permafrost are affected by so many factors that
they are understood only in general terms. For example, thermal conductivity
varies with material composition and may cause freezing or thawing to pro-
gress unevenly from the surface downward. Thus, as freezing proceeds after
a thaw season, unfrozen lenses may develop between the permafrost table and
the solidly frozen ground surface. This phenomenon causes unusual pressures
in the soil water and depresses its freezing point. In addition, heat of fusion
is generated when the zone around the ice-free lenses freezes. The result is
that pockets of free water may exist in the ground for considerable time after
the surface is frozen, possibly for several months.
Further complications in the distribution of frozen ground occur because
the permafrost table tends to mirror the surface topography, rising beneath
and lowering under valleys. In addition, the permafrost table is influenced
hills
in variousways by the position of surface water (fig. 11.4), depending on

whether the rivers or lakes freeze over in the winter (Ferrians et al. 1969).
Distribution, Thickness, and Origin

Where annual temperatures average 0° 0°C, ground freezing during
or below
the winter will penetrate deeper than the depth of summer thawing. Each
passing year will produce another increment of perpetually frozen ground, and
C
Chapter 11 410
Mean annual ground temperature

Figure 11.3.
Ground temperature change with
depth in permafrost regions.
-20 +20'
(After J. R. Williams 1970) i
Permafrost table Maximum annual

temperature
Minimum annual
temperature —Level of zero
annual amplitude
Base of permafrost
Small lake Large lake

Figure 11.4.
Schematic cross section showing Large river
the effect of surface water on the
distribution of permafrost. (PF is
permafrost; T are taliks.)
the permafrost zone will gradually thicken. The process, however, cannot go
on indefinitely because the Earth's thermal regime will exert controls on the
depth to which freezing can penetrate. Heat affecting permafrost issues from
the sun and from the interior of the Earth as geothermal heat flow. Therefore,
cold filtering downward from the surface is counteracted by heat escaping
upward from inside the Earth. The interaction results in the geothermal gra-
i.e., the rate of temperature increase with depth below the surface. Thus,
dient,
under a constant climate, at some depth the temperature will be kept above
0°C by the geothermal heat flow (fig. 1 .5), and the total thickness of perma-
1
frost will relate to the position of this temperature level.
The geothermal gradient, and therefore thickness of permafrost, can be

quite variable under areas having identical surficial climates (fig. 1 1.5). This
occurs because the geothermal gradient is affected by the thermal conduc-
tivity of the parent material. Where mean surficial temperatures are the same,
permafrost will extend to greater depths in materials of higher conductivity.
Temperature (°C)
Figure 11.5.
10 4 -2 +2 Generalized profiles of measured
temperature on the Alaskan arctic
coast (solid lines) Dashed lines
represent extrapolations
100
J
200
300 V
400
Q.
Q
500
Cape
Barrow •Simpson
• Prudhoe Bay
600 • Cape
Thompson
700 Northern Alaska
The variations in geothermal gradients along northern Alaska, shown in figure

1 1.5, are explained by this phenomenon.
Permafrost today may extend to formidable depths (table 11.1). It av-
erages between 245 and 356 m in North America and tends to be slightly
thicker in Eurasia. The thickest known permafrost is in Siberia, where a depth
of approximately 1500 m has been reported. The great thicknesses of perma-
frost probably reflect both past and present climate conditions.
Several lines of evidence indicate that some permafrost must have orig-
inated during the Pleistocene (see Washburn 1980, pp. 60-61), especially in
areas that were not glaciated. In areas covered by Pleistocene glaciers, perma-
frost probably formed after the ice dissipated. In fact, in polar regions perma-
frost is forming today where glacial retreat has exposed unfrozen ground
(Washburn 1980). This indicates that some present-day climates are cold
enough to accumulate modern permafrost and certainly maintain ancient
permafrost. Permafrost does, however, respond to climatic change (Mackay
1975b). For example, in the upper parts of the temperature profiles from
northern Alaska (fig. 1 1.5), the distinct curvatures shown are caused by cli-
matic warming after the permafrost was formed. The equilibrium geothermal
gradient is shown by the straight parts of the curves, and their upward pro-
jection indicates the prevailing surficial temperatures at the time of perma-
frost formation (Lachenbruch and Marshall 1969; Gold and Lachenbruch
1973).
Table 11.1 Selected Northern Hemisphere permafrost thicknesses
Location Mean Annual Thickness of

Air Temperature Permafrost
Alaska
Prudhoe Bay (70°N, 148°W) -7 toO°C 609 m (2000 ft)
(20 to 32° F)
Barrow (71 °N, 157°W) -12to7°C 405 ma (1330 ft) 16 km
(10 to20°F) (10 miles) inland
Umiat (69°N, 152°W) -12 to -7°C 322 m (1055 ft)
(10to20°F) 235 m (770 ft) under
Colville River
Cape Thompson (68°N. 166°W) -12 to -7°C 306 ma (1000 ft)
(10to20°F)
Bethel (60°N. 161 °W) -7 toO°C 184 m (603 ft)
(20 to 32°F) 13 m (42 ft) under
Kuskokwim River
Ft Yukon (66°N, 145°W) -7 toO°C 119 m (390 ft)
(20 to 32°F) 5.5 m (18 ft) under
Yukon River
Fairbanks (64° N, 147°W) -7 toO°C 81 m (265 ft)
(20 to 32 °F)
Kotzebue(67°N. 162°W) -7 to0°C 73 m (238 ft)
(20 to 32°F)
Nome(64°N, 165°W) -12 to -7°C 37 m (120 ft)
(10 to20°F)
McKinley Natl Park-East extremely variable 30 m (100 ft)
Side (64°N, 149°W)
Canada
Melville Island. N.W.T 548 m (1800 ft) near coast
(75°N, 111 °W) probably thicker
mtenorward
Resolute. N.W.T. -162°C(2.8°F) 396 m (1300 ft)
(75°N, 95°W)
Port Radium. N WT -7 1°C(192°F) 106 m (350 ft)
(66°N. 118°W)
Ft Simpson, N.W.T. -3 9°C (25.0°F) 91 m (300 ft)
(61 °N, 121°W)

Yellowknife, N.W.T -54°C(22.2°F) 61-91 m (200-300 ft)
(62°N, 114°W)
Schefferville, PQ. -4.5°C (23.9°F) 76 m (250 ft)
(54°N, 67°W)
Dawson, YT (64°N, 139°W) -4.6°C (23.6°F) 61 m (200 ft)
Norman Wells, N.W.T. -6.2°C (20.8°F) 46-61 m (150-200 ft)
(65°N. 127°W)
Churchill, Man (58°N. 94°W) -7 1°C(192°F) 30-61 m (100-200 ft)
U.S.S.R.
Upper Reaches of Markha 1500 m (4920 ft)
River (66°N. 111°E)

Udokan (57°N, 120°E) -12°C(104°F) 900 m a (2950 ft)
Bakhynay (66°N. 124°E) -12°C(10.4°F) 650 m (2130 ft)
lsksi(71°N, 129°E) -14°C(68°F) 630 m (2070 ft)
Mirnyy (63°N, 114°E) -9°C(158°F) 550 m (1805 ft)
Usf-Port (69°N. 84°E) -11 C(122°F) 425 m (1395 ft)
Salekhard (67°N. 67°E) -7°C (194°F) 350 m (1150 ft)

Norilsk (69°N. 88°E) -8°C(17 6°F) 325 m (1070 ft)
Yakutsk (62°N. 129°E) 195-250 m (650-820 ft)
Vorkuta (67°N, 64°E) 130 m (430 ft)
Association of American Geographers Commission on College Geoc-

'
L W Price
*A calculated depth not actually measured
Figure 11.6.
Extent of continuous and
discontinuous permafrost in the
Northern Hemisphere. (After
Fernans et al. 1969)
Continuous permafrost
/
^Discontinuous permafrost
The destruction and/or protection of permafrost varies with factors other

than climate, such as the type of surface cover. The presence of ice or dense
vegetation has a decided influence on permafrost thickness. In Antarctica, for
example, some areas beneath the ice sheet may have no permafrost (Ueta and
Garfield 1968), while zones that are glacier free have permafrost up to 150 m
thick (J. R. Williams 1970).
Permafrost underlies 26 percent of the world's land surface (Black 1954)
and so is not, as we may tend to think, an unusual phenomenon. In fact, it is
known to exist beneath the ocean in many nearshore polar areas, although it
probably formed on land and was submerged during a subsequent rise in sea
level. As the map in figure 1 1.6 shows, most of the known permafrost exists
in the polar regions of the Northern Hemisphere, extending to a southernmost
limit at a latitude of approximately 55°N in both North America and Eurasia.
In China, however, latitudinally controlled permafrost is known to exist as far
south as 46°N. Permafrost in these regions is usually divided into continuous
Chapter 11 414
and discontinuous types (Ray 195 1). Continuous permafrost usually consists
of thick layers of perennially frozen ground that are spread rather evenly under
a wide areal surface. The continuous nature of the permafrost alters only where
it thins under deep, wide lakes or rivers (fig. 1 1.4). In contrast, discontinuous
permafrost is shallower and contains unfrozen zones within the frozen ground
or wide gaps that remain unfrozen (fig. 11.7). These unfrozen areas, called
taliks, increase in size and number southward until true permafrost exists only
as isolated patches. Taliks appear as islands exposed at the surface, lenses or
layers within the permafrost, or unfrozen ground beneath the permafrost (figs.
%Wi 1 1.2,
way with
the -1°C
1 1.6). The southern
the — 6°C
limit of continuous permafrost coincides in a general
annual isotherm and the discontinuous boundary with
isotherm (Brown 1970).
It should not be assumed from the above that permafrost exists only in
high latitudes. Isolated zones of frozen ground, called sporadic permafrost (see
fig. 1 1 .7), can be found far south of latitude 55°N and are probably relicts of
a once colder climate. Pockets of permafrost persist also where elevations are
sufficiently high. In the United States, for example, permafrost is found near
the summit of Mt. Washington in New Hampshire (Goldthwait 1969) and in
the Rocky Mountains above the tree line (Ives and Fahey 1971). Permafrost
has even been reported near the summit of Mauna Kea, Hawaii (Woodcock
et al. 1970; Woodcock 1974), at an elevation of 4170 m. In some areas (for
example, the high-altitude plateaus of China), permafrost is so widespread

that it is appropriately considered as being continuous or discontinuous rather
Figure 11.7. than sporadic.
Types of permafrost zones:
The general but imprecise relationship between average air temperature
(A) continuous, (B) discontinuous,
(C) sporadic. AL = active layer;
and permafrost boundaries suggests that modern climate determines the dis-
PF = permafrost, T = talik. tribution of frozen ground (Harris 1983) even though most of the included ice
formed at an earlier time. This conclusion fits the interpretation that the re-
cent retreat of the discontinuous permafrost boundary in Manitoba (Canada)
is related to a climatic shift that began only about 120 years ago (Thie 1974).
As Washburn (1980) suggests, this probably indicates that discontinuous
permafrost may be in such a delicate equilibrium with the present climate that
only minor changes in climate or surface condition will produce drastic effects.
All other things being equal, the climate most amenable to preservation
of permafrost is one with cold, long winters followed by cool, short summers
and low precipitation in all seasons (Muller 1947). Other variables, such as
vegetation type and density, composition of surface materials, topography, and
surface water, exert an influence on the spatial character of permafrost. For
example, most permafrost lies beneath the northern boreal forests or, north of
the tree line, under a tundra vegetation dominated by low sedges, grasses, and
mosses. The boundary between continuous and discontinuous permafrost often
parallels the southern limit of the tundra. Because the thermal properties of
vegetation control how efficiently temperature can penetrate the ground (Corte
1969; Price 1972), not clear whether the climate or the vegetation pro-
it is
duced by the climate is the determining factor of permafrost distribution.

Probably all factors exert some control, and therefore we cannot expect a pre-
cise correlation between air temperature and permafrost boundaries.
41 5 Periglacial Processes and Landforms
Among students of periglacial phenomena, the consensus is that the regime Periglacial Processes
isdominated by the processes of frost action and mass movement. Although
these were discussed in an earlier chapter, an examination of their prowess in
the periglacial setting can give a different perception and a greater appreci-
ation of their geomorphic significance. Frost action encompasses a group of
processes —wedging, heaving, thrusting, cracking — that all serve to prepare
bedrock or soil for erosion. Mass movments transport the loosened debris. The
two groups of processes function in periglacial environments as they do in tem-
perate zones. As suggested earlier, however, major differences arise because
frost action is considerably more severe in periglacial zones, and because mass
movements may be intensified during thawing because the material is satu-
rated with excessive water that cannot drain downward through the system.
The combination of these factors results in geomorphic features that are unique
to periglacial regions.
Frost Action
The driving force in all processes included in the realm of frost action is the
growth of ice within a soil or rock. Intuitively we might expect the explanation
of freezing in a porous substance to be relatively simple, but in fact, the ther-
modynamics of the process are very complex (see Miller 1966; Anderson 1968,
1970). In addition, other factors within the system influence the final geo-
morphic effect of frost action because they respond to freezing in different
ways. For example, segregation of ice into discrete lenses generates a stress
from that produced by freezing of disseminated porewater,
field that differs
and the resulting processes also are dissimilar. Whether ice lenses actually
form depends mainly on ( 1 ) the rate of freezing, (2) the ability of water to be
drawn to a central freezing plane or point, and (3) the size of the pore spaces.
Generally, lenses of ice are smaller and less common as depth increases
because the greater pressure lowers the freezing point. However, there are
exceptions to even this generalization when the soil properties are conducive
to ice-lens development. The water in fine-grained sediment (small pore spaces)
freezes at lower temperatures and also has a greater propensity to suck water
to a central freezing plane. Because of these factors, along with the tendency
of fines to hold more water, segregated ice masses are most common in fine-
grained sediment, and coarse sediment usually contains interstitial ice. On the
other hand, extremely fine-grained sediment may be impermeable. Therefore,
ice lenses form preferentially where suction potential and permeability are
most advantageously combined, and this usually occurs in deposits of unce-
mented silt (discussed in Washburn 1980).
In detail, the movement of water to a freezing plane is more complicated
than the foregoing discussion might suggest and depends on many factors other
than grain size (Konrad and Morgenstern 1983; Rieke et al. 1983). As a re-
sult, Konrad and Morgenstern have identified a parameter called the segre-
gation potential to indicate whether a soil will readily form ice lenses (see
Konrad and Morgenstern 1983 for discussion and references). The segrega-
tion potential is the ratio of the rate of water migration to the temperature
Chapter 11 416
gradient near the frost front. It assimilates all of the controlling factors and
provides an index that can be applied in predicting the susceptibility of soils
to frost-related problems.
The expansion associated with freezing exerts pressures that can produce
results as variable as shattering of solid rock or physical lifting of the ground
surface. Precisely what event will transpire during freezing and subsequent
thawing is determined by a multitude of complex interdependent variables
that operate within the system.
Frost Wedging Frost wedging is the prying apart of solid material by ice
(Washburn 1973). It is synonymous with splitting, riving, and shattering. The
precise mechanisms of frost shattering are extremely complex (for a discus-
sion, see McGreevey 1981) and well beyond our introductory treatment. Suf-
fice it to say that shattering will occur when the stress generated by changing
water into ice exceeds the tensile strength of rocks. Exactly what conditions
(rate of freezing and absolute temperatures) will result in shattering are quite
variable because the process is also controlled by the internal properties of the
rocks (Douglas et al. 1983). Most evidence for the wedging process has been
derived from laboratory experiments or is based on theoretical models. It is
doubtful that these are directly applicable to the field situation; in fact. Thorn
(1979) found little correlation between the experimental work and measure-
ments of shattering in the natural setting. Furthermore, he suggests the pos-
sibility that porosity, saturation in the rocks, and freezing intensity may be
combined to generate a number of disruptive mechanisms. This tends to rein-
force the idea that hydration shattering (White 1976a; see chapter 10) may
be an important partner wedging mechanics.
in frost
In porous substances that are saturated, frost wedging is facilitated by

rapid freezing of water in the near-surface pores. This tends to close the system,
allowing the buildup of some extra pressure that is needed to cause shattering.
Slow freezing seems to be an effective wedging process only if water is free to
migrate to a freezing plane where excess pressure is generated by growth of
large ice crystals. As discussed in chapter 10, wedging in cracks takes place
only if freezing proceeds from the surface downward (Battle 1960).
Other factors influence wedging besides the rate of freezing and the
frost
closure of the system. It is known, for example, that a high water content
increases the strain on rocks (Mellor 1970) and that most rocks shatter more
readily if they are immersed in the fluid (Potts 1970). This probably explains
why wedging is often more pronounced at the base of sheer cliffs where
groundwater is available than at the top.
Given an abundant supply of water, characteristics of parent material will
control the extent of wedging. For example, larger pore size and porosity in a
rock will increase its susceptibility to wedging, and sedimentary rocks with
fissility of micaceous minerals generally allow easier migration of water and
promote greater wedging. In addition, the effect of frost wedging seems to be

linked to the number of freeze-thaw cycles. Some caution is required, however,

in correlating freeze-thaw frequency and wedging because air temperature is
not always a good predictor of rock surface temperature or the temperature
within cracks in those rocks (Douglas et al. 1983). Furthermore, a knowledge
of rock properties is important because water may freeze at a temperature
other than 0°C depending upon those characteristics. Thus, prolonged freezing
with extremely low temperature may have more geomorphic significance than
frequent freeze-thaw cycles (Rapp 1960; Washburn 1973).
The products of frost wedging are notably angular and range in size from
blocks as large as buildings to fine-grained debris. It has been traditionally
accepted that the terminal size of frost shattering (Hopkins and Sigafoos
is silt
1951; Taber 1953). Although this is probably true in most situations, some
workers now believe that clay-sized particles may be formed under certain
conditions (McDowall 1960) or that any terminal size is only rarely attained
(Potts 1970). In any case, coarse angular debris is common in periglacial re-
gions, accounting for the prevalence of talus rubble at the base of alpine slopes.
Wedging also breaks particles loose from bedrock covered by a soil, and these
progressively make their way to the surface, creating additional problems for
farmers working the land.
Frost Heaving and Thrusting Frost heaving refers to vertical displacement

of matter in response to freezing, while frost thrusting connotes horizontal
movement (Eakin 1916). In the natural setting the two processes are virtually
indistinguishable, and their designation as separate processes is probably more
imagined than real. Heaving is directly responsible for several phenomena that
are commonplace and accentuated in periglacial environments. First, heaving
causes the ground surface to move vertically as ice formation expands the
ground material. The extent of this displacement depends on the physical vari-
ables within the system (Rieke et al. 1983), and the surfaces of adjacent local
areas may be lifted at disparate velocities and amounts. Such differential
heaving causes building foundations and other types of construction to crack.
Second, heaving has the ability to sort heterogeneous debris by forcing larger
particles to migrate surfaceward relative to their finer counterparts. The par-
ticles may move upward as much as 5 cm a year (Price 1972). It is well to
recognize that any object inserted into the ground or dispersed within the near-
surface mass is subject to the mechanics of heaving — not only stones of var-
ious size that emerge farmlands or on slopes but also fenceposts, telephone
in
poles, pilings, or anything else that people might conceivably shove into the
ground.
The heaving process also functions in indurated solids and is capable of
displacing joint-bounded blocks of bedrock, even though these are held rather
tightly by the surrounding mass. These upheaved blocks may project up to
1.5 m above the ground and, where soil is absent, give the surface a jagged
appearance.
Chapter 11 418
In several classic studies of heaving mechanics, Taber (1929, 1930) showed

that heaving in a closed system is limited to that produced by the 9 percent
volume increase caused when water freezes. In open systems, heaving is much
greater because additional force is gained from crystal growth along the
The amount of extra heaving depends on how much water can
freezing plane.
be drawn to the point of freezing from the surrounding material, a function
of the segregation potential. Taber also showed that heaving stress due to crystal
growth is exerted normal to the freezing isotherm and not in the direction of
least resistance. Larger particles of varying compositions will have different
thermal conductivities and thus may locally influence the orientation of the
S = frozen silt I = ice layers cooling surface. This certainly introduces lateral movements in the expansion
V= voids that are not perpendicular to the freezing isotherm. Some workers also dis-
agree with Taber's conclusion that differences in resistance to expansion do
Figure 11.8. not influence the direction of the heave (Beskow 1947).
Displacement of stones during the
freezing process (After Taber
Any theory attempting to explain why large particles are moved surface-
1943 Courtesy of The Geological ward in a nonsorted soil mass must consider not only the forces that lift the
Society of America) clast but also the reasons does not return to
it its original position during con-
traction. Details of the two plausible theories concerning heaving mechanics,
called frost pull and frost push, are discussed by Washburn (1973). Briefly,
the frost-pull hypothesis suggests that stones are lifted vertically along with
the fines when the ground expands during freezing. The fine sediment, being
more cohesive, is brought downward quickly upon thawing and collapses around
the larger clasts while the bases of the stones are still frozen. Cavities formed
when a large particle is heaved (fig. 1 1.8) may also be partly closed by thrusting
into this zone of lowered resistance.
The frost-push hypothesis is based on the fact that individual stones are
better conductors of heat than porous soil. As a result, stones will cool more
quickly, and the form along the freezing plane will be adjacent to
first ice to
and at the base of the stones, thereby pushing them upward. The phenomenon
of ice forming preferentially near stones embedded in a silty soil has been
observed and is not conjecture; nor is it conjecture that thawing occurs first
around the stones. Presumably, however, material thawed adjacent to the top
of the stone will collapse against the upper part of the clast while the base
remains frozen and thus prevent its return to the original position. The shape
of the stone may become important in this procedure; for instance, wedge-
shaped particles with the narrow edge projecting downward will have greater
difficulty returning to their original level.
Evidence exists to support both the frost-pull and frost-push theories, and
probably both processes function simultaneously in a heaving environment.
Washburn feels, however, that rapid heaving or movement that breaks a veg-
etal cover probably requires that frost push be dominant. Slow ejection of clasts
can probably be accomplished by frost pull with little necessity for rapid ice
buildup beneath the stones.
Other factors magnitude of frost heave and its continuation in
affect the
the upper part of the Experiments conducted in Greenland by Washburn
soil.
(1967) showed that moisture content, vegetation, and depth are all critical
variables in the heaving process. The greatest heave occurred in zones that
41 9 Periglacial Processes and Landforms
had abundant moisture, and deeper objects in an active layer were displaced
farther than shallow ones. However, if objects are inserted into the permafrost
itself, heaving becomes negligible. In addition, ice tensing and heave are prob-
ably dependent on soil properties (thermal conductivity and grain size) and
external factors other than surface temperature, such as overburden pressure
(Gilpin 1980). Heaving is generally less when the vegetation cover is dense,
presumably because of the insulation provided by the vegetal mat. At or near
the surface another form of ice, called needle ice (or piprake), aids in the
lifting process. Needle ice consists of groups of slender ice crystals that are
usually 1-3 cm long, although they have been observed to attain lengths up
to 40 cm (Troll 1958). The ice clusters are capable of lifting stones up to cobble
size. Needle ice, which forms best in moist, loamy soils with no vegetation,
also is very important as an aid in mass movement.
Frost Cracking Frost cracking is the development of fractures at very low

temperatures. The process seems to function best in permafrost regions, al-
though it has been reported from other environments. Frost cracking is usually
considered as a frost action phenomenon, but Price (1972) points out that in
fact it is different because it results from thermal contraction rather than ex-
pansion associated with freezing.
At very low temperatures frozen ground often evolves into a polygonal
network of contraction fractures, but cracking may depend more on the rate
of cooling than on the absolute temperature value at the moment of fracture
(Lachenbruch 1966; Black 1963, 1969). Initial cracking at the surface may
m and can become progressively wider and deeper. The
extend to a depth of 3
results of frost cracking will be discussed more in a later section when we
examine polygonal ground features.
Mass Movements
Any variety of mass movement (discussed in chapter 4) can occur in a peri-
glacial environment, but there is little doubt that two types, frost creep and
solifluction, dominate the cold regime. The movement of debris occurs si-
multaneously with the frost action processes, and it is questionable how ef-
fective the erosion would be if it operated alone. Mass movements, however,
are responsible for a variety of landforms that are unique to periglacial re-
gions.
Frost Creep Frost creep, like any form of near-surface creep, is the down-
slope movement of particles in response to expansion and contraction and under
the influence of gravity. It is unique only because freezing and thawing gen-
erate the cycle of expanding and contracting. As explained earlier, the freezing
front in the soil usually parallels the ground surface. Individual soil particles
are heaved perpendicular to the surface, but during thawing they are affected
only by gravity and should contract in a vertical direction (shown earlier in
fig. 4.20). Actually, a component of upslope migration has been observed in
the contracting phase (Washburn 1967; Benedict 1970). This motion, called
Chapter 11 420
Figure 11.9.
Route of downslope movement
followed by a particle under
combined frost creep and
solifluction
retrograde movement, probably stems from the attraction of fine-grained par-

ticles for one another and the cohesive strength developed between them. On
the other hand, some downslope flowage may occur during the thaw, and so
the overall route followed by any soil particle involved in frost creep probably
resembles the path shown in figure 1 1.9.
Detailed studies in northeast Greenland (Washburn 1967) showed frost

creep to be the dominant process where soils are silty and slopes stand between
10° and 14°. It was noted, however, that in any given year other processes
could dominate the system. Therefore, it may not be safe to generalize about
the conditions that promote frost creep because soil texture, moisture content,
vegetation, and freeze-thaw frequency all have some bearing on the efficacy
of the process.
Solifluction (Gelifluction) The term solifluction was first proposed to describe

the action of slow flowage in saturated soils. The original meaning carried with
it no climatic restrictions, but over the years the term has generally been used
to connote a process that functions in periglacial regions. We will follow this
practice here, although you should realize that it is technically incorrect be-
cause solifluction can occur anywhere. The term gelifluction (Baulig 1957)
refers to soil flowage associated with frozen ground and as such is a specific
type of solifluction.
Solifluction is most dramatic in permafrost zones because water released
in the active layer during thawing cannot penetrate below the permafrost table.
Soils in the summer are often saturated, and the concomitant loss of friction
and cohesion causes them to behave like viscous fluids. Such materials can
flow on slopes as gentle as 1°, but maximum
displacement seems to occur
where the gradient is between 5° and 20°. Soils on slopes steeper than 20°
tend to drain easily, and water escapes as surface runoff.
It is clear that abundant moisture is the overriding factor in the solifluc-
tion process (Washburn 1967; Chambers 1970; Price 1972). Manx students
of periglacial geomorphology feel that significant flow can occur only when
the moisture content equals or exceeds the liquid limit, but this interpretation
Table 11.2 Representative values of combined frost creep and gelifluction. a
Rate (cm/yr) Slope Source
2 15° Rapp 1960

0.9-3.7 10°-14° Washburn 1967
1.0-3.0 3°-4° Jahn 1960
5.0-120 7°-15° Jahn 1960
10 20° P. J. Williams 1966°
2 5 6°-13° Benedict 1970
26 2 5°-4° French 1974
"tended as random sampling ana . j~ility exists depending on differences in ground

illation depth slope face direction gram size and soil texture etc
"Pebbles on bare slope surface
cannot be accepted unequivocally because solifluction has been observed at

lower moisture contents (Fitze 1971). Obviously, other factors such as grain
size, slope angle, and vegetation play a modifying role on a local basis (Harris,
1973). For example, gravel and coarse sands are so permeable and easily
drained that they virtually never flow. Clays, on the other hand, are extremely
cohesive. Thus, silty soils, especially those with a bimodal size distribution,
are most susceptible to solifluction. Nonetheless, the overwhelming impor-
tance of water content was clearly demonstrated in Greenland (Washburn
1967), where the highest rates of movement commonly occur on the most
densely vegetated areas of the slopes and at gradients that are lower than the
dry portions of the slopes. This indicates that moisture can overcome both the
binding effect of vegetation and the apparent deficiency of a transporting gra-
dient.
The rates of combined frost creep and gelifluction in northeast Greenland
and other regions arc shown in table 1 .2. It is difficult to ascertain how much
1
of the total movement is accomplished by frost creep and how much can be
attributed solely to flow.
Landforms Associated with Permafrost Periglacial Landforms

Some periglacial features are so closely allied to the distribution of permafrost
that a genetic relationship can hardly be denied, and observation of these fea-
tures preserved in temperate regions affords the best evidence known for re-
constructing the areal extent of ancient permafrost boundaries. The basic

processes involved in the development of these features are no different from
those that form other periglacial landforms; that is, they rely on frost action
and/or mass movement. The permafrost layer, however, places a distinctive
imprint on the system and allows us to consider the geomorphic features as a
separate group.
Ice Wedges and Ice-Wedge Polygons When thermal tension exceeds the
strength of the surface materials, frost cracking begins and fractures pene-
trate into the active layer and the upper portion of the permafrost. At first the
cracks are only millimeters wide, but they may extend vertically downward
to depths of several meters. The actual moment of cracking seems to occur
Chapter 11 422
sometime between January and March when temperatures in the ground reach
their annual minima (Mackay 1974) and the rate of temperature drop may
be extreme. During spring and early summer the crack may be partially or
totally filled with snow that has filtered down from the surface and with water
released by the onset of thawing in the active layer. As the diagrams in figure
1 1 .9 show, this mixture freezes below the permafrost table and occupies the
initial crack as an ice veinlet. Because the initial fracture now represents a
zone of weakness within the permafrost, subsequent cold winters produce

cracking along the trace of the original opening, and new ice is added as be-
fore. Over the years, this accumulation of relatively pure and vertically ori-
ented ice takes the form of a downward-tapering wedge, called an ice wedge.
shown in the diagrams (fig. 11.10) and pictured in figure 11.1. In dry perma-
frost, the material from the contraction fissures may be sediment, especially
wind-transported sand or loess associated with extreme aridity (for example,

see Carter 1983). The resulting feature is commonly called a sand wedge (Pewe
1959).
Ice wedges range from 1 cm to 3 m wide and from 1 to 10 m deep, with
the depth to top-width ratio normally ranging from 3:1 to 6:1. They are best
1st Winter 1st Fall

Figure 11.10.
Evolution of an ice wedge by
thermal contraction (After
Lachenbruch 1966)
meters
500th Winter 500th Fall
Fro2Efe=
developed in fine-grained soils with a high ice content (Black 1976). The host
sediments are usually upturned within 3 m of the wedge, in response to the
horizontal compression generated as the wedge grows. Ice in wedges contains
oriented air bubbles and a complicated fabric of ice crystals (Black 1974).
The growth of an ice wedge does not necessarily proceed on an annual
basis. Mackay (1974, 1975a) observed that only 40 percent of the ice wedges
on Garry Island (Northwest Territories, Canada) cracked in any given year,
the frequency of cracking apparently controlled by the depth of snow cover.
Even if cracking does occur, the openings often narrow before ice-veinlet growth
begins. The size of a veinlet thus is dependent on the timing between closure
of the winter cracks and the appearance of meltwater (Mackay 1975). In most
cases the thickness of an annual increment is considerably less than the size
of the thermal crack.
wedges are commonly connected in a polygonal pattern that is similar
Ice
Figure 11.11.
in to the more familiar mud cracks. They differ mainly in that
most respects
Polygonal markings on ground
ice-wedge polygons are much larger, having diameters that range from several surface caused by ice-wedge
meters to more than 100 meters (fig. 11.11), and they contract and expand polygons in the vicinity of Meade
River, about 35 miles southeast of
according to the laws that govern temperature effects on ice (Black 1974).
Barrow, Alaska. Barrow district,
Northern Alaska region, July 1949.
-. v-"-r- v .
'3-'-
Chapter 11 424
The perimeters of some polygons are accentuated as minor ridges, so that the
boundary is elevated relative to the center of the feature (low-center ice-wedge
polygons). Others (high-center polygons) are depressed along their edges by
melting and erosion of the ice wedges, leaving the central core higher than the
rim (Pewe 1966; Black 1974, 1976).
Actively growing ice wedges are restricted to areas of continuous perma-
frost,and their southern boundary seems to be the — 6°Cto— 8°C isotherms
(Pewe 1973). Farther south in the discontinuous permafrost zone, they be-
come inactive and disappear. As ice wedges melt, the surrounding sediment
may collapse into the space opened by the disappearance of the ice, thereby
creating an ice-wedge cast, a feature that is best preserved in gravels (Black
1976). Although ice-wedge casts are probably the best proof of an earlier
permafrost condition, other wedgelike features unrelated to permafrost look
very similar. For example, features called soil wedges are commonly reported
in the Soviet literature. They apparently originate by frost cracking in the
active layer and in-filling with local sediment. Therefore, it is easy to misin-
terpret these features as ice-wedge casts, which incorrectly suggests that a
permafrost environment existed in the past and has subsequently been de-
stroyed. In fact, it is become part of a
possible for an old ice-wedge cast to
modern wedge if frost cracking operates in the present active layer. To
soil
avoid confusion, Washburn (1980) suggests that any wedge or wedge-cast type
should be designated as being either fossil or contemporaneous.
Pingos The term pingo (Eskimo for "mound" or "hill") was first suggested
in 1938 by Porslid for large, ice-cored, domelike features that exist only in
permafrost regions. They are most perfectly developed in areas of continuous
permafrost, although they do form in the discontinuous zone (Holmes et al.
1968). Active pingos rise from a few meters up to heights of 70 meters and
have basal diameters up to 600 meters (Washburn 1980). They are oval or
circular in plan view, as in the photograph in figure 11.12. The ice core in a
pingo is typically massive and is often exposed by tension fractures that de-
velop at the summit of the mound as it rises. Sometimes the exposed ice melts,
and small freshwater lakes occupy the craters bounded by the tensional cracks.
Many varieties of pingos are known (Pissart 1970; Flemal 1976) but most
can be placed in two major genetic categories. Closed-system pingos (Mac-
kenzie type) develop in level, poorly drained, shallow lake basins. They are
most distinctly preserved Mackenzie delta (Canada) where a clutch of
in the
1,450 pingos gives evidence that pingos tend to form in fields rather than as
single individuals. In general, the draining of a lake allows the permafrost
As the table rises, water is
table to rise to the level of the former lake floor.
trapped in The cryostatic pressure (pressure from ice for-
the saturated soil.
mation) displaces the water upward until it also freezes and becomes the core
of the pingo (see Flemal 1976). The size and shape of the pingo often reflect
that of the original residual pond; the pingo may grow in distinct stages
(Mackay 1973).
&
73T
Open-system pingos (East Greenland type) develop most readily on slopes Figure 11.12.
The Discovery pingo. A dense
where free water under artesian pressure is injected into the site of the pingo.
stand of large birch trees on the
As the water approaches the surface, it freezes. The continuing introduction pingo contrasts sharply with the
of water from below under hydrostatic pressure builds up the ice mass and open stand of small black spruce
in the surrounding muskeg.
domes the surface material into the pingo shape (see Miiller 1963). The initial
Goodpaster district, Yukon region,
growth of some pingos may involve ice lenses above the water table, but their Alaska, Aug. 1960.
continued growth requires a proper combination of hydrology and soil texture
(Ryckborst 1975). Open-system pingos are much more common in discontin-
uous permafrost where taliks and freely circulating water are not unusual.
The rate of pingo growth is variable, ranging from a few centimeters to
more than a meter a year (Washburn 1980; Mackay 1973). This indicates that
many large pingos are quite old. In fact,
,4
C dates have shown two closed-
system pingos to be 4,000 and 7,000-10,000 years old (Miiller 1962). On the
other hand, some pingos are growing today (Mackay 1973), suggesting that
pingo formation did not occur in a unique or specific time in postglacial history
but is probably a continuing process.

Fossil pingos (Wayne 1967; Mullenders and Gullentops 1969; Flemal et
al. 1976) provide good evidence of an extinct permafrost environment in re-
gions that are now temperate. Proving that features are fossil pingos, however,
requires detailed field and laboratory analyses. The DeKalb mounds (northern
Illinois), for example, consist of approximately 500 circular or elliptical mounds
shaped from late Pleistocene sediments. Good evidence for their pingo origin
Chapter 11 426
is the large cluster of features and the fact that their centers consist of lacus-
trine silts and clays. Topographically, however, most of the mounds are less
than 5 m higher than the surrounding ground level, and many are up to 300 m
in diameter. Their recognition as fossil pingos could not have been suggested
without careful field examination.
Thermokarst Thermokarst features are a variety of topographic depressions

that result from thawing of ground ice. They are more abundant where ice
wedges are present and where the thermal equilibrium is somehow disturbed.
The destruction of ground ice is due either to lateral degradation (back-
wearing) where lateral erosion of surface water exposes the ice, or to vertical
degradation (down-wearing) when the surface thermal properties are altered
(discussed in Czudek and Demek 1970). These features are similar in many
respects to those found in normal karst regions, but the fundamental process
is melting rather than solution of rock material.
Many of the processes that culminate in thermokarst features are initi-
ated by broad climatic changes, but minor events such as fires, clearing of
forest vegetation for farming, shifting stream channels, etc., probably also can
upset the thermal balance enough to be the impetus. Mackay (1970), for ex-
ample, reports a remarkable case which trampling of vegetation by a dog
in
led to subsidence of 18-23 cm in a small area within two years. Observations
such as this demonstrate the remarkably fragile equilibrium of the periglacial
system.
Patterned Ground
Periglacial regions are often characterized by a peculiar arrangement of sur-
face materials into distinct geometric shapes. The features, collectively known
as patterned ground, include such diverse shapes as polygons, circles, and
stripes. They are common in permafrost regions, but perennially frozen ground
is not a prerequisite for their development. They can be found anywhere within
the periglacial regime or even in other morphogenetic regimes. Since pat-
terned ground was first described in the late 1800s, almost every worker in
periglacial areas has noted its and theories concerning

striking appearance,
its origin are almost as numerous as the features themselves. Perhaps the most
extensive review of patterned ground and its origin was provided by A. L.

Washburn in 1956, and that excellent work still stands as the cornerstone of
our knowledge about these forms.
Classification According to Washburn (1956), patterned ground features can

be placed in a descriptive classification on the basis of two primary criteria:
(1) the geometric shape (circle, polygon, etc.). and (2) whether the material
composing the feature has been sorted (table 11.3). Sorting separates the
larger-size fractions from the fine particles and usually generates a feature
rimmed by stones and centered with fines (for a review see Goldthwait 1976).
The most common forms seem to be circles, polygons, and stripes, with steps
and nets as minor transitional types. For a more detailed discussion see Wash-
burn (1980).
Table 11.3 Simplified classification of patterned ground features
Types Subtypes Processes
Circles Sorted Features subdivided on basis of necessity of cracking (frost

Nonsorted cracking, permafrost cracking, dilation, joint control)
Where cracking not essential, heave and mass
displacement important
Polygons Sorted Includes ice and sand wedges. Cracking of all types
Nonsorted Heave, mass displacement, and thaw processes
important in noncracked types.
Nets Sorted Includes earth hummocks Cracking of all types except joint
Nonsorted controlled types. Heave and mass displacement in
noncracked types Thaw also important in sorted
varieties.
Steps Sorted Cracking unimportant Heave, mass wasting, and

Nonsorted displacement in nonsorted Frost sorting and thaw also
important in sorted varieties Terracette form.
Stripes Sorted All types of cracking important Heaving, mass wasting, and
Nonsorted displacement in nonsorted Frost sorting and thaw also
important in sorted types.
After Washburn 1970. in Acta Geographies Lodziensia Used with permission of Institute of Geography, Poland
Polygons Polygons are best developed on flat, nearly horizontal, surfaces.

The bounded by straight segments composed of stones that
sorted variety is
surround a central core of finer material. They range in diameter from a few
centimeters to over 10 m and always occur in groups rather than individually.
The stones marking the boundary are often oriented parallel to the direction
of the rim. The stones increase in size with the dimension of the polygon and
decrease in size with depth.
Nonsorted polygons differ in several ways from the sorted type: (1) the
polygonal borders are devoid of stones, the geometric form being marked by
furrows, (2) they can be considerably larger than the sorted variety, often
reaching 100 m
in diameter, and (3) they have been found on slopes as steep
as 31°. Nonsorted polygons are usually associated with the ice-wedge poly-
gons discussed earlier. As such they are primarily related to permafrost. How-
ever, they can also form by dessication and even in solid rock (Walters 1978).
Circles Sorted circles are stone-rimmed circular forms that range from a
few centimeters to several meters in diameter (fig. 11.13). As with polygons,
the stone size correlates with the dimension of the circle and decreases with
depth. Circles, unlike polygons, can exist singly or in groups. Nonsorted circles
lack the coarse, bouldery rim and are bounded instead by vegetation sur-
rounding a central core of bare soil. In some cases the barren core develops
because scouring and winnowing action by the wind preferentially erodes and
initiates the circular form (Fahey 1975).
Stripes Stripes are linear alignments of stones, vegetation, or soil on slopes.

The sorted stripes are elongated strips of stones separated by intervening zones
of fine sediment or vegetation. In general, the pattern changes as the slope
Chapter 11 428
increases, from polygons through sorted nets or steps into stripes. Stripes usu-
ally range in width from several centimeters to several meters. Their length
varies but is sometimes greater than 100 m. As in other sorted features, larger
stones are associated with bigger stripes, and the largest particles are found
at the surface of the accumulation. Nonsorted stripes consist of vegetation or
soil with intervening zones of bare ground. They are usually not as long as the
sorted variety and are sometimes discontinuous.
Origin Patterned ground probably originates in a variety of ways; in fact,

Washburn ( 1 956) discusses 1 9 major hypotheses and concludes that patterned
ground is polygenetic and that some forms result from a special combination
of processes. Despite the conflicting opinions, the properties of the major fea-
tures do imply that patterned ground exists in a range of transitional types.
The characteristics of the different types reflect variations in a few controlling
Figure 11.13. factors rather than completely different origins. Although details are still
Sorted stone circles caused by
frost action, in Alaska Range,
missing, it does seem certain that several processes are basic to the formation
south central Alaska June 1968
of patterned ground: (1) cracking or dessication, (2) heaving and its associ-
ated sorting, and (3) gravity as expressed by the slope inclination. Cracking
is instrumental in developing polygonal geometry and is primary to both sorted
and nonsorted types. Heaving separates the coarse and fine sediment fractions.
In the usual case, repeated freezing and thawing move the stones upward and
outward. This creates a fine-grained nucleus as the smaller grains with greater
cohesion contract farther inward and downward during the thaw phase. The
process continues until adjacent cells coalesce, and the coarse stones are con-
centrated as polygonal or circular rims along the lines of interference. This
model is accepted by most periglacial students although some evidence exists
to suggest that texturally unsorted debris does not always respond to frost
action in the same way (Ballard 1973).
The rate of lateral movement of large clasts is not well documented. Re-
cent detailed work by Vitek (1983) suggests that particles larger than 1.3 cm
in length are transported to the polygonal margins at a rate ranging from 0.45
to 0.89 cm/yr. Larger particles move at a slower rate. The clasts are moved
by and the growth of needle ice and/or lacustrine ice.
frost action
In recent analyses (Ray et al. 1983a, 1983b), the regularity in sorted pat-
terned ground has been theoretically explained as resulting from convection
of water in the active layer. Convection cells are presumed to form when ice
melts in an uneven manner during the summer thaw. This creates an undu-
lating frozen ground surface beneath the active zone and temperature and
density gradients in the water of the active layer. The model does not suggest
that stones are physically moved by convecting water; instead, it contends that
circulating water influences the shape of the underlying ice front. This, in turn,
determines where sorting mechanisms will work most effectively until the sur-
face pattern is a mirror image of the undulating ice front. In limited field
studies, the model has correctly predicted the width to depth-of-sorting ratio
for sorted polygons.
The origin of stripes is clearly related to slope angle. Polygons and circles
are much more prevalent on flat surfaces, and stripes become dominant as
slope angle increases. This does not mean that stripes never form on gentle
slopes; they do (Evans 1976). However, such occurrences are rare, and as gra-
dients increase, stripes are the most common form of patterned ground. Usu-
ally transitional forms of nets or steps develop at angles between 2° and 7°;
the precise angle at which stripes develop depends on other factors such as
soil texture and moisture and the formation of needle ice (Mackay and Mat-
thews 1974; Washburn 1980).
Landforms Associated with Mass Movement

In contrast to patterned ground features, which are shaped primarily by frost
action, certain periglacial forms result when the dominant mechanism is
transportation. This does not mean that frost action is absent. On the contrary,
frost action is a necessary prelude to the mass movement, and probably the
two processes function simultaneously to fabricate the ultimate landform.
Nonetheless, the final geometry of these features reflects the motion of rock
and soil particles rather than a static, in situ disintegration of the parent ma-
terial (for reviews see Benedict 1976 and S. E. White 1976b).
Chapter 11 430
Gelifluction Features The process of gelifluction operates on gradients as low

as l° to 2°. Where the process is active a number of deposits and surficial
forms result (gelifluction sheets, gelifluction benches, gelifluction lobes and
gelifluction streams). Although these features can be separated on the basis
of individual morphology (see Washburn 1980), they share common genetic
mechanics. Variations in form result from differences in texture, gradient, and
soil moisture. Material in gelifluction deposits is usually poorly sorted, but a
crude stratification may be present in some deposits. Angular clasts are nor-
mally oriented with long axes arranged parallel to the direction of movement.
In regions covered by arctic or alpine tundra the most common feature
seems to be the gelifluction (solifluction) lobe. The large, tonguelike masses
of surface debris may occur as a single lobe, usually 30 to 50 m wide, or as
one of many individuals that are joined laterally into a much broader field. In
long profile an individual lobe ismarked by a pronounced scarp at its leading
edge that ranges from 1 to 6 m high. Commonly, however, a succession of lobes
develops on a slope in which each upstream lobe overlaps the rear of the next
downslope lobe, giving the entire sequence a staircase profile. Internally the
lobes consist of angular unsorted debris, with particle sizes varying from silt
to coarse boulders. The long axes of the boulders are usually oriented parallel
to the direction of movement. Various types of gelifluction lobes and related
forms (turf-banked lobes and terraces, stone-banked lobes and terraces) are
discussed in detail by Benedict (1970, 1976).
Gelifluction lobes, as you might expect, move primarily by gelifluction.
but other processes are undoubtedly involved. Flow mechanics is indicated by
the fact that the greatest movement occurs near the surface, and essentially
no movement is perceived below a depth of 25 cm (Price 1972). As the ma-
terialmoves downslope, it encounters zones of high resistance to flowage. These
retard the forward motion and cause the soliflucting layer immediately up-
slope from the resistant zone to bulge slightly, producing the scarplike front.
As Price (1972) points out, however, the front is often stabilized, and only a
thin surface layer slides to and cascades down the stationary front. The ad-
vance of the lobe in this manner is much slower than one might expect, prob-
ably averaging 1-3 cm a year, but the rate varies according to where the
measurement is made. Maximum velocities are found along the longitudinal
axis and decrease progressively toward the lateral margins (Washburn 1973).
but rates are significantly affected by moisture content and gradient (Benedict
1970) in addition to vegetation cover and depth of freezing (Gamper 1983).
Blockhelds Blockfields (often known as felsenmeer) are usually thought of

as broad, relatively level areas covered by moderate to large angular blocks
of rock (Sharpe 1 938). Similar accumulations of blocky debris also are found
on slopes and are variously referred to as block slopes, block streams, rubble
sheets, and rubble streams (see Washburn 1973). All these terms have been
used interchangeably, and some semantic confusion exists. Our concern here
is with block accumulations that occur on slopes, either concentrated in val-
leys or disseminated across a wide area. Although we may employ the term
"blockfield" in reference to these accumulations (as have many others), some
workers would consider this technically incorrect. True blockfields as origi-

nally described are not the product of mass movement but result in situ from
frost wedging and heaving of the underlying bedrock. In this discussion we
will actuallybe treating block streams or block slopes, but the term "block-
field" has been so widely applied to these features that we will employ it to
mean any accumulation of blocky debris, including those found on slopes and
subjected to mass movement.
Blockfields on slopes tend to be elongate with widths averaging between
60 m and 120 m and lengths from 350 m to 1.3 km. Maximum and minimum
values vary drastically depending on the block size, the slope angle, and the
distribution of the bedrock source. The gradient on the accumulation ranges
widely,from 1° to 20°, although it most commonly is between 3° and 12°.
Some workers feel that an appropriate upper limit for these features is 10°
to 15° (Caine 1968; White 1976). Blocky slope accumulations normally dis-
play an internal fabric indicative of movement; i.e., long axes of individual
blocks are oriented downslope. The orientation may become transverse to the
slope near the distal end of the deposit (Caine 1 it may diverge locally
968), or
from the norm. Accumulations seem to be 4-20 m thick, but very few obser-
vations of the total thickness have been made. One deposit that was totally
exposed displayed an internal layering characterized by an open-textured sur-
face zone and a general increase in matrix with depth (Caine 1968). This type
of fabric distribution, however, cannot be accepted as a general trait until more
observations become available. Individual blocks contained in boulder masses
are striking in their enormity, most being 1-3 m in intermediate diameter and
some as large as 6 m (fig. 1 1.14).
A number of mass-movement processes — ranging from avalanches or
landslides to catastrophic floods —have been proposed development of
for the
block slopes or streams. Precisely which mechanism or combination of pro-
cesses leads to the features is mostly conjecture, and the distinct probability
remains that block accumulations have more than one mode of origin. The
most commonly cited transporting mechanism is gelifluction. Caine (1968),
for example, concluded that in Tasmania the block slope material moved when
the entire deposit contained abundant matrix. The open texture in the surface
zone (up to 3 m deep) could have been created after movement ceased when
the matrix was removed by ground or surface water (Caine 1968; Potter and
Moss 1968). The gelifluction model will work efficiently only if the lower por-
tions of the accumulation contain a high content of fine-grained matrix. Such
a layer was observed by Caine (1968) and suggested by Denny (1956) and
Sevon (1969). However, it has not been proven as a universal property of these
features, and until more conclusive data are found, the solifluction mechanism
can be only a viable hypothesis. Nonetheless, it would explain the terrace and
lobe development noted on some block streams (Denny 1956; Potter and Moss
1968). Further, the orientation of blocks is similar to that characteristically
produced by gelifluction transport.
Chapter 11 432
Some investigators have suggested that slope blockfields are emplaced by

movement as rock glaciers, i.e., an accumulation containing interstitial ice
rather than a soil matrix (Patton 19 10; Kesseli 1941; Blagborough and Farkas
1968). Such mechanism has the advantage of easily accounting for the open
a
texture in the blocky debris, since no matrix sediment must be removed. The
ice simply melts away as the rock glacier becomes inactive. A rock glacier
transport, however, fails to explain why slope blockfields lack such normal
characteristics of rock glacier deposits as steep fronts and pronounced trans-
verse ridges and furrows. In addition, a greater shear stress is required in
moving a rock glacier than in gelifluction of a fine soil, and so, to attain the
proper mobility, rock glacier accumulation should be considerably thicker than

blockfield deposits (Wahrhaftig and Cox 1959; Potter and Moss 1968).
In comparison with inactive deposits, active blockfields conspicuously lack
lichen cover, and the clasts show no signs of weathering. Because of the wide
distribution of active block features in modern periglacial regions, inactive
deposits in temperate zones or at altitudes well below functioning periglacia-
tion have been widely interpreted as evidence of earlier periglacial conditions.
Figure 11.14. For example, many blockfields have been described in the middle Appalachian
River of Rocks blockfield (block Mountains (H. T. U. Smith 1953; Denny 1956; Potter and Moss 1968) where
stream), Berks County, Pa Slope
present climates are patently not periglacial.
is 5° Blocks in foreground are as
much as 5 m long
Rock Glaciers Rock glaciers are features having a bouldery surface com-
posed of angular rock fragments that are commonly up to 3 m in diameter.
This texture is somewhat misleading because many rock glaciers contain finer-
grained debris at depth. The features usually originate in cirques or in high,
steep-walled recesses.
Rock glaciers are usually classified as tongue-shaped or lobate on the basis
of their plan-view shape (see White 1976b for discussion). Tongue-shaped forms
often head in cirques and extend downvalley, maintaining an external shape
that is similar to a true glacier. In contast, lobate rock glaciers are short and
broad and develop below talus accumulations along valley sides having steep
cliffs.
Tongue-shaped rock glaciers are several hundred meters to more than a

kilometer long and vary in width, narrowing toward their source where the
valley sides are more constricting (fig. 11.15). The deposits contain ice several
meters below the surface. According to Potter (1972), the ice exists either as
an interstitial cement (ice-cemented type) or as an extensive mass that con-
tains widely dispersed rock fragments (ice-cored type). In either case the fea-
tures may represent a transitional phenomenon between a nonglacial process
and a glacier with a thick mantle of surface debris or, alternatively, an ice- Figure 11.15.
Rock glacier on Sourdough
cored moraine (for discussion see Barsch 1971; Ostrem 1971). Topographi-
Mountain near McCarthy, Wrangell
Mountains, Alaska
\*
Chapter 11 434
cally, any rock glacier may have furrows, crevasses, and lobate or transverse
ridges similar to those found on true glaciers. Their front is normally near the
angle of repose and therefore quite steep.
Modern, active rock glaciers are known in polar or subpolar regions and
in the high mountains of the mid-latitudes. They vary in thickness from 15 to
50 m and move continuously at rates usually ranging from centimeters to a
meter per year, depending on local conditions. In some cases they are known
to be advancing in areas where glaciers are presently retreating, perhaps be-
cause the surface rubble insulates the underlying ice and slows its response to
climatic change (Osborn 1975). Inactive or fossil rock glaciers are widely dis-
tributed in regions that now have temperate climates. Whether they have pa-
leoclimatic significance depends on precisely how such features originate.
Although a variety of origins have been suggested, only those processes that
produce a continuous forward motion are acceptable. The main controversy
surrounding the genesis of rock glaciers revolves around the question of whether
or not a true glacier is a required precursor (Outcalt and Benedict 1965). In
other words, is the ice contained within the rock glacier necessarily glacier
ice, or could it have accreted in a periglacial setting by gradual freezing of
porewater? The fact that some fossil rock glaciers in the southwestern United
States developed in regions having no evidence of glaciation (Blagborough and
Farkas 1968; Barsch and Updike 1971) seems to support a completely peri-
glacial origin for at least some of the features. It seems reasonable, however,
that most tongue-shaped rock glaciers are somehow related to true glaciers.
Lobate types may be ice-cemented and move by creep generated by the weight
of the bouldery mass on the interstitial ice (White 1976b; Washburn 1980;
Wayne 1981). Nonetheless, enough disagreement exists about the origin and
transport mechanics of these deposits to refrain from sweeping conclusions at
this time (see Whalley 1983 for discussion).
Environmental and As mentioned human expansion into the arctic and subarctic regions
earlier,
Engineering is impending population growth has generated a growing con-
inevitable. This
Considerations cern as to whether our perception of geomorphic systems operating in these
areas is sophisticated enough to develop the regions and still preserve their
fragile environmental balance. The concerns are most evident in the engi-
neering community that deals with construction problems in arctic regions on

a day-to-day basis (for many references see National Academy of Sciences
1 983). The technical problems are real; our ability to cope with them depends
on how well we understand the processes functioning within the system and
how those processes are ultimately related to the environmental variables.
Most engineering problems are associated with permafrost. According to
Brown ( 970), any one of four basic engineering approaches may be employed
1
when dealing with the permafrost condition: (1) disregard the permafrost;
(2) use an active approach in which the permafrost in the near surface is elim-
inated by keeping the soil continuously thawed or by removing the natural soil
and replacing it with permafrost-resistant material; (3) use a passive ap-

proach in which permafrost is preserved by keeping the soil frozen at all times
and thereby eliminating the annual thaw; and (4) design structures to with-
stand frost action.
Which of the above methods is best depends on the local situation. In
discontinuous permafrost, it may be possible to ignore the condition if the
parent material is bedrock or a sandy or gravelly soil. Active approaches, such
as simply removing the vegetation and its insulating effect, are also feasible
on a local scale. In zones of continuous permafrost, the potential problems
cannot be disregarded. There the most practical method is to keep the area
continuously frozen, primarily by ventilation or insulation.
Some heaving and settling will occur in the first several years after con-
struction regardless of the method employed, because the permafrost requires
some time to adjust into a new equilibrium condition. Thus, any project should
prepare for the initial response in the system with a design that can withstand
the stresses generated. Often the periglacial features described earlier give
tangible clues about the near-surface ground conditions at any locality (table
1 1.4), and their recognition is an important first step in a site analysis (Thomas
and Ferrell 1983). Air photos and remote sensing data interpreted by a geo-
morphologist can provide a valuable contribution in the planning stage.
Building Foundations
The bearing strength (ability to support load without plastic deformation) of
a surface varies significantly with the type of material, both when it is frozen
and when thaws (Swinzow 1969). The loss of strength during thaw is most
it
dramatic in clay and silt soils that are not permeable and so retain a large
portion of the water released by melting. Bedrock or coarse-grained deposits
are reasonably stable during the thaw phase because they drain easily. Be-
cause water cannot permeate below the permafrost table, the active layer in
fine-grained soils becomes a quagmire that cannot support weight and there-
fore allows differential settling of overlying buildings, bridge abutments, etc.
In addition, the lack of lateral drainage and the high segregation potential
will accentuate differential heaving during the next freeze. Building founda-
tions, therefore, present enormous problems. In the case of small dwellings it
may be adequate simply to live with the frost activity by pursuing an annual
maintenance program. Large buildings, however, become unsafe if they are
allowed to endure substantial heaving and settling, and preventive measures
must be taken. Lobacz and Quinn (1963) were able to show that residual thaw
zones persist beneath buildings that have no ventilation at their base, and that
when buildings are elevated, seasonal frost will extend from the surface to the
permafrost table.
Because of these conditions, most building constructions today in areas
of continuous permafrost follow the passive approach and attempt to maintain
the permafrost intact. The floor of the building is built on top of pilings driven
into the permafrost layer, so that it normally stands 1-2 m above the ground
Chapter 11 436
Table 11.4 Landforms that indicate ground conditions in arctic and near-arctic
regions.
Feature and Description Associated Ground Conditions
Polygonal ground (ice wedges) Usually — Typically indicates relatively fine-grained

indicates the presence of a network of unconsolidated segments with
ice —
wedges vertical wedge-shaped ice permafrost table near the ground
masses that form by the accumulation of surface; also known from coarser
snow, hoarfrost, and meltwater in ground sediments and gravels where wedge ice
cracks that form owing to contraction is less extensive.
during the winter Wedge networks are

also common in wet tundra where no
surface expression occurs. (Subject to
extreme differential settlement when
surface disturbed.)
Stone nets, garlands, and stripes Frost — Indicate strong frost action in moderate well-
heaving in granular soils produces drained granule sediments that vary from
netlike concentrations of the coarser silty fine gravel to boulders. Surficial
rocks present. If the area is gently material commonly susceptible to
sloped, the net is distorted into garlands flowage.
by downslope movement. If the slope is
steep, the coarse rocks lie in stripes that
point down hill
sheets and lobes Sheets or

Solifluction — Indicate an unstable mantle or poorly
lobe-shaped masses of unconsolidated drained, often saturated sediment that is
sediment that range from less than a moving downslope by seasonal
largely
foot to hundreds of feet in width and frost heaving On steeper slopes they
may cover entire valley walls, found on often indicate bedrock near the surface,
slopes that vary from steep to less than and on gentle slopes, a shallow
3°. permafrost table
Thaw lakes and thaw pits Surface — Usually indicate poorly drained, fine-grained
depressions form when local melting of unconsolidated sediments (fine sand to
permafrost decreases the volume of ice- clay) with permafrost table near the
rich sediments. Water accumulates in surface
the depressions and may accelerate
melting of the permafrost Often form
impassible bogs
Beaded drainage — Short, often straight Indicates a permafrost area with silt-rich
minor streams that pools or small
join sediments or peat overlying buried ice
lakes. Streams follow the tops of melted wedges.
ice wedges, and pools develop where
melting of permafrost has been more
extensive
Pingos — Small ice-cored circular or elliptical Indicatesilty sediments derived from the
hills occur in tundra and forested
that slope or valley, also groundwater with
parts of the continuous and some hydraulic head that is confined
discontinuous permafrost areas. They between the seasonal frost and
often lie at the juncture of south and permafrost table or is flowing in a
southeast-facing slopes and valley floors thawed zone within the
and in former lakebeds permafrost Those in former lakebeds
indicate saturated fine-grained
sediments
From Femans el al 1969

surface, and air can circulate freely in the open space between the floor and
the ground. This keeps the ground frozen in the winter because the heat ra-
diating from the building is removed by cold air moving through the open
space. Thawing does occur in the summer, but it is minimized because the
surface is shaded by the building. Nonetheless, the piles supporting the building
are susceptible to some heaving in the active layer; Johnston (1963) suggests
they should be driven to a depth at least twice the thickness of the active layer
in order to keep the heaving effect at a minimum. Other methods of air cooling
or liquid cooling (Cronin 1983) have been utilized to maintain the frozen con-
dition.
Whenever possible, buildings should be located on a sand and gravel sub-
some communities, built
strate, for reasons explained above. In coastal regions
prior to any understanding of periglacial processes, have been built on silty,
deltaic plains. Further development may be impossible in such a setting; in
fact, some small towns may have to be relocated if the problems associated
with the permafrost cannot be controlled. Relocation was actually accom-

plished for one community that originally existed on the deltaic plain of the
Mackenzie River. The Canadian government physically moved the commu-
nity 56 km to a new townsite that was chosen because it had abundant sand
and gravel for building sites and roads. Although construction of the new town
was costly, its success demonstrated that a reasoned approach and proper reg-
ulations can allow development to proceed without harmful degradation of the
system (for details see Pritchard 1962; Price 1972; Cooke and Doornkamp
1974).
Roads and Airfields

The outbreak World War II made the U.S. and Canadian governments aware
of
that defending Alaska and the Northwest Territories from attack would be a
monumental task. In an attempt to rectify this vulnerability, many necessary
but hastily conceived projects were completed to facilitate the transportation
of men and supplies to the area. Railroads, pipelines, and airfields were con-
structed, and the Alcan Highway, stretching almost 3000 km, was completed
in only eight months, a truly remarkable engineering feat. The highway, how-
ever, in permafrost mechanics, as the construction
proved to be a case study
and maintenance problems were often baffling and disruptive (Richardson
1942, 1943).
We now know that problems associated with roads or airfields vary greatly
because they are affected by microenvironmental conditions. In general, the
most common ailments are (1) differential heaving and settling that creates
a washboard surface (fig. 11.16), (2) sinking of portions of the surface,
(3) destruction of bridges by spring floods, (4) burial of the pavement by
slumping or other mass movements, and (5) icing. Icing is by far the worst
problem. It occurs when water runs onto road surfaces during the winter,
quickly freezes over large areas, and often accretes to a considerable thick-
ness. Groundwater under hydrostatic pressure may continue to flow through
Chapter 1 1 438
most of the winter and, with improper planning, may also run onto the road
surface if road cuts open springs and seeps. Even something as innocuous as
a river freezing may result in icing of a nearby highway surface. This happens
when the river freezes from the surface downward and creates a pressure head
on the bottom water. The water, thus mobilized, moves to the channel sides
where it and overflows the channel banks onto the highway.
rises to the surface
The best road protection requires wise planning and carefulpavement de-
sign. These can reduce frost action in the active layer and prevent excess water
from reaching the road surface. The highway route should avoid seeps and
springs by diverting around hills. If cuts must be placed in hillslopes, the pave-
ment can be protected from icing by integrating a system of culverts and
drainage ditches that will divert free water away from the road.
Figure 11.16.
Pavement design usually begins with a layer of gravel 0.6-1.5 m thick
Gravel road near the Umiat
showing severe differential
Airstrip
spread directly on the tundra to insulate the underlying surface and so min-
subsidence caused by thawing of imize thawing. A layer of insulating material such as peat (McHattie and
ice-wedge polygons in permafrost
Esch 1983) or polystyrene (Johnston 1983) may also be inserted within or
Anaktuvuk district, Northern
Alaska region, Alaska, Aug 1958
beneath the gravel cover. Some melting must be expected, nonetheless, and
the design should allow for reduced strength in the summer and for heaving
in the winter (Linell and Johnston 1973; Hennion and Lobacz 1973). Even
painting the pavement surface white may help to reduce the thaw of the sub-
strate.
Utilities — Water and Sewage

Some of the most burdensome problems facing communities in permafrost
regions involve utilities and services that are taken for granted in the mid-
latitudes.Water supply, for example, is complicated by deep winter freezing,
and large lakes may be the only sufficient, continuous, and dependable water
source. Taliks within the permafrost may provide some water, but their dis-
tribution is not regular and they may not contain enough water to support
even a small settlement. Although taliks beneath the permafrost can provide
large amounts of water, it is very expensive. Drilling through the permafrost
is costly,and pipes must be cased to prevent freezing. In addition, some deep
subpermafrost water is brackish or highly mineralized (J. R. Williams 1970).
Locating an adequate, continuous water supply solves only half the
problem, because the water must still be delivered to the place where it is to
be used. People in tiny settlements have traditionally obtained water by car-
rying it in tanks from a nearby river or lake. In the winter, when the source
freezes over, blocks of ice are melted. Larger communities cannot rely on such
primitive methods and most have piped-in water. Either the pipes must be
insulated or the water must be continuously circulated to prevent freezing.
Many areas now utilize an enclosed heated and insulated conduit, called a
utilidor, for water delivery. The utilidor complex, placed above the ground, is
designed to hold water pipes, electric cables, and sewage disposal pipes (see
Zirjacks and Hwang 1983), but its installation is very expensive.
Sewage disposal is still another perplexing problem. Some small villages
utilize individual buckets as collectors of all kinds of waste, human and oth-
erwise. These are carried to a specified location on a lake or river where they
are dumped in expectation of the spring thaw. Essentially, nature cleans the
system at winter's end. For sanitation alone, larger communities must be more
sophisticated and most now employ the utilidor system. In practice, however,
utilidors solve only theimmediate necessity of removing wastes from indi-
vidual doorsteps. The wastes is a most difficult problem. Few
final disposal of
communities have efficient treatment plants, and most wastes are ultimately
dumped as raw sewage into the nearest waterway. The problem is not nearly
as acute in discontinuous permafrost areas, where septic tanks and cesspools
are common. In continuous permafrost zones, however, alternative techniques
must be developed. Sewage lagoons (large dug-out hollows) do allow waste to
decompose anaerobically (Brown 1970) and are becoming a common disposal
method. One novel idea, tried at a military base, was to use fuel oil instead of
water sewage system. When the wastes are collected, they are suffi-
in the
cientlymixed with the oil to be injected directly into a furnace where incin-
eration eliminates the problem and simultaneously generates heat and
electricity for the base (Alter 1966).
Chapter 11 440
Pipelines
It is perhaps fitting to end our discussion of periglacial geomorphology with
a brief description of the one project that focused so much attention on this
environment in recent years. The discovery of oil on the north shore of Alaska
near Prudhoe Bay was exciting news to an energy-addicted nation. The con-
flict between energy and environment that ensued created a furor that ran
from reasoned arguments to pure emotionalism. Nonetheless, it did force us
to analyze in a systematic way the problems associated with construction of
a pipeline in a permafrost region.
The Trans-Alaska pipeline was not the first pipeline constructed by the
United States in a periglacial environment. The Canol pipeline system was
built during World War II across 2575 km of discontinuous permafrost. Its
purpose was to pump crude oil along the main pipe to a refinery located along
the Alcan Highway near Whitehorse, B.C. From there gasoline was pumped
to airfields and other military establishments. The Canol project, like the Trans-
Alaska pipeline, was steeped in controversy (Richardson 1944), and its use
was discontinued in May 1945 after only 13 months of operation, presumably
because of the high cost of maintenance and its decreased importance in the
war effort.
The Trans-Alaska pipeline is different in most respects from the older

Canol project, and little applicable knowledge was derived from the earlier
work. The Trans- Alaska line is much farther north, and much of its 1270 km
route from Prudhoe Bay to Valdez crosses zones of continuous permafrost.
The 1.2 m diameter pipe is larger than the Canol pipe and, fully loaded with
oil, weighs over 900 kg per meter (Harwood 1969). The crude oil enters the
pipe at a temperature of about 58 °C, but friction along the flow path increases
thistemperature considerably. Cooling devices are placed along the route to
keep the temperature at a constant 63 °C, but the oil is still hot and so presents
the most difficult problem in preserving the permafrost environment intact.
Given the above, the planning had to consider the possibility of major
changes in the physical environment as well as a host of engineering problems
(Lachenbruch 1970; Kachadoorian and Ferrians 1973). A pipe resting on the

surface and carrying hot fluids would certainly cause thawing of the under-
lying permafrost. Without preventive measures, the thaw would probably ex-
pand outward with time as shown in figure 11.17, perhaps at a decreasing rate,
but possibly never reaching an equilibrium state (Lachenbruch 1970). The
soil could also become liquefied and unable to support the heavy pipe, and any
gelifluction induced by the liquefaction might carry the pipe along with the
soil, producing ruptures and oil spills. As the pipe crosses materials of diverse
texture and composition, differential heaving and settling are inevitable, and
manner could bend and rupture the pipe. This problem
stresses generated in this
is where the pipeline crosses slopes underlain by ice wedges
especially pertinent
that melt more quickly and undermine the support rapidly. In addition, the
actual construction could have inadvertently altered the delicate environ-
mental balance. For example, destruction of the surface vegetation or diver-
sion of river courses would have disrupted the thermal regime and changed
the thickness of the active layer.
Figure 11.17.
Theoretical growth of thawed
cylinder around heated pipe
placed in silty soil. Pipe is 1.2 m
in diameter and kept at a
temperature of 80°C. The curves

at the left apply to conditions near
the Arctic coast, those on the
right to the southern limit of
permafrost. (After Lachenbruch
1970)
-40 -30 -20 -10 -0

m
10 20 40
Horizontal distance from centerline depth (feet)
-Arctic case 4« Interior case-
Faced with these potential hazards, the designers of the pipeline antici-
pated the problems based upon the best available knowledge of how the system
would respond. In general three approaches are employed in different places:
(1) the pipe is buried, (2) the pipe is suspended above the ground, and (3) the
pipe is placed along the edge of a road that parallels the route of the pipeline.
Burying the pipe is perfectly safe in coarse-grained soils that are well drained
and contain little ground ice. In other zones a refrigerant is run along the pipe
to keep the ground frozen. Suspending the pipe above the ground is necessary
where the soil is where the route
particularly susceptible to thawing and also
crosses the path of migratory animals such as caribou. The above-ground sus-
pension requires a dense network of pilings to provide the needed strength,
and these are inserted to great depth. In addition, some of the supporting piles
are refrigerated and others are distributed in a zigzag fashion to absorb the
effect of differential movement by heaving, gelifluction, or earthquakes (fig.
11.18).
Only time can tell us whether the design of the pipeline and the careful
planning of the overall project will provide a continuous supply of crude oil
with a minimum of hazards to the environment. Initial operation of the pipe-

line has been encouraging, although some minor leaks have occurred (Stanley
and Cronin 1983).
Chapter 11 442
Figure 11.18.
Trans-Alaska pipeline crossing the
Tolovana River valley near
Livengood. Alaska
Summary The processes and landforms found in periglacial environments have been
briefly examined. In general, periglacial conditions are those existing in any
cold, nonglacial setting regardless of latitude, but most regions of intense peri-
glaciation are polar or subpolar. In many cases periglacial zones possess a
unique property of permanently frozen ground, called permafrost. Permafrost
adds a complicating factor in the activity of periglacial processes and leads to
the formation of diagnostic features such as ice-wedge polygons, thermokarst.
and pingos that develop only where permafrost is present.
The driving processes in periglaciation are frost action and mass move-
ments. Frost action includes wedging, heaving, and cracking. Mass move-
ments usually involve frost creep and a form of soil flowage called gelifluction.
Both processes act in the development of most periglacial features, although
certain forms seem to be more common where one or the other process is dom-
inant.
The rigorous climate and the presence of permafrost lead to unusual dif-
ficulties in urban development that require engineering techniques not em-
ployed in The expected population expansion in arctic zones
other regions.
demands that we develop a sophisticated understanding of periglacial pro-
cesses in order to maintain the fragile environmental balance.
Benedict, J. B. 1976. Frost creep and gelifiuction features: A review. Quat.

Res. 6:55-77.
Black, R. F. 1976. Periglacial features indicative of permafrost: Ice and soil
wedges. Quat. Res. 6:3-26.

Embleton, C, and King, C. A. M. 1975. Periglacial geomorphology. New
York: Halsted Press.
Ferrians, O. J.; Kachadoorian, R.; and Greene, G. W. 1969. Permafrost and
related engineering problems in Alaska. U.S. Geol. Survey Prof. Paper
678.
Flemal, R. C. 1976. Pingos and pingo scars: Their characteristics,
distribution, and utility in reconstructing former permafrost
environments. Quat. Res. 6:37-53.
Goldthwait, R. P. 1976. Frost sorted patterned ground: A review. Quat Res.
6:27-35.
National Academy of Sciences. 1983. Proc. Permafrost 4th Internat. Conf.
Washington, D.C.: National Academy Press.
National Research Council of Canada. 1978. Proc. Permafrost 3rd
Internat. Conf Ottawa: National Research Council of Canada.
Pewe, T. L. 1969. The periglacial environment. Montreal: McGill-Queen's
Univ. Press.
Price, L. W. 1972. The periglacial environment, permafrost, and man.
Assoc. Am. Geog., Comm. on College Geog. Resource Paper 14.
Tricart, J. 1969. Geomorphology of cold environments. Translated by

Edward Watson. New York: St. Martin's Press.
Washburn, A. L. 1980. Geocryology. New York: John Wiley & Sons.
^
Karst — Processes and Landforms
12
I. Introduction IV. Surficial Landforms
A. Definitions and A. Closed Depressions
Characteristics 1. Dolines
II. The Processes and Their a. Solutional Dolines
Controls b. Collapse Dolines
A. Karst Rocks — The Resisting 2. Doline Morphometry
Framework 3. Uvalas and Poljes
1. Lithology B. Karst Valleys
2. Porosity and 1. Allogenic Valleys
Permeability 2. Blind and Dry Valleys
B. The Driving Mechanics and 3. Pocket Valleys
Controls C. Tropical Karst
1. Climate, Vegetation, V. Limestone Caves
and Biogenic C0 2 A. Cave Physiography
2. The Solution Process 1. Entrances and
III. Karst Hydrology and Drainage Terminations
Characteristics 2. Passages, Rooms, and
A. Surface Flow Patterns
B. Karst Aquifers and B. The Origin of Limestone
Groundwater Caves
C. The Relation Between VI. Summary
Surface and Groundwater VII. Suggested Readings
1 Springs
2. Morphometry of Karst
Drainage
445
Chapter 12 446
Introduction In regions of carbonate rocks and evaporites, weathering and erosion produce
unique landforms called karst or, when widespread, karst topography. In con-
trast to most processes studied before, karstification (the processes that de-
velop karst topography) is not easily observed because much of the geomorphic
work is accomplished well below the ground surface. In fact, some modern
textbooks of geomorphology completely ignore karst or treat it in a few de-
scriptive paragraphs because it is primarily driven by the solution process and
can justifiably be considered under the topic of chemical weathering. How-
ever, themagnitude of solution in the development of karst is so great, and
the unique topography resulting from the process so widespread, that it seems
to deserve special treatment.
Several excellent books deal specifically with the topic of karst (Jennings
1971; Herak and Stringfield 1972; Sweeting 1973, 1981; Bogli 1980; Milan-
ovic 1981), and information contained in those sources was used as a frame-
work for this chapter. In general, however, syntheses of literature dealing with
karst, and published in English, are rather uncommon.
Definitions and Characteristics

Karst is defined by Jennings (1971) as "terrain with distinctive characteristics
of relief and drainage arising primarily from a higher degree of rock solubility
in natural water than is found elsewhere.'' The definition stresses two main
points: (1) distinctive landforms and other surface characteristics developed
on highly soluble rocks, and (2) a unique type of drainage pattern resulting
from the karst processes. As all rocks are soluble to some extent, karst must
develop only on those rocks that are particularly susceptible to solution. In
such situations, the solution process can create and enlarge cavities within the
rocks. This leads to the progressive integration of voids beneath the surface
and allows large amounts of water to be funneled into an underground drainage
system while simultaneously disrupting the pattern of surface flow. Because
hydrology exists in physical "symbiosis" with solution, it becomes a very im-
portant aspect of karst phenomena. Solution integrates spaces, allowing pro-
nounced underground circulation of water that, in turn, promotes further
solution. A true karst area, therefore, possesses a predominantly underground
drainage with a poorly developed surface network of streams. As our interest
is primarily in processes, the basic concepts may be applicable in many areas
that have surface drainage and have not developed a strong enough topog-
raphy to be considered as karst. Importantly, however, the area may still be
affected by karst-forming processes.
The term "karst" is a German adaptation of the Slavic word kras or krs
and theItalian word carso, which literally mean "a bleak waterless place**
(Monroe 1970) and also connote a bare rock surface. The early and classical
description of karst was derived from a high plateau area near the Adriatic
Sea between northwest Italy and Yugoslavia. This region is characterized by
irregular topography containing many closed depressions and interrupted
stream valleys; thus, early observers of the region considered karst as a geo-
morphic freak with chaotic and disordered topographic expression. Most ge-
ologists and geographers thought of karst as a curiosity rather than a topic
447 Karst — Processes and Landforms
Figure 12.1.
Major karst areas of the United
States (Boundaries are
generalized, and some regions
shown black contain
in solid some
nonkarst areas.)
was not until 1893, when Jovan Cvijic

for serious scientific investigations. It
(name rhymes with "screech") published his book Das Karstphanomen, that
karst geomorphology was given true scientific status. Cvijic's work clearly de-
fines karst landforms and demonstrates the predominance of solution in their
development, even though the solution effect had been alluded to in earlier
studies (Sawkins 1869; Cox 1874).

Karstlands (or karst landforms) are found in almost every region of the
world, including arctic and arid zones, but they are most likely to occur in
temperate or tropical climates. In the United States, karst has developed
wherever conditions are favorable, but the major concentrations of features
exist in several general areas (fig. 12.1): (1) the Valley and Ridge province of
the Appalachian Mountains (Pennsylvania, Maryland), (2) central Florida,
(3) the plateau region of east-central Missouri, (4) a belt extending from south-
central Indiana into west-central Kentucky, (5) the Edwards Plateau (Texas).
The field of karst geomorphology has its roots in central and eastern Eu-
rope, and contributions by English-speaking geomorphologists were rather
meager during the formative years of the discipline. In recent years, however,
these geomorphologists have added greatly to our knowledge of the discipline,
especially by their recognition that analyses of karst processes and karst land-
scapes require the integration of hydrologic, chemical, and mathematical con-
cepts as well as traditional geomorphology (Palmer 1984). In addition, the
modern emphasis in karst geomorphology is largely on process and the ap-
plication of conceptual models on practical problems involving engineering,
water supply, and economic geology. Many articles espousing this thrust are
now appearing in scientific journals such as the Journal of Hydrology, Water
Resources Research, and the National Speleological Society Bulletin.
The historical development of karst science explains the proliferation of
terminology that has arisen for karst features and processes. Although the
Yugoslavian terrain inspired the study of karst, not all the names for indi-
come from there, and almost every country has developed
vidual karst features
Chapter 12 448
Table 12.1 Abbreviated glossary of common karst terminology.
Term Definition
Aggressive water Water having the ability to dissolve rocks. Especially water
containing dissolved C0 2 .
Blind valley A ends suddenly where its stream disappears

valley that
underground
Cave A natural underground room or series of rooms and passages
large enough to be entered by a person
Chamber The largest order of cavity in a cave or cave system
Closed depression Any closed topographic basin having no external drainage.
regardless of origin or size
Cockpit 1. Any closed depression having steep sides 2 A star-shaped
depression having a conical or slightly concave floor
Cockpit karst Tropical karst topography containing many closed depressions

surrounded by conical hills Similar to cone karst.
Cone karst Tropical karst with star-shaped depressions at base of many
steep-sided, cone-shaped hills.
Doline A basin or funnel-shaped hollow in limestone ranging in diameter

from a few meters to a kilometer and in depth from a few to
several hundred meters. May be distinguished as "solution"
or "collapse" if precise origin is known In U.S.. most dolines
referred to as sinks or sinkholes
Exsurgence Point at which underground stream reaches the surface if the
stream has no known surface headwaters
Karren Channels or furrows caused by solution on massive bare
limestone surfaces Synonym lapies
Karst plain A plain on which closed depressions, subterranean drainage, and
other karst features may be developed Also called karst
plateau.
Karst topography Topography dominated by features of solutional origin
Karst valley 1. Elongate solution valley 2. Valley produced by collapse of a

cavern roof
Karst window Depression revealing a part of a subterranean river flowing
across its floor, or an unroofed part of a cave.
Karstic Adjective form of karst
Karstification Action by water, mainly chemical but also mechanical, that
produces features of a karst topography
Mogote A steep-sided hill of limestone generally surrounded by nearly flat
alluviated plains Generally used for karst residual hills in the
tropics. Synonym pepmo.
Polje A very large closed depression in areas of karst topography.
having flat floors and steep perimeter walls.
Resurgence Point at which underground stream reaches the surface
Reemergence of a river that has earlier sunk upstream
Room A part of a cave system that is wider than a normal passage
Similar to chamber
Speleothem A secondary mineral deposit formed in caves
Swallet, swallow hole A place where water disappears underground in a limestone
region. A swallow hole generally implies water loss in a
closed depression or blind valley A swallet may refer to water
loss in a streambed even though there is no depression Also
ponor. sink, sinkhole, stream sink
Terra rossa Reddish-brown soil mantling limestone bedrock; may be residual
in some places
Tower karst Karst topography characterized by isolated limestone hills
separated by areas of alluvium Towers generally steep-sided
and forest-covered hills, often with flat tops
Uvala Large closed depression formed by the coalescence of several
dolines. compound doline
449 Karst —Processes and Landforms
a particular terminology of karst geomorphology. Table 12.1 presents an ab-

breviated glossary of karst terminology adopted by the U.S. Geological Survey
and followed in this discussion. The complete listing of terms can be found in
Monroe (1970).
Karst Rocks — The Resisting Framework The Processes and

It is tempting to say that karst develops primarily on limestones and leave it Their Controls
there without further elaboration. Although technically correct, such a state-
ment is grossly misleading because some limestones are not potential harbin-
gers of karst topography. Only certain limestones have the unique combination
of properties that allow them to succumb to karstification and foster karst
topography. We will briefly examine what rock properties are most conducive
to karstification and why.
Lithology As a general rock group, limestones show great variability, but

the accepted definition is that a limestone is a rock containing at least 50 per-
cent carbonate minerals, most of which occur in the form of calcite (CaC0 3 ).
Although very young limestones may contain some aragonite, the two most
common carbonate minerals in limestones are a low-magnesium calcite, con-
taining 1-4 percent magnesium, and dolomite (Sweeting 1973). If more than
50 percent of the carbonate minerals are calcite, the rock is called limestone;
if more than 50 percent of the carbonate minerals are dolomite, the rock is
called dolomite (Leighton and Pendexter 1962). The purer the limestone is
with respect to CaC0 3 the greater will be its tendency to form karst. Corbel
,
(1957), for example, suggests that 60 percent CaC0 3 is needed before any
karst will form, and about 90 percent is required to expect a fully developed
karst region. It should be noted, however, that even pure limestones may not
produce a karst terrain because the processes also depend on other factors.
Rocks other than limestone can produce karst if ancillary conditions are
proper and the material is sufficiently soluble. Dolomites are commonly karst-
but unless they are very pure their porosity and permeability tend to be
ified,
somewhat low. Evaporites such as gypsum and halite are also prone to karst-
ification. In general, however, the occurrence of karst in dolomites and evap-
orites is minor compared to the widespread distribution of limestone karst
regions.
Porosity and Permeability In addition to lithology, the creation of karst also

depends on how much water any rock can hold and how easily the water moves
through the rock system. Porosity is a measure of the water-storing capacity
of a given rock, and it is usually expressed as the percentage of void spaces in
the rock:
P = j* X 100,
where P is porosity in percent, Vv is the volume of voids, and V is the total

volume of the material. Presumably, open textures and higher porosities should
facilitate the solution process and the development of karst because the rock
Chapter 12 450
will hold more water. Some evidence exists to support this contention (Sweeting
1973). Porosity, however, consists of two types. Primary porosity relates to
intergranular void spaces created during formation of the rock. This type of
porosity, however, is commonly decreased with time by precipitation of ce-
ment, recrystallization, and change in mineralogy. In older limestones, for ex-
ample, calcite tends to be replaced by dolomite, a process that usually decreases
the primary porosity (Powers 1962). In addition, metamorphism may de-
crease carbonate porosity by inducing an increase in calcite grain size and
reducing void size by pressure. Thus, primary porosity may be an ephemeral
characteristic of limestones and perhaps is not as important in karstification
as secondary porosity and permeability.
Secondary porosity comes from openings in rocks that occur along bed-
ding-plane partings or as fractures, such as joints and fault zones. It is gen-
erally agreed that the nature and pattern of these openings may be the single
most important factor Not only do they allow the rocks to
in karstification.
hold more water, but they also promote circulation within the system by in-
creasing permeability. Because permeability (the capacity to transmit water)
depends on the continuity of voids, even rocks with high primary porosity may
develop little karst if no secondary avenues of flow are present (Tricart 1968).
Porosity, then, has real importance only if the system is also permeable.
Zones of weakness increase porosity but, more important, they have a decided
influence on the permeability. This explains why permeability rates vary by
as much as five orders of magnitude depending on the size and interconnection
of fractures and partings.
Joints are important avenues of water transport and enhanced perme-
ability. Usually joints occur in patterns with one dominant direction and a
secondary set of joints intersecting the main set at angles between 70° and
90° The spacing of the joint sets can be very significant in the genesis of karst.
.
If intersecting planes are too close, the rock may be highly permeable but too
weak to allow the fulldevelopment of karst.
Faults also transmit water effectively, but their precise role in karstifi-
cation varies according to local conditions (Stringfield and LeGrand 1969).
For example, fault zones sometimes have low permeability where voids are
occupied by secondary mineralization associated with ore deposits. Such a de-
terrent to karstification, however, may be partly offset if the ore includes sul-
fide minerals, because oxidation produces sulfuric acid, which makes the
permeating fluids more aggressive in the solution process (Pohl and White
1965; Morehouse 1968).
In summary, it seems likely that the full development of karst depends
primarily on whether water capable of solution passes through a rock sequence
along discrete flow paths with enough discharge to create significant solution
openings. Given that requirement, the process is aided by having a thick se-
quence of pure, crystalline limestones that is not interrupted by major insol-
uble beds. In addition, some relief should be available to permit free circulation
of the water in the system.
The Driving Mechanics and Controls
Climate, Vegetation, and Biogenic C0 2 Karstification requires abundant

water that is free to circulate through the karst rocks. The water not only
serves as the solvent in the development of karst but also encourages the growth
of vegetation and the soil microbial activity that add extra C0 2 to the system.
Regions of low rainfall coupled with high temperature and evaporation rates
will, therefore, be less susceptible to karst development. This is not to say that
karst features never form in arid or semiarid regions. They do Carlsbad —

Caverns in New Mexico
and karst of the Nullarbor Plain in Australia attest
to that fact. Normally, however, the topography produced is quite subdued,
and features such as collapses and deranged surface drainage are not as striking.
It is also possible that the karst in these areas formed at an earlier time when
the climate was more humid than at present or, in the case of Carlsbad, may
be the result of deep-seated acidic fluids rising into the rocks (Davis 1980;
Egemeir 1981; Hill 1981).
In extremely cold arctic or subarctic regions, the full development of karst
is hindered by the presence of permafrost and by the low vegetal productivity
that retards microbiologic activity (D. I. Smith 1969). Although water is
present, it is often frozen and no free circulation exists. The important aspect
of biogenic C0 2 (which we will discuss later) is usually missing. Karst is there-
fore a rather anomalous phenomenon in these areas.
At humid climate, with the ideal combina-

the other extreme, a tropical
tion of temperature and precipitation to drive the solution process, should be
most conducive to karstification provided enough relief exists to promote
downward and circulatory movement of groundwater. Chemical reactions
proceed more rapidly at higher temperatures, and lush vegetal cover combined
with intense microbial activity impresses the tropical soil water with high par-
tial pressures of carbon dioxide (P (0 )• The subsurface water in such regions
is very aggressive in solution, and a great abundance and variety of karst fea-
tures are found in tropical regions (Jennings and Bik 1962; Monroe 1976).
The large variation in the controlling factors of karst in different climatic
regimes has perpetuated the climatic geomorphology thrust initiated in the
1930s (Lehmann 1936). For example, Corbel (1959) analyzes rates of karst
erosion in relation to climatic parameters, and others (P. W. Williams 1963;
Douglas 1964) have modified the details of Corbel's original approach while
still maintaining its dependence on climatic variables. Nonetheless, it seems
certain that climatic factors are important mainly because they provide free
water and create the soil and vegetal characteristics that produce high bio-
genic C0 2 values. Thus, even though enhanced solution is generally related
to climate, it is the production of biogenic C0 2 that represents the important
control (Smith and Atkinson 1976; Trainer and Heath 1976; Drake 1980; Brook
et al. 1983). In fact, minor vegetation changes within the same climatic zone
(even arctic or subarctic) will generate areas of greater solution (Woo and
Marsh 1977; Brook and Ford 1982). Consequently, recent work has placed
less emphasis on climatic variables and more on solution mechanics.
Chapter 12 452
The Solution Process The solution process itself is in reality the critical
function in the entire analysis of karst. Regardless of how conducive climate,

lithology, fractures, and other variables are to karstification, karst topography
would never develop if the solution process were somehow rendered inopera-
tive. Its function or malfunction fundamental to the topic, and we must
is
attempt to understand its mechanics, at least in its simplest terms.

Laboratory studies tell us that the mineral calcite, like all common min-
erals, is soluble in pure water. At saturation it is soluble to the extent of about
12- 15 ppm depending on the temperature of the water. This solubility is rather
when compared +2
startling to natural river waters where concentrations of Ca
and bicarbonate (HC0 3 ~) are much greater (Livingstone 1963), indicating a
substantial increase in the solubility. Since we are considering the same sub-
stance (calcite), it is obvious that the solvent in the natural system is not pure
water. Rainwater, in fact, is not pure because it incorporates a variety of chem-
ical constituents as it passes through the atmosphere. The most important of
these is carbon dioxide (C0 2 ), which is soluble in pure water: some of the
dissolved C0 2 reacts rapidly with the water to form a weak acid (H 2 CCm.
called carbonic acid
C0 2ldissohedl + H 2
- H 2 CO,. (1)
This acid, however, is always dissociated into its ionic state, and the above
reaction can be expressed more realistically as
C0 2(dissoliedl + H 2
- H + + HC03- (2)
The amount of C0 2 actually dissolved in water depends on the partial pressure

of carbon dioxide (Pco ) in the air standing at the air-water interface and on
the temperature of the"water. The air in contact with the water can be in the
atmosphere, or in spaces within the soil, or in subterranean cavities such as
caves. In any case, the amount of C0 2 dissolved in the water increases as the
Pco of the air increasesand as the temperature of the water decreases. Colder
water will dissolve more C0 2 than warm water at any given Pco value. In
the atmosphere Pco israther small, having values averaging about 0.03 per-
cent of volume (3 X 10~ 4 bar); similar values are found in most caves (Hol-
land et al. 1964). Anomalously high Pco values are found in air contained
within soil or the vegetal litter covering it. Values of 1-2 percent of volume
are common, and some poorly ventilated tropical soils may contain as much
as 20-25 percent (Jennings 1971). The abnormal CO ; values in soil air. and
the resulting large amounts of dissolved C0 2 in the soil water, stem from mi-
crobial action involved in the decomposition of vegetal matter. This represents
the biogenic C0 2 briefly discussed in the previous section. It is regarded by
most karst experts as being the prime ingredient in the solution process.
Calcite itself is dissociated into an ionic state such that
CaCCWi.e, - Ca +2 + CO," 2
-
453 Karst —Processes and Land forms
However, the C0 3 ion produced quickly reacts with the H + formed when C0 2
is dissolved in water (reaction 2), and thus the dissociation of calcite also pro-
duces a bicarbonate ion because
C(V + H + ^ HC0
2
3
-.
(4)
It is clear from these reactions that the solution of limestone revolves

around the CaC0 3 -C0 2 -H 2 chemical system. This system is extremely
complicated, and its mechanics much more sophisticated than this introduc-
tory treatment can show. We are not dealing with a single reaction that pro-
duces solution of calcite, but a process that involves a series of reversible and
mutually interdependent reactions, all proceeding at different rates, and each
regulated by different equilibrium constraints. We will therefore explain the
process only in the most general terms.
If we combine reaction (2) and (3), a general form of the process can be
expressed by the following:
CaC0 3 + H 2 + C0 2(dissolvedl
- Ca +2 + 2HC0 3 ". (5)
The bicarbonate ions (HC0 3 ~) are derived from two sources shown in equa-
tions (2) and (4). In equation (4), the reaction of carbonate ions (from the
dissociation of CaC0 3 ) and the H+ (from the dissolving of C0 2 in water) pro-
duces a disequilibrium between the Pco in the air and the Pco in the water.
This causes more C0 2 to diffuse from the air into the water and allows further
solution of the calcite because reaction (5) is driven to the right side of the
reversible equilibrium.
To a large extent, the amount of C0 2 dissolved in water regulates the
solubility of limestone. Where the amount of C0 2 dissolved in water is high,
the fluid will agressively attack the calcite. Soil water with its high PCo is
therefore the most effective solution agent (Drake and Wigley 1975).
Another important mechanism in the solution process is mixing corrosion
(Bdgli 1964). Essentially, when two bodies of water at equilibrium with dif-
ferent C0 2 contents are mixed, the resulting fluid needs less C0 2 to establish
equilibrium than the sum of the C0 2 contained in the two original fluids. Some
C0 2 is released and becomes available to promote further solution. The pro-
cess is very significant where seawater mixes with fresh water in carbonate
regions. In that situation, mixing produces brackish water that is subsaturated
with respect to calcite and enhances the solution process (Runnells 1969;
Plummer 1975; Hanshaw and Back 1980). Even there, however, the phenom-
enon is very complex, and the degree of saturation with respect to calcite may
depend on other factors (Plummer et al. 1976).
In normal karst situations, the effect of mixing really depends on the type
of water masses being mixed (Thrailkill 1968). For example, water slowly
percolating through the zone of aeration, a fluid called vadose seepage, may
not show a significant change when it mixes with phreatic water at the water
table. However, two types of vadose seepage are known (Thrailkill and Robl
1981), and each may respond differently upon mixing. In addition, water rap-
idly introduced to the subsurface, called vadose flow, may trigger considerable
Chapter 12 454
mixing corrosion when it meets the liquid standing at the water table (see
Mixing of vadose flow and the proper vadose seepage may

Thrailkill 1968).
cause pronounced undersaturation of the underground water and promote the
solution process (Thrailkill 1972).
Solution rates are usually estimated by direct measurement of the dis-
solved load in streams leaving a karst area, although other techniques can be
employed (for an excellent discussion, see Jennings 1983). The rate is equal
to the concentration of dissolved load multiplied by the stream discharge such
that
S = CQ,
where 5 is the solution rate in units of mass/time, C is the solute concentration
(mass/volume), and Q is stream discharge (volume/time). Many measure-
ments can be integrated over time and converted into rates of surface lowering
or karst denudation. Values of karst denudation, adjusted for subsurface con-
tribution, seem to range from about 10 mm/ 1000 yr to over 100 mm/ 1000
yr, depending primarily on the amount of runoff (Atkinson and Smith 1976;
Jennings 1983).
The controls on solution rates involve a complex maze of interrelated fac-
tors. Sweeting (1973) suggests that nearly all the possible limestone solution
is accomplished within the first minute of contact, but for all the reactions to
establish equilibrium throughout the system it may take anywhere from 24 to
60 hours. In addition, Jennings (1983) shows that higher discharge produces

greater solute load. In light of this, it is generally assumed that water moving
through the system rapidly and with some turbulence will ultimately dissolve
more limestone, whereas stagnant water may become supersaturated with re-
spect to calcite as equilibrium conditions are attained throughout the various

reactions in the system (Kaye 1957; Weyl 1958). Other analyses, however,
suggest that the dependence of solution rates on velocity and turbulence only
holds when the fluid is very acidic (see Palmer 1984). Where pH is greater
than 4, the solution rate is primarily dependent on the reaction rate at the
bedrock surface and may be almost independent of velocity.
In addition to flow velocity and pH, as water moves through underground
fractures the solution rate decreases with flow distance. This occurs because
the water gets closer to saturation with respect to calcite the farther it travels.
In the past,arguments have been made that the decrease in solution rate is
linear with percentage of saturation and therefore travel distance. Palmer
(1984), however, points out that under such controls the transport distance for
complete saturation would be so short that long, linear caves would be diffi-
cult, if not impossible, to develop. Based on earlier studies of calcite solution
by water near saturation (Berner and Morse 1974; Plummer et al. 1978).
Palmer (1984) presents cogent arguments that solution rates do not decrease
linearly with percentage of saturation except at lower values. At some point
above 65 percent saturation, the solution rate drops rapidly (fig. 12.2). This
allows solution to continue for a greater time and over a greater distance, and
it helps to explain the origin of elongated cave systems.
Max
Figure 12.2.
Solution rate versus degree of
saturation. Instead of decreasing
linearly, the solution rate drops
sharply to a low level at 65-90
3 percent saturation
50% 100%
Percent saturation
Surface Flow Karst Hydrology

Karst regions are unique because their drainage networks characteristically and Drainage
are disrupted, and few rivers can traverse such areas in a continuous and un- Characteristics
segmented manner. The reason for this strange fluvial behavior is the facility
with which surface flow can be diverted into the underground system. De-
pending on soil types, vegetation, and joint spacing, overland flow on hillslopes
may be drastically reduced by infiltration, and in extreme cases no water will
enter the nearby channels.
Rivers also lose water when some of the flow descends into swallow holes
or swallets, shown on the map in figure 12.3. These are nothing more than
open cavities on the channel floor that are capable of pirating a portion of a
river's discharge or even the entire river (especially during low flow) into the
underground system. Thus, a large part of the total flow of rivers in karst
regions may follow a subsurface route that may or may not parallel the path
of the river valley on the surface. Rivers that are able to cross a karst terrain
as continuous surface entities have distinct hydrologic characteristics. For ex-
ample, flood records for 114 basins of different sizes show that the mean an-
nual floods (recurrence interval of 2.33 years) in carbonate basins of
Pennsylvania were considerably lower than those in basins underlain by dif-
ferent rock types (White and Reich 1970). Commonly, however, the peak flow
is spread over a longer period as the subsurface water is slowly released to the
rivers. It appears likely that the precise hydrologic character of surface rivers
in karstic areas depends greatly on the state of development of the under-
ground drainage, especially the degree of interconnection between subsurface
passageways (Ede 1975).
Karst Aquifers and Groundwater

Our brief look at normal hydrogeology (chapter 5) allows us now to examine
the differences between karst aquifers and the groundwater conditions in other
porous and permeable materials.
Chapter 12 456
Historically, karst geomorphologists have viewed the hydrology in karstic

systems in discordant ways. Most conflicting opinions concerned the type of
water movement, the depth to which groundwater would penetrate, and most
important, whether or not a water table as we normally conceive it exists in
karstic terrains. In the first two decades of the twentieth century, two schools
of thought developed concerning the hydrologic system.
In one model, groundwater in karst regions is considered to be similar to
the normal groundwater found in unconfined aquifers composed of other rocks
or unconsolidated debris. The water presumably exists in a vertical stratifi-
cation. Water in the zone of aeration moves downward to a pronounced water
table by seepage or by flow in various types of fissures and cavities. Below the
water table the phreatic zone is completely saturated, although the mechanics
of how water moves through this zone are very controversial. The water table
itself is subject to significant fluctuations, rising and falling on a seasonal basis.
Figure 12.3.
The distribution of swallets, sinks,
and springs in a karst region of
southeast New York. (From Baker
1976)
The second major conceptual model of karstic groundwater completely

denies the existence of a water table and with it the distinct vertical zonation
of the water. Workers accepting this view believe that the distribution and
movement of groundwater are controlled entirely by the spatial characteristics
of the network of interconnected passageways within the rocks. Several points
support this conclusion: (1) adjacent wells drilled into limestone often have
hydrostatic levels that are significantly different in elevation; (2) tunnels re-
veal dry fissures immediately next to cracks that are filled with water;
(3) tracing of water with dyes shows that paths of movement often cross one
another (obviously flowing at different levels) and even pass under a surface
stream from one side of a valley to the other, a physical impossibility in normal
unconfined aquifers; and (4) poljes (see table 12.1) at the same level do not
behave in a similar manner, some flooding in winter while others remain dry.
Thus, the underground system is thought to be a collection of conduits func-
tioning like three-dimensional rivers.The passages may be totally intercon-
some cases they may function like a single confined aquifer having
nected, or in
a recharge area, a discernible pathway, and a separate point where the water
emerges once again to the surface. Most geomorphologists believe that these
two basic models can simultaneously exist in the same karst area.
A relatively recent development has been to recognize the importance of
a hydrologic zone existing in the weathered, upper portion of the bedrock but
beneath the soil cover (see Williams 1983). This zone, the subcutaneous zone,
stands above the phreatic zone, but it stores water and is periodically satu-
rated, especially after storms. The subcutaneous zone develops by enhanced
solution immediately beneath the soil; this enlarges cavities and fractures and
Figure 12.4.
creates high permeability and porosity. The solution enlargement decreases
Subcutaneous zone in limestone
with depth, however, and at some level, openings are too small to transmit along Route 55 near Perryville,
water at the same rate as in the overlying more permeable rock (fig. 12.4). Mo.
ftt _v a •***' *&/$•£. «-* <J5 : i^

•
M-. _
Chapter 12 458
Figure 12.5.
Classification of flow types in
karst aquifers (From W B White Stream

1969) Karst landforms rare and/or
subdued
T
— Generally low relief surface
Random arrangement of
small cavities
Relatively deep flow
(A) Diffuse flow
Sinking
stream
Shaft formation at edge

of cap
Stream in open channel
under large overburden
Slope of channel determined

by slope of recharging bed
(B) Freeflow
Recharge
Water table
Water percolates through

imestone under confining beds
(C) Confined flow
As a result, water is stored in the form of a perched water table. The perched
water table slopes toward points of rapid vertical percolation (major joints,
faults, shafts, etc.), and a lateral component of flow develops in the subcu-
taneous zone and converges at those points. Subcutaneous flow has a signifi-
cant influence on the hydrologic characteristics of the system and on the pattern
of landforms developed in karst regions.
W. B. White (1969) has classified carbonate aquifers according to their
hydrogeologic properties as diffuse flow, free flow or confined flow, each type
having subtypes; the main classifications are depicted in figure 1 2.5. In diffuse
flow, cavities are limited in size and numbers, caves are rare, a well-defined
water table is present, and flow obeys or nearly obeys Darcy's law. In free-
flow aquifers, the water moves through integrated conduits under the influ-
ence of gravity, often attaining turbulent flow. The flow is capable of trans-
porting sediment as discharge is enlarged by surface runoff sinking into
fractures (swallets). Discharge of the groundwater is usually through large

springs that accumulate, at a single outlet, the water flowing through vast
areas of underground drainage. Confined-flow aquifers are characterized by
water that moves in response to pressure. They may be true confined aquifers
as described earlier and may contain water under considerable hydrostatic
head; thus, artesian flow conditions are possible in these aquifers.
It is now becoming clear that the different types of aquifers recognized
by White may also be distinguishable on the basis of the chemical behavior
of their water. Shuster and White (1971), analyzing 14 springs in the central
Appalachians, suggest that diffuse-flow aquifers maintain a constant hardness
and have spring water nearly saturated with respect to CaC0 3 Aquifers that .
utilize conduits are undersaturated and show considerable variability in hard-

ness (Thrailkill 1972; Drake and Harmon 1973). More recently, Kroethe and
Libra (1983) have identified two different flow types in the southern Indiana
karst region on the basis of sulfur chemistry in spring waters.
The Relation Between Surface and Groundwater

Assuming that flow through conduits is the unique characteristic of karst hy-
drology, we should briefly examine how water is exchanged between the sur-
face and the underground system and vice versa, and how the properties of
the passageways control the movement. In karst areas, water is diverted into
the groundwater system through vertical openings and swallets. In addition,
interconnected spaces in the zone of aeration permit slow percolation of va-
dose seepage, and this water is added to the total accumulation of ground-
water. In closed depressions of a karst landscape, water input to the
underground system is through a highly permeable point at the base of the
depression. The point input usually represents an intersection of vertical joints
or cylindrical solution openings called shafts. Precipitated water reaches the
input point by various transmission routes acting on the sloping surfaces of
the enclosed depression. These routes are similar to those discussed in chapter
5, with the added process of subcutaneous flow. Gunn (1981, 1983) suggests
that various possible combinations of flow types will have a decided effect on
the extent of solution and the morphologic evolution of the karst.
Springs After entry into the underground system, water moves as one or
more of the types introduced above. The manner in which groundwater leaves
the system is and flow type. In diffuse aquifers,

also dependent on the aquifer
water will emerge according to the rules that govern normal groundwater flow.
Some question lingers about the depth to which phreatic water will penetrate,
and we will discuss this problem when we review the origin of caves. In conduit
systems, however, water usually reaches the ground surface in the form of
springs,and the characteristics of these interesting geomorphic features com-
monly reveal the distribution and properties of the aquifer itself.
Springs in karst areas are usually larger and more permanent than those
in other regions because of the high infiltration associated with karst rocks.
Those springs or seeps stemming from flow in diffuse aquifers are called
exsurgences (fed by seepage), in contrast to those fed by groundwater moving
Chapter 12 460
Fault
Figure 12.6.
on the position of
Fault control
springs Rainwater enters shales
exposed at the surface along
faults Water follows fault zones
into and through limestones until
itemerges as springs (Arrows
show movement paths of water.) Spring
through distinct conduits, which are called resurgences. This distinction is often
confused since normal diffused flow may encounter cavities during its move-
ment and ultimately emerge mixed with water derived from sinking streams
(Thrailkill 1972). In general, however, resurgences are much more inconsis-
tent in their discharge and chemistry. In some humid areas, springs respond
quickly to changes in surface flow (White and Schmidt 1966: Baker 1973a).
Baker (1973a), for example, found high, turbid discharges in New York springs
during the spring season when snowmelt and rainfall produced high flow in
the surface streams. During fall and winter, however, the discharge from springs
was less by two orders of magnitude, and the water was clear. In other cases,
however, the response of spring discharge to precipitation is significantly af-
fected by storage and flow in the subcutaneous zone (Williams 1983).
The physical controls on springs that rise from conduit systems are them-
selves quite variable. Some springs represent the emergence of an aquifer
stream that flows through a cave system under the influence of gravity. Other
springs reach the surface under pressure by rising up cavities or fractures. In
most cases, however, springs are ultimately controlled by large structural fea-
tures or by stratigraphic relationships. Faults sometimes serve as the locus of
large artesian springs, especiallyif the structural relationships produce a large
potential differencebetween the recharge area and the fault (Burdon and Sa-
fadi 1963). In other situations, faults may deflect underground flow until it
emerges into a valley that stands well below the level of the recharge zone (fig.
12.6). An excellent example of this is in the area of the Kaibab Plateau in
Arizona (Huntoon 1974). The plateau north of the Grand Canyon supports
no surface streams, and all consumed by evaporation infil-
precipitation not
trates into porous surface rocks. Huntoon (1974) observes that water collected
in those rocks is transmitted to faults that act as drains, funneling the water
downward to carbonate rocks that stand 3000 feet below the plateau surface.
The water is then discharged through springs at the base of the canyon.
Regional stratigraphy and the dip of beds also may control the position
of springs. Baker (1973a) showed that in New York State, spring placement
often results from water moving down dip in well-integrated conduits floored
by rocks with low permeability (fig. 12.7).
Morphometry of Karst Drainage A trend in karst geomorphology has been

the attempt to describe karst drainage in quantitative terms (P. W. Williams
1966. 1971, 1972a, 1972b; LaValle 1967. 1968; Baker 1973a). Results of these
Figure 12.7.
Stratigraphic control on the
movement of groundwater Water
Bedrock moves down dip at contact of
Karst Plateau
Monawk -
1400'
benches limestones and impermeable
Barton Hill (recharge area)
Hudson
1200'
^y Fox Creek
m/ft^-
Lowland shale (From Baker 1976)
Valley
1000' .^—gj^k Spring +. ""* '
_
800'
^^_
600' ^j^fij wSk ^J

400'
4000 8000 Feet
Glacial deposits j| Shale 1220 2440 Meters
Datum is mean
sea level vertical
Siltstone I Shaly limestone exaggeration x 10
Limestone Direction of ground water movement
Figure 12.8.
Components of karst
morphometry and hydrology in
New Guinea as ordered by P. W.
Williams.
\ 2 \ f] ^\ I \ V. \« \ ' 2 *• A
f/O !
W \ as ( -/Vv /
Km.
*V> Topographic Divide. • Stream-sink.
Intermittent Stream Channel. 1.2,3 Basin Order.
Summit.
studies show karst drainage to be well organized rather than having the cha-
otic nature assumed by early investigators. Williams devised a method, shown
in figure 12.8, of ordering stream segments following the Strahler method such
that every swallet would accept the drainage of a particular order. After each
sink is ordered, measurements of morphometric parameters are made and
plotted against the order hierarchy to demonstrate a statistical relationship.
Baker (1973a) extended the Williams methodology by including in the anal-
ysis underground links between the swallets and the springs that were iden-
tified by mapping and dye tracing. Stream orders so designated also showed
Chapter 12 462
100
Figure 12.9.
Morphometry relationships of the
karst hydrology, including
subsurface segments, in an area
of eastern New York state
100
c
Qj
E
o
«J
E
CO
CD
CO
-D
E
3
2 3 4 6 8 10
Stream order
a high correlation to the number of streams and total length of streams (fig.
12.9). In addition, streams flowing in conduit systems seem to maintain lon-

gitudinal profiles that are similar to those of surface stream channels in the
same area (White and White 1983).
Surficial Landforms The solution process, working on rocks with diverse properties, results in a
number of surficial landforms that define a true karst. The features range in
size from tiny modifications of exposed limestone outcrops to large depressions
and hills that dominate the topography. However, they all manifest the cor-
rosional process. Unless otherwise noted, the features discussed originate in

the humid-temperate climatic regime; a special section will consider humid-
tropical karst because of the widespread occurrence and anomalous topo-
graphic forms generated under tropical conditions.
Closed Depressions
If someone were to ask what kind of landform best typifies a karst terrain, the
answer would have to be closed depressions. Although these depressions range

in size from tiny holes to those covering wide areas, they all have in common
the property of supporting no external surficial drainage. In addition, it seems
unlikely that widespread surficial depressions can develop unless they are con-
nected to an underground conduit system in which water is free to flow to a
spring outlet at a lower level (Palmer 1984).
Figure 12.10.
Ground surface pitted by dolines.
Dolines By far the most common karst landform is a closed hollow of small Monroe County, III
or moderate size called a doline (in the United States often called a sink or
sinkhole). Dolines are usually wider than they are deep, having diameters
ranging from 10 to 100 m and depths between 2 m and 100 m. In plan they
are circular or elliptical, but their cross-profile shape can vary considerably
from the normal funnel-like form to shapes resembling a disc, bowl, or cyl-
inder. Occasionally isolated dolines occur, but more commonly they are abun-
dant enough to provide a karst terrain, such as the one in figure 12.10, with
a strongly pitted appearance. For example, Malott (1939) estimates that as
many as 300,000 dolines exist in the karst region of southern Indiana. As figure
12.1 1 illustrates, dolines can justifiably be considered as the fundamental ele-
ment of karst because when present in large numbers they substitute for the
valleys that dominate the normal fluvial environment.
Chapter 12 464
Figure 12.11.
Area of southern Indiana showing
well-developed karst topograph^
Note the predominance of dolines
and the absence of surface
drainage (Corydon East
Quadrangle, Indiana. US Geol.
Survey 7 1/2' quadrangle Contour
interval = 10')
38" 07-30-'
86* 07'3O"
The term "doline" has suffered through enough different connotations to

prompt some authors to call for its elimination, but its use is so widespread
that its removal is virtually impossible. In fact, Cvijic (1893) used the term
in his classic book on karst, and classifications of doline types are firmly es-
tablished (Cramer 1941). A variety of doline types have been described, in-
cluding solutional dolines, collapse dolines, alluvial dolines, and solution
subsidences (Cramer 1941). We will focus on the solutional and collapse do-
lines, which are the predominant forms. These are treated as separate features
even though most dolines are a combination of the solutional and collapse types.
Solutional Dolines Waters infiltrating into joints and fissures enlarge the
cracks by solution and create a closed surface depression called a solutional
doline. Many reports have documented the surface-downward origin, and it
seems clear that this form represents the paramount doline type. As joints
enlarge, especially at the intersection of joint trends, the surface is depressed,
and internal drainage fosters continued development of the feature.
Like all geomorphic features, solutional dolines develop best where con-
trolling factors are combined in a particular way. The factors most conducive
to the formation of solutional dolines are these:
Figure 12.12.
Doline near Dongola, Union
1. Slope. Since ponding or retardation of flow accelerates infiltration, County, III
the frequency of solutional dolines is inversely proportional to the

surface gradient. Steep slopes promote rapid flow across the
surface, and so valley floors or gently undulating plains are the best
places for solvent action to initiate the process. Dolines formed on
steeper slopes tend to be asymmetric; however, that phenomenon
can also be produced in other ways.
2. Lithology and structure. Porous limestones are less susceptible to
solutional doline formation than dense limestones that are well

jointed. The joints allow selective solution rather than a uniform
corrosion over the entire surface. Solutional dolines can form in
dolomites, but the features are usually deeper with steep, rocky
sidewalls. Structures tend to align and elongate the dolines parallel
to the major trends (LaValle 1967; Matschinski 1968; Kemmerly
1976), but the degree of control depends on many variables.
3. Vegetation and soil cover. Soil and vegetal cover usually increase
solution activity because of the C0 2 factor. Other factors being
equal, solutional dolines will develop more rapidly under an
organic-rich cover than where surfaces are bare. Trees seem to be
especially important in this process (fig. 12.12).
Chapter 12 466
Figure 12.13.
Sequential development of
collapse dollne
Arch before collapse
r
— i
Collapsed debris
(C)
Collapse Dolines Collapse dolines differ from solutional dolines in that the
depressions are initiated by solution that occurs beneath the surface. Expan-
sion of caverns, caused by corrosion and by the roof material falling under
gravity, decreases the support of the overlying rock material. In a study of
Tennessee karst, Kemmerly (1980a) suggested that most collapse dolines occur
in the partially weathered residuum overlying the solid bedrock. In this pro-
cess, vertical fractures beneath the residuum are gradually widened by solu-
tion. This creates a bridge of unconsolidated debris, a regolith arch, that is
supported by pinnacles of the underlying bedrock (fig. 12.13). Further wid-

ening of the arch supports by solution and simultaneous spalling of debris from
beneath the arch of sediment makes it impossible to support the overlying
mass, and collapse occurs. Collapse dolines tend to have greater depth/width
ratios than solutional dolines. Their sidewalls are characteristically steep and
rocky, and the bottom is filled with fragments of the collapsed debris.
The collapse process is facilitated in humid environments where under-
ground drainage is well established, and especially in the downstream portion
of that drainage, where dolines seem to be deeper and more numerous (LaValle
1967). Some evidence suggests that collapsing can be initiated by rapid low-
ering or repeated fluctuation of the water table (Kemmerly 980a). Although
1
geological processes such as river entrenchment can cause this, the association
of drawdown with collapse dolines demonstrates that human action can be a

geomorphic agent, commonly with catastrophic results. A classic example of
human intervention into the natural balance was reported by Foose (1967).
Near Johannesburg, South Africa, a mining company lowered the water table
by an extensive pumping program in order to have access to deeper ore that
existed in the phreatic zone. Several years after the project's completion, large
collapse dolines, some 125 m in diameter and 50 m deep, began to form sud-
denly and with disastrous results. In December 1962, an ore refining plant
dropped 30 m into a doline as the surface suddenly collapsed, and 29 men
were killed in one terrifying moment.
Doline Morphometry The belief that karst forms are controlled by climate
was the prevailing concept in karst geomorphology after World War I. This
reliance on climate to explain karst processes and form maintained its dom-
inance until investigators became concerned that too much variability of form
existed in the same local region where climatic controls were identical (Jen-
nings and Bik 1962; Verstappen 1964). Since the 1960s, investigators have
attempted to explain karst in terms of geologic and hydrologic variations, and
studies of spatial morphology of the karst features have appeared. For ex-
ample. Hack concluded as early as 1 960 that doline density varies significantly
with rock type in the Shenandoah Valley of Virginia (Hack 1960a).
Although analyses of linear karst features have been made (White and
White 1979), most morphometric work has concentrated on the spatial char-
acteristics of dolines, relating those to some geologic or hydrologic variable
(LaValle 1967, 1968; Matschinski 1968; Williams 1971, 1972a, 1972b; Kem-
merly 1976; Palmquist 1979; White and White 1979; Mills and Starnes 1983).
These studies have been made in many different parts of the world and under
a variety of geologic and climatic conditions. For example, Williams (1966b,
1971, 1972a, 1972b) developed a sophisticated spatial analysis of tropical karst
topography. Using aerial photographs or topographic maps, the karst terrain
was separated into divides, summits, channels, and stream sinks. Each closed
depression was given a number representing the highest stream order of the
drainage that disappears into the sink (fig. 12.8). The topographic divides sur-
rounding the depressions form a polygonal network, indicating that the terrain
is completely partitioned into separate and adjoining basins. This topography,
which Williams calls polygonal karst, is dominated by the hills but dynami-
cally controlledby the position of the sinks. Williams analyzed the pattern of
sinks by measuring the average distance from each sink to its closest neighbor
and comparing that value to the expected mean distance, which is determined
from a density analysis of the sink population. The index ratio La/Le (where
La is the mean actual distance and Le is the mean expected distance) tells
whether the sinks have a random or uniform distribution. In New Guinea the
stream-sink dispersion is highly uniform, which Williams (1971, 1972a,b) in-
terprets as the best accommodation that can be made as depressions compete
for space when the topography evolves under the processes of doline forma-
tion.
Chapter 12 468
The example just given reveals the primary assumption of all morpho-
—
metry studies that derived relationships will provide insight into the basic
factors controlling the morphometry and, through that insight, an under-
standing of process. In fact, Palmquist (1979) identified the independent vari-
ables that initiate and enlarge dolines as ( 1 ) groundwater recharge,
(2) secondary permeability, (3) regolith thickness and shear strength, and
(4) hydraulic gradient. From these he proposed a process-response model in
which the initiation and enlargement of primary dolines would subsequently
lead to the generation of secondary dolines. The implication is that mixed do-
line populations can exist within the same karst area.
In a recent study, Kemmerly (1982) demonstrated that such a mixture of
primary and secondary dolines is present in the Western Highland Rim area
of Kentucky and Tennessee. In that region, one population consists of large,
joint-controlled dolines having second-order (or higher) internal drainage and
wide spacing. A second population has smaller, first-order swallet depressions
that exhibit no joint control. The large dolines are apparently primary in na-
ture and reflect high groundwater recharge, permeability, and hydraulic gra-
dients. Presumably, these may even develop in a manner similar to the
subcutaneous flow model proposed by Williams (1983) and depicted in figure
12.14. The smaller dolines, however, were probably developed where recharge,
permeability and hydraulic gradient were considerably less.
Assuming the preceding models are correct, we are led to several perti-
nent observations about karst depressions. First, the initiation of secondary
dolines in the manner suggested by Palmquist (1979) requires that those
depressions must be linked to an unbroken conduit system. It also follows that
no depression can form until sediment covering the swallet is flushed into and
through that system (Palmer 1984). Second, the distinction of primary and
secondary dolines indicates that time is an important factor in the develop-
ment of a full karst, a fact that has been emphasized by many researchers
(Cooke 1973; Palmer and Palmer 1975; Wells 1976; Kemmerly and Towe 1978;
Kemmerly 1980).
The emphasis on morphometry does not mean that climate is irrelevant
in karst processes. Obviously, it is an important factor, mainly because it pro-
duces the water needed for karst processes to work. Given that, it is local quirks
of the environmental setting that determine karst characteristics.What mor-
phometry should demonstrate is that given sufficient time, karst landforms,
like most surface features, may approach and possibly attain a dynamic equi-
librium, the properties of which are determined by the relationship between

process and local controlling factors.
Solutional doline :*-

Figure 12.14. v. Soil
Formation of solutional doline I
aided by subcutaneous flow that Perched
"
V^f-
is directed to zones having subcutaneous
1/
enlarged fractures and high , water
|
permeability
v
1
t III
Rapid flow down enlarged fractures
Uvalas and Poljes Closed depressions of larger size than dolines are called
uvalas or poljes. Uvalas form as dolines enlarge and coalesce into hollows with
undulating floors, the irregularity being produced by the differences in size of
the integrated dolines. Jennings (1967) reports a single uvala that was con-
structed from 14 separate dolines of diverse size and shape. Uvalas have no
specific size requirements as they range from 5 to 1000 m in diameter and
from 1 200 m in depth. Their plan shape can be highly irregular as a result
to
of their unique origin.
Poljes are relatively large closed depressions with flat bottoms and steep
sides. They are irregular in plan and usually are elongated along the strike of
bedding or some zone of structural weakness. Thus, they can be structurally
or lithologically controlled, and some expand by pronounced lateral corrosion
when they are temporarily filled with water. Gams (1969) places a rather vague
minimum size requirement on poljes, suggesting that they must be at least
several square kilometers in area.
Poljes often abut against impermeable and nonsoluble rocks, and rivers
flowing through those rocks may extend partly across the polje surface before
they sink. In times of high flow the sink may not be able to absorb the dis-
charge, and shallow lakes occupy the polje basin. In dry seasons evaporation
may destroy the lakes. As discussed earlier, polje lakes may also form and
disappear with changes in the underground hydrology.
Karst Valleys
The second major topographic group in karst topography is karst valleys, which
show a variety of charcteristics. In general they can be divided into several
types with clearly discernible properties.
Allogenic Valleys Allogenic valleys head in impermeable rocks adjacent to

the karstic area. As the surface flow originating in the nonkarstic rocks enters
the karst region, it forms spectacular gorges with steep, canyonlike walls. For
instance, the Tarn gorge in the Grands Causses area of France is 2 km wide
and 300 m deep. Such magnificent valleys obviously require considerable dis-
charge to develop, with a combination of solution and fluvial abrasion as the
driving mechanics.
Blind and Dry Valleys Most surface rivers traversing a karst surface, no Dry valley Blind valley
matter where they originate, eventually sink into the underground system. They
may disappear into holes within the channel (swallets or swallow holes), or
the entire drainage may be inside a doline and so all surface water (confined
in a channel or not) sinks into the base of the doline. As figure 12.15 shows, Sink
rivers flowing across a karst surface in well-defined valleys tend to lower the
valley floor upstream from the sink faster than the reach downstream from
the sink because the flow is drastically diminished at the point of infiltration.
Figure 12.15.
Eventually a limestone scarp may develop, separating the two channel reaches.
Longitudinal profile of a blind
This scarp can vary in size from a few meters in small streams to tens of meters valley and dry valley.
in larger rivers (Malott 1939). Usually the cliff increases in height with age.
Rivers terminating at the cliff face are said to be occupying blind valleys.
Chapter 12 470
Dry valleys have all the properties of normal fluvial valleys except, as the
name implies, they have no well-defined watercourses or carry only ephemeral

flow in response to massive floods upstream. They probably represent the most
common form of karst valleys and, as mentioned earlier, are the most favor-
able locale for doline formation because rainwater tends to pond on the valley
surface and sink into the rock fractures. Like all karst valleys, their sides tend
to be steep, but factors of lithology and age cause some variability in their
cross-profile shape. The complex and varied, but
origin of dry valleys can be
in the simplest and most common case they represent the downstream reach
of a blind valley that absorbs all the surface flow at a particular sink (fig.
12.15).
Pocket Valleys Pocket valleys are essentially the opposite of blind valleys
in that they begin where groundwater resurges rather than where it sinks.
They are normally associated with large springs that resurge on top of an im-
permeable substrate at the foot of a thick exposure of karst limestone.
Pocket valleys, sometimes called steepheads, are usually U-shaped in cross-
profile, having steep sidewalls. They characteristically have a steeply inclined
headwall that may be recessed by spring sapping that undermines the over-
lying limestone. Their dimensions vary depending on the spring discharge and
the nature of the karst rocks, but some are 8 km long, 1000 m wide and 300-400
m deep (Sweeting 1973).
Tropical Karst
The normal karst features just described can be compared with the spectac-
ular karst topography developed in humid-tropical climates. Once again, we
find that studies of tropical karst carried on in diverse regions of the world
have resulted in a confusing array of terms describing the same features.
Nonetheless, it is safe to say that every landform recognized in karst terrains
of temperate regions is present in tropical karst. What sets tropical karst apart
seems to be the fact that the general landscape is dominated by residual hills
rather than the closed depressions so characteristic of the temperate karsts.

Most of the residual forms occur as steep-sided, cone-shaped hills that, com-
bined with surrounding depressions, constitute a particular type of karst to-
pography known as cone karst (Ger. kegelkarst). The subforms within the
general category of cone karst are numerous and transitional, but two topo-
graphic components called cockpits and towers seem to be the uniquely di-
agnostic elements of tropical karst; in fact the terms "cockpit karst" and "tower
karst" are deeply entrenched in the literature.
Cockpits are similar to temperate-climate dolines except that they are
usually irregular or star-shaped depressions that surround the residual hills
(fig. 12.16). Cockpit karst was first named in Jamaica and was
described and
originally ascribed to solution along joints and faults (Lehmann 936; Sweeting 1
1958). Later work by Aub (as quoted by Sweeting 1973) and P. W. Williams
(1971) suggests that the star shape of cockpits arises from gulley erosion of
centripetal streams flowing into the depressions. It seems certain, however,
that cockpits, like most normal dolines, develop by solution from the surface
'
Figure 12.16.
i -*£**
X
nr^
Aerialview of cockpit karst
W%
3L* ;
.^"^%Sv.
•
y.
;
-^
^^Nk
-•- v t~ »i%
*F*V «T| t
ft 1 %-W
i '* '*
V u*
W.v
•
-N,
Chapter 12 472
downward. Indeed, Monroe (1976) found that circular solutional dolines are
the most common closed depression in Puerto Rico, although collapse dolines
and uvalas are also present. Where cockpit karst ("cone karst" of Monroe
1976) is prevalent, the residual hills are often joined together to form linear
The align-
ridges that in turn are cut by gullies into a sawtooth configuration.
ment of the ridges and the cockpits are apparently so random that joint control
seems to be ruled out, although the gullies may be related to zones of struc-
tural weakness.
Towers are steep-sided hills and can be vertical or even overhanging. The
steep inclination can change where surface erosion piles talus at the slope base
(McDonald 1975). According to Jennings (1971), tower karst differs from
cockpit karst in the steepness of the residual hills and the presence of swampy,
alluvial plains (often similar to poljes) surrounding the towers rather than the
depressed cockpits. Towers can be of dramatic size, sometimes rising several
hundred meters above the surrounding plain (Wilford and Wall 1965). Towers
have also been called pepinos, haystacks, and commonly, mogotes, a term used
in Cuba, Vietnam (Silar 1965), and Puerto Rico (Monroe 1976). In fact,
Monroe (1976) suggests that the term "mogote" is probably more appropriate
for the feature and should be universally accepted.
Mogotes in Puerto Rico rise from blanket sand deposits as hills between
30 m and 50 m high (figs. 12.17, 12.18). Some have solution rock shelters
Figure 12.17. etched into the mogote sides, but caves passing entirely through the hill, a
Large mogote (tower) in Puerto
Rico
H
.
common phenomenon in towers of other regions, are rare. Many of the Puerto
Rican mogotes are asymmetric in their cross-profiles (fig. 12.19), which Monroe
(1976) attributes to case-hardening of the limestones on the windward slope
by repeated solution and reprecipitation of CaC0 3 The downwind slopes, being
.
less resistant, are oversteepened and even overhanging due to slumping of the
less-indurated rocks. Day (1978), however, observes that asymmetry occurs
in only 35 percent of the mogotes and is probably related primarily to erosion
at the tower base, with induration being a secondary factor. The basal erosion
is most likely solutional, but mechanical erosion may be possible (McDonald
1979).
Some controversy exists as to what process causes mogotes or towers to
rise above the alluvial
plain. Theories suggested have been collapse of caves,
river erosion,and differential solution of the karst rocks; probably the features
are polygenetic (Panos and Stelcl 1968). In Puerto Rico, studies show clearly
that the residual hills are composed of the same limestone that underlies the
blanket sands. The one significant factor, however, is that the material holding
up the mogotes has been indurated by solution and reprecipitation of CaC0 3
during alternating episodes of wetting and drying (Monroe 1969, 1976; Miotke
1973). The original limestone surface beneath the alluvial plain was probably
irregular, and so the sand deposits preferentially collected in minor depres-
sions during their initial accumulation. As soon as plants became fixed on the
Figure 12.18.
sands, the limestones were corroded more rapidly, and this, combined with the
Mogotes rising from pineapple
case-hardening of exposed limestone knobs, produced the mogote topography. fields in Puerto Rico.
W *£&.
Chapter 12 474
Direction of wind
Figure 12.19.
Diagram showing characteristics
of an asymmetric mogote in
f= ^^
Puerto Rico (From Monroe 1976)
u » }rnz
Blanket sand
il?5E _* l
i
j
i
,
Limestone indurated by solution and precipitation
2Z Z Soft 1 chalky limestone
<? Empty solution cavities
* Solution cavities partly filled with sinter
' Solution cavities entirely filled with sinter
% Stalactite
Limestone Caves Any survey of karst processes and landforms must include a brief discussion
of limestone caves. Caves are natural underground cavities that include en-
trances, passages, and rooms that can be traversed by a human explorer (W. B.
White 1976). Technically, caves, being underground features, are not part of
karst topography. They may, however, create surface topography by facili-
tating collapse, and they both influence and reflect the mode of karst hy-
drology that exists in any particular region. So caves can justly be considered
as part of the karst system.
Cave Physiography
As defined, caves have entrances, passages, rooms and blockages called ter-
minations. The assemblage of these components in different combinations pro-

duces caves with a variety of shapes and patterns.
Entrances and Terminations Cave entrances can be found in places such as

doline bottoms, hillsides, spring mouths, roadcuts, and quarries. Perhaps the
most spectacular entrances to cave systems are openings into vertical voids
called shafts or chimneys. Shafts (Pohl 1955) are cylindrical in shape and
evidently formed by solution when water moves rapidly down their walls
(Brucker et al. 1972). They are most common in flat-lying rock sequences
such as those in the Appalachian Plateau, where shafts sometimes are more
than 100 m deep and 1 5 m wide. Shafts characteristically drain at the bottom
"
through a narrow opening that is connected to a larger cave system. "Chimney
is a term used by White (1976) to encompass all vertical or nearly vertical
openings that do not have the cylindrical shape of a shaft. Commonly chim-
neys follow steeply dipping bedding planes.
475 Karst —Processes and Land forms
Plan view Cross section

Figure 12.20.
Plan view and cross-section
shapes of cave passages
(Adapted from W. B White 1976)
Canyon Elliptical tube

Hydraulic control
Combined
Structure control
Caves terminate when the passages narrow to the point that a person can
no longer follow the opening. These terminations may be due to collapse of
overlying rocks, narrowing of the voids into permeable but impassable units,
or clay and silt deposits that fill to the ceiling of the cave. Recognize, however,
that all caves are part of an integrated network of flow paths even if termi-
nations exist. Water following caves routes still continues through termina-
tions to a spring outlet.
Passages, Rooms, and Patterns Passages are the main physiographic com-
ponent of caves. They have various shapes, sizes, and patterns, shown in figure
12.20, that are developed at the same or different times by a single source of
water or by a complexly integrated drainage network. In plan view individual
passages are either linear, angulate, or sinuous depending on the structural
and hydrologic controls.
Chapter 12 476
In cross section, the shape of a passage represents an accommodation be-

tween hydraulics and rock properties such as lithology, bedding, and jointing.
Where flow velocities are high and limestones are thick, the passages are usu-
by hydraulics and exist mainly as narrow, vertical slits called
ally controlled
canyons or more circular voids up to 30 m in diameter known as elliptical
tubes. Canyons form in the vadose zone, and tubes usually develop in the
phreatic zone. Passages that are controlled by rock variations are much more
irregular in cross section (fig. 12.20). Realistically, all variations exist in be-
tween the different types, and strict categorization of passage shapes is prob-
ably meaningless. Rooms are usually nothing more than zones of enlarged
passages caused by intersection or enhanced solution.
Almost all passages show a variety of small sculptured markings on their
walls, ceilings, and floors. The features are too numerous to describe here (see
Bretz 1942 for details). Some are formed by solution and others by abrasion.
In addition, many rooms and passages have an abundance of speleothems
(chemically precipitated dripstone deposits) that inspire the popular fasci-
nation with caves. These features (stalactites, stalagmites, etc.) are com-
posed of a CaC0 3 substance called travertine. They form when downward-
permeating water, saturated with respect to calcite, reaches the cave passage.
At this point C0 2 is diffused from the water to the cave atmosphere because
the P(0 in the water is considerably greater (Holland et al. 1964). Some water
may also be lost by evaporation. In either case (loss of water or loss of C0 2 ),
the CaCO, must precipitate as shown in the general reversible reaction (5)
shown earlier in the chapter.
Passages are usually integrated into a three-dimensional conduit system
referred to as cave patterns. In general, the two major types are the branch-
work pattern and the maze pattern. Branchwork patterns, which are by far
the most common, are formed by tubular or canyon passages that join as trib-
utaries in the downflow direction much like a normal surficial stream network.
The pattern develops because, of all the possible flow routes, only a few are
sufficiently open to transport high discharges on low hydraulic gradients. These
evolve into the major passageways. Smaller openings are less efficient in hy-
draulic transmission, and the slow, diffuse nature of the flow causes saturation
in distances that are too short for significant passage enlargement (Palmer
1975). An example of the sequence involved in the development of
excellent
a branchwork pattern is given in Palmer (1981b).
In contrast to the branchwork pattern, the maze system develops as a re-

sult of simultaneous, rather than sequential, enlargement of openings. A maze
pattern consists of many intersecting passages that form closed loops. The pat-
tern has been divided on the basis of geometry into three subtypes (fig. 12.21):
network maze, anastomotic maze, and spongework maze (Palmer 1975). The
specific maze type developed seems to depend on local groundwater recharge
type and structural setting. Maze patterns develop only when the branching
tendency is somehow suppressed. Palmer (1975) suggests that this occurs
mainly in two situations: (1) where aggressive recharge takes place uniformly
in all fractures and (2) where floodwater recharge is so variable in character
that no stable passage configuration is allowed to develop.
Figure 12.21.
Subtypes ot the maze pattern of
cave development.
Anastomotic Spongework
maze maze
The Origin of Limestone Caves

The origin of limestone caves has been steeped in controversy since the fea-
tures were first recognized, and disagreement still rages today. The differences
of opinion probably stem from several factors. (1) Historically, theories about
cave formation have tried to fit all caves into genetic models with little regard
for geologic differences. For example, there is no compelling reason that caves
in highly deformed rocks of the Alpine region should develop in precisely the
same way as those in the plateau areas of southern Indiana. (2) Most cave
models are based on hydrological schemes, yet until recently there has been
a remarkable lack of hydrologic data in the arguments. In fact, most theories
were founded on physiographic evidence. (3) Very few caves have been ex-
amined in great enough detail to substantiate a local origin, let alone an all-
encompassing genetic model.
All classical theories of cave development are concerned primarily with
the position of the solvent water relative to a water table. Three main ideas
have evolved over the years, suggesting that caves form ( 1 ) above the water
table by the corrosive action of vadose water, (2) beneath the water table by
deep circulation of phreatic water, or (3) at the water table or in the shallow
phreatic zone, often associated with fluctuations of the water table itself.
The concept that caves develop above the water table by vadose water
appeared early in this century and has been complicated by many researchers
adopting slightly different variations of the theme ( Martel 1921; Malott 1921,
1938; Piper 1932; Gardner 1935). In general, most workers recognized the
and many suggested that rivers flowing with
solution effect of vadose flow,
some velocity under hydrostatic head are able to enlarge caves by abrasion.
Some saw importance in the water table level since free-flowing water is es-
sentially on the top of the water table, but others (Malott 1921; Addington
1927) denied its importance.
In 1930 W. M. Davis published his classic "two-cycle theory" of cave de-
velopment, which depends on deeply circulating phreatic water. Davis argued
that most caves are now in a phase of active deposition rather than solution,
as evinced by the abundance of dripstone. Thus, cave erosion must have oc-
curred at some earlier time when the water table was higher than the present.
According to Davis, solution of the caves took place along curving flow lines
in the phreatic zone (fig. 12.22a) that descended deep below the surface and
emerged under the major river. Subsequently, the region was rejuvenated by
Chapter 12 478
Figure 12.22.
Models of groundwater flow in the
development of caves
Swinnerton (1932)
(B)
Water table Stream
10 km -
Corrosion
(C) Thrailkill (1968)
uplift, rivers entrenched into the surface, and the water table was lowered to
join with the new level of the rivers. This, of course, exposed the already formed
caves and allowed dripstone deposition to begin.
The major model, and its many variations, place the main zone of
third
cave enlargement in the reach where the water table fluctuates with the sea-
sons. Swinnerton (1932) accepted earlier ideas that stressed the importance
of the water table. He believed that if water moves along all possible paths
beneath the water table, the zone just below the table would be the most un-
dersaturated and would have the greatest discharge (fig. 12.22b). Thus, cave
formation would proceed most efficiently in the shallow phreatic zone and
should be related to the surface stream level. Other studies seem to support
this model (Sweeting 1950; Davies 1960; White 1960: Wolfe 1964) but the
most convincing evidence was presented by Thrailkill (1968).
Thrailkill provided a sound theoretical base for the shallow phreatic model
by pointing out that where no water crosses the water table, the flow will be
nearly horizontal and water just beneath the water table will follow very shallow
flow routes, especially if the recharge occurs from a number of discrete point
sources (fig. 12.22c). The water at the water table becomes undersaturated
when it mixes with vadose water (mixing corrosion), when it becomes cooled
(dissolves more C0 2 ), or when water is introduced by backflooding of nearby
rivers. The combination of lateral flow and undersaturation dissolves more
limestone during floods, and therefore, the position of cave formation is most
likely to be the average level of the flood water table, which he considers to
be the upper limit of the shallow phreatic zone.
There seems to be a general sympathy for the shallow-phreatic mode of

cave origin (Moore 1960; LeGrand 1983). However, any attempt to restrict
cave origin to a specific groundwater zone may be too limiting. Caves can
probably form anywhere sufficient groundwater flow and suitable lithologic
characteristics exist (White 1969, 1977). Given sufficient groundwater, whether
a cave develops in a shallow or deep situation may depend on local controlling
factors (for example, Ford and Ewers 1978). The complete integration of caves
may also be time dependent; i.e., the location of major solution may shift with
time. In addition, it now seems certain, as discussed earlier, that the transport
distance of groundwater flow and the percentage of saturation also control
where caves may and may not develop. Palmer (1981a) suggests that a great
disparity exists in solution rates of karst rocks. This eventually leads to the
creation of only a few major passages, and the result is a dendritic (branch-
work) pattern. However, beyond a certain discharge there is a limit to the
solution rate. Where many different fractures are exposed to flow at or greater
than this limiting discharge value, a maze pattern will develop. Therefore, the
type of cave pattern and the position of its development may be a function of
the mode of groundwater discharge and the local structural condition.
Karst develops by extensive solution of limestones and other soluble rock types. Summary
Limestones that are most susceptible to karstification are crystalline, high in
calcite content, and intensely fractured. The limestones should also have no
impermeable units in a stratigraphic sequence that stands well above base
level. The karst processes function best in humid-temperate and humid-trop-
ical climates where abundant water is available and biogenic CO z is freely
added to the water from a thick vegetal cover.

Karst areas are characterized by a distinct hydrology in which surface
flow is partially or totally disrupted when water is diverted into the under-
ground system. Groundwater movement in karst aquifers differs from normal
situations in that the flow is often controlled by the orientation of fractures
and interconnected solution cavities.
The most common karst landforms are enclosed depressions called dolines
(sinks or sinkholes in the U.S.). Other distinct forms are larger depressions
and a variety of surface valleys that are unique to karst regions. Tropical karst
differs from temperate karst in that hills rather than depressions dominate the
topography. In most regions, karst landforms and hydrologic parameters seem
to possess regularmorphometric relationships that probably reveal some type
of dynamic equilibrium between process and form. Limestone caves are sub-
surface manifestations of karstification. The mechanics of cave formation are
not completely understood. Most interpretive disagreement has centered on
where the solution process is most efficient in relation to the water table.
The cave pattern developed depends primarily on the mode of ground-
water recharge and the local structures in the rocks.
Chapter 12 480
Jennings, J. N. 1 97 1 Karst. Cambridge, Mass.: M.I.T. Press.

.
Kemmerly, P. R. 1982. Spatial analysis of a karst depression population:

Clues to genesis. Geol. Soc. America Bull. 93:1078-86.
Monroe, W. H. 1976. The karst landforms of Puerto Rico. U.S. Geol.
Survey Prof. Paper 899.
Palmer, A. N. 1984. Recent trends in karst geomorphology. Jour. Geol.
Educ. 32:247-53.
Sweeting, M. M. 1973. Karst landforms. New York: Columbia Univ. Press.
Thrailkill, J. 1968. Chemical and hydrologic factors in the excavation of
limestone caves. Geol. Soc. America Bull. 79:19-45.
Williams, P. W. 1983. The role of the subcutaneous zone in karst hydrology.
J. Hydrol. 61:45-67.
Coastal Zones — Processes
and Landforms
13
I. Introduction IV. Shoreline Configurations and
II. Coastal Processes Landforms
A. Waves A. Beach Cusps
1 . Wave Generation B. Large-Scale Rhythmic
2. Waves and Shoaling Topography and Capes
3. Tsunamis and Seiches V. Erosional Landforms and
B. Tides and Currents Rates
1. Tides VI. Depositional Shorelines
2. Nearshore Currents A. Spits and Baymouth Bars
a. Rip Currents and B. Barrier Islands
Cell Circulation 1. Distribution and
b. Longshore Currents Characteristics
from Oblique Waves 2. The Geomorphic
C. Coastal Storms Processes
III. Beaches VII. Summary
A. The Beach Profile VIII. Suggested Readings
B. Beach Morphodynamics
481
Chapter 13 482
Introduction In the preceding chapters we have examined a variety of processes that mold
the exposed surface of the Earth into diagnostic landforms. However, 70 per-
cent of our planet is covered by ocean water, and the sea possesses immense
energy. The application of this energy on the adjacent land can produce rapid
and enormous changes in the nearshore physical environment. Therefore, the
and depositional processes that function at the interface between the
erosional
ocean and the land represent another major realm of geomorphology.
Coastal zones respond to forces like any other geomorphic system. Viewed
on a short-term basis, parts of the coast might be considered as quasi-equi-
librium forms. In that sense, repeated movement of sediment and water con-
structs a beach profile that reflects some balance between the average daily
or seasonal wave forces and the resistance of the landmass to wave action.
Considered over a longer time span (graded time), however, beaches or entire
coastlines may be imperceptibly changing toward a larger equilibrium form.
On the even larger geologic time scale, marine transgressions and regressions
may represent dramatic alterations of the position and character of the coast.
Our attention here will focus mainly on the coastal zone during steady and
graded time spans.
The study of beaches and coasts is not strictly academic. Most of the largest
U.S. cities are near the ocean, and three-fourths of the people in the country
live in coastal states. The concentration of people in coastal regions and the
pressures involved in using this land for recreation and development have placed
a genuine strain on the system. Whether we can utilize the environment com-
pletelyand still prevent damaging changes in its character depends on how
well we understand the functional processes. Engineers, geographers, ocean-
ographers, geologists, and other scientists have made significant contributions
to our present understanding of the coastal environment. As usual, however,
interdisciplinary communication has not been great because each discipline is
engrossed in specific problems. Nonetheless, if demands on the coastal system

continue to increase, cooperative efforts of all coastal workers will be needed
to avert environmental disaster.
For those interested in coasts or beaches, there exist a monumental wealth
of data and many The following references treat
exceptionally good syntheses.
the topic with a number of different approaches and at many levels of so-
phistication: D. W. Johnson 1919; Guilcher 1958; Steers 1962; Shepard 1963;
Wiegel 1964; Bascom 1964; Ippen 1966; Zenkovich 1967; Manley and Manley
1968; Bird 1969; Muir Wood 1969; Shepard and Wanless 1971; King 1972;
Davies 1973; CERC 1973; Komar 1976, 1983b; Davis 1978. Much of the ma-
terial presented here has been drawn from these excellent sources.
A coast is a relatively large physiographic zone that extends for hundreds

of kilometers along a shoreline and often several kilometers inland from the
shore. Like all large physiographic entities, coasts have not escaped the ir-
repressible urge of scientists to classify, and many attempts have been made
483 Coastal Zones — Processes and Landforms
to categorize their features (see King 1972). Coasts are particularly suscep-
tible to classification because their regional properties are clearly documented
on available maps and aerial photographs. In some cases, no extensive field
study is needed to systematize coastal properties into a viable classification
scheme.
Some of the commonly used classifications of coasts combine, in different
ways, a few basic ingredients that serve as the fundamental criteria. Most
important of these are ( 1 ) form of the land-sea contact, i.e., the configuration
of the shoreline; (2) stability or relative movement of sea level; and (3) influence
of marine processes. Some of the classifications are purely descriptive and may
have little application in dynamic geomorphology. Others are genetic and so
are allied closely with the processes involved in developing the diagnostic coastal
properties. For example, Inman and Nordstrom (1971) classify the mor-
phology of coasts on the basis of modern plate tectonic theory. On the other
hand, as Russell (1967a) suggests, perhaps we have not yet collected enough
precise data concerning coastal properties to warrant any classification at this
time. In addition, all coasts have a past as well as a present, and so time also
becomes an important consideration in coastal classification (Bloom 1965).
That is, coasts may reflect evolution and contain relict parts that are not in
equilibrium with modern, or even Holocene, processes. For instance, marine
terraces standing well above the ocean are not related to modern wave attack
but do indicate a coast that has undergone movement relative to sea level.
Because of all these considerations, no specific classification will be ad-
vocated here, and no attempt will be made to analyze entire coastal regions.
Instead, we will examine only parts of the entire coast that are being actively
affected by modern processes. In that sense we will talk about "erosional" or
"depositional" coasts, but these terms are applied only on a local basis and
have no regional connotation.
The processes that initiate change in the coastal zone are extremely difficult Coastal Processes
to study because they are driven by interrelated forces of high energy, each
of which may produce a different response in the same coastal environment.
Thus, different processes are not easily studied without permanent installa-
tions designed to provide precise measurements over various time intervals.
An example of such a research center is operated and maintained at Duck,

North Carolina, by the U.S. Army Corps of Engineers. This installation, known
as the Coastal Engineering Research Center-Field Research Facility (CERC-
FRF) collects data on various oceanographic and meteorological factors, in
addition to conducting regular and event-initiated bathymetric surveys of the
nearshore environment (table 13.1). Measurements are made from a 561 m
steel pier oriented perpendicular to the shoreline and a mobile sam-
(fig. 13.1)
pling machine called the CRAB (Coastal Research Amphibious Buggy). The
CRAB (fig. 13.2) can be driven as far as a kilometer offshore into water depths
up to 9 m.
Chapter 13 484
Table 13.1 Data type and frequency of measurements at the Coastal

Engineering Research Center-Field Research Facility, Duck, North
Carolina.
Data Type and Collection Technique Frequency of Measurement
Meteorological
Temperature Continuous
Rainfall Continuous
Barometric pressure Continuous
Wind velocity and direction Continuous
Wave Data
Baylor staff gages on pier 20 mm. every 6 hr
Wave rider buoy 2500 m off pier 20 min. every 6 hr
Radar derived wave height, period, angle On call; at least 1/da
Currents
Longshore surface current direction determined Daily
by dye packets at pier end, nearshore and
beach
Oceanographic
Surface water temperature Daily
Water density
Water visibility
Water Levels
Tide levels at pier end gage Every 6 minutes
Nearshore Bathymetry
Pier lead lines Daily
CRAB surveys
Four lines parallel to pier; Biweekly and after storms
500-600 m N and S of pier; up
to 1 km offshore
Figure 13.1. Complete surveys along 14 lines within Monthly

Sampling pier at the Coastal ± 600 m of pier
Engineering Research Center,
Duck, N.C.
(A) (B)
Figure 13.2.
(A) CoastalResearch Amphibious
Buggy (CRAB) used at CERC for
offshore sampling (B) CRAB
taking bottom samples near the
CERC pier at Duck, N.C.
Chapter 13 486
Wavelength
Figure 13.3.
Fundamental parameters of —r
waves.
^^ ^^ ^^^ ^^ Wave height
(H)
The description of CERC-FRF is not meant to suggest that coastal re-

search in the absence of such heroic efforts is inconsequential. On the contrary,
many excellent studies have been accomplished without access to such re-
search facilities. Realistically, however, continuous data collection at perma-
nent stations may be needed to integrate information about diverse phenomena
into reliable syntheses.
Waves
The paramount driving forces in shoreline processes are waves, which expend
the energy they obtained in the ocean against the margins of the land. A thor-
ough analysis of wave theory is steeped in complex physics and mathematics.
Here we can introduce only a simplified version of the basic concepts needed
to understand the geomorphic role of waves, recognizing that such a discussion
is a broad generalization of a very complicated phenomenon.
Wave Generation Most waves that do geomorphic work are generated by

strong winds blowing across large portions of the open ocean. Precisely how
energy is transferred from the wind into ocean waves is not completely under-
stood (discussed in Komar 1976). Nonetheless, empirical studies indicate that
wave properties reflecting this energy exchange (fig. 13.3) depend primarily
on the wind velocity, the wind duration, and the fetch (the distance over which
the wind blows). The fetch is particularly important in determining the height
of waves and their period, a parameter that is merely the time interval between
two successive wave crests passing a fixed point. Extremely high waves with
long periods can be generated only when all three controlling factors are at a
maximum.
In the area where they are generated, waves are irregular, and individual
crests are discernible for only short distances before they disappear in a maze
of interfering waves. This chaotic state occurs mainly because small and large
waves generated during the same storm are out of phase. Waves passing through
the system tend to interfere with one another, adding to the height of some
waves and subtracting from others. When waves are enhanced, the heights
developed can be truly awesome. Storm waves are commonly more than 20 m
high under severe winds. The greatest documented wave height, 34 m ( 1 1 2 ft),
was measured in a February 1933 storm generated in the open water in the
deep part of the South Pacific.
As waves move away from their source area, they begin to separate from
one another according to their various periods. This process, called wave dis-
persion, causes regularly spaced successions of waves with rounded crests to
Wave direction
Figure 13.4.
The orbital motion of water
particles in a wave of oscillation.
Zone of wave
Figure 13.5.
generation
Dispersion of waves from an area
Ocean swell of wave generation Waves having
different periods separate from
one another to create ocean
swell.
Dispersion
migrate from the source zone. Emerging waves typically have a low ratio of
wave height to wave length (a parameter known as the wave steepness) and
appear as the long, low waves commonly referred to as swell. In swell, water
particlesassume the circular orbital paths that characterize deep-water waves
of oscillation (fig. 13.4). Although individual water particles in oscillatory
waves have little forward motion, the waves themselves advance with a typical
trochoid form. That is, the wave form is moving forward, not the ocean water.
The dispersive process, shown diagrammatically in figure 13.5, works be-
cause wave length and velocity are both a function of the period such that
and
v-f, Z7T
where g and w are the well-known constants, L is the wave length in meters,
V is velocity in m/sec, and T is the period in seconds. Because of these rela-
tionships, long-period waves with greater length and velocity will separate from
short-period waves. For example, a wave with a period of 6 seconds will have
a length of 56 m (184 ft) and a velocity of 9 m/sec (30.7 ft/sec). Waves gen-
erated in the same storm that have a period of 14 seconds have a length of
303 m(1000 ft) and a velocity of about 22 m/sec (71.5 ft/sec). It is apparent
that waves disperse from the generation zone simply because longer waves
outrun the shorter ones during any given interval of time.
Chapter 13 488
Breaker
Figure 13.6.
zone
The transformation of oscillation Swel Waves of
waves into waves of translation as translation
swell approaches the nearshore
environment
Ocean floor
Waves with identical periods travel away from their source as distinct
groups. It is important to note that storm-generated waves are not the same
as waves produced by a point-source impulse such as a pebble tossed into a
pond or a tsunami created by rock displacement beneath the ocean floor. In
such cases, waves radiate in all directions from the point source. In contrast,
ocean swell follows a pathway that encompasses a substantial area of the ocean
surface but also has finite lateral boundaries; the direction of the well-defined
corridor is determined by the direction of the generating wind. Although the
wave path does spread somewhat, it is quite possible for swell to strike a few
score kilometers along a coast while leaving nearby reaches virtually unaf-
fected by the storm.
Studies show that swell can travel great distances across the open ocean
without losing much of the original energy (Snodgrass et al. 1966). Most en-
ergy loss, especially in the shorter period waves, occurs near the generation
zone. Once the swell condition is attained, the long-period waves experience
little additional dissipation of energy and may traverse an entire ocean basin.
Waves and Shoaling As swell approaches a landmass, the waves begin to

"feel" the effect of the ocean bottom. When the water depth is approximately
half the wave length, the oscillatory waves begin a transformation into steplike
forms called waves of translation (fig. 13.6). Translation waves develop when
the orbital path of water particles is intercepted by the ocean floor. The orbits
begin to flatten noticeably and their axes of rotation rise to higher levels. Even-
tually the orbital path is destroyed, producing the different wave type. Waves
of translation are different from waves of oscillation in that the water particles
have a distinct forward motion without the corresponding backward move-
ment that characterizes oscillation. Once in the form of translation, the ve-
locity and length are determined as
v= Vg/T
and
L=T yfgh,
where h is the depth of the water.

The change in wave types is accompanied by the phenomenon of breaking
waves. Deep-water oscillatory waves remain stable only if the wave steepness
(H/L) is lower than 1/7 (.14). As waves approach the shore, water depth is
progressively decreased (a phenomenon called shoaling), and waves begin to
"feel bottom." Their heights increase as the rounded crests become more
489 Coastal Zones —Processes and Landforms
Figure 13.7.
Types of breaking waves.
Spilling
Plunging
Collapsing
Surging
peaked, and their velocities and lengths decrease as the wave crests bunch up;
only the wave periods remain constant (fig. 13.6). The combined effect of an
increasing height (H) and decreasing length (L) causes oversteepening of the
wave; the critical H/L value is attained and the wave breaks. The distance
over which breaking occurs depends primarily on the bottom slope, i.e., how
rapidly the water depth decreases toward the shoreline.
Three common types of breakers, called spilling, plunging, and surging,
have been observed and are depicted in figure 13.7 (Wiegel 1964). In spilling
breakers, the top of the wave crest becomes unstable and flows down the wave
front as an irregular foam. In the plunging type, the wave crest curls over the
front face and falls with a splashing action into the base of the wave. In surging,
the wave crest remains essentially unbroken, but the base of the wave front
advances up the beach. A fourth type of breaker, called collapsing, has been
identified by Galvin (1968); its characteristics are intermediate between those
of the plunging and the surging types (fig. 13.7). Actually, breaking waves
probably occur in a complete spectrum of types that depend on the bottom
slope, H/L, and the period. In addition, breaker style at any given locality
varies from periods of storms to periods of relative quiet. This indicates that
the type of breaker may be subject to change irrespective of bottom topog-
raphy.
Chapter 13 490
The significance of breaker style rests in the fact that most types tend to
push sediment toward the shore and therefore are relatively benign with re-
gard to beach erosion. Plunging breakers, however, are dominantly erosive and
will produce large (often destructive) changes in the beach environment.
Breakers mark the oceanward limit of a zone, called the surf, in which
the original energy given to the waves receives its final transformation. High
velocitiesand substantial impacts occur under breakers, and thus the creation
of translation waves provides the water with kinetic energy that is capable of
doing geomorphic work. Once formed, the waves of translation (and the water
they contain) move forward to their inevitable collision with the landmass.
The surf ends when the wave form is lost as it impinges on the beach face.
From there, water carried by its own momentum continues to slide up the
beach as swash until, at its highest encroachment, the force gathered in the
open ocean is finally and totally dissipated. The swash zone is alternately cov-
ered as the water rushes up the beach face (swash) and exposed as the water
moves back down the beach (backwash) under the influence of gravity.
The position of the surf zone may change frequently because breaking
waves not only reflect the bottom topography but also use their energy to re-
align it. That is, swell approaching a shoreline will break according to the
existing shoaling configuration, but at the same time, the breakers may shift
the bottom sediment so as to produce a new sea-floor topography for the next
train of storm-generated waves. In addition, the surf zone changes in response
to vertical displacement of the ocean level due to tides, storms, and so forth.
In some cases two sets of swell generated in different areas arrive at a
shore simultaneously. This produces a systematic variation in the heights of

waves striking the shore, a phenomenon known as surf beat (Munk 1949;
Tucker 1950). In surf beat, successive waves gradually increase in height until
they reach a maximum, then systematically decrease in height to a minimum
value. Thus a pronounced periodicity develops whereby one large wave will
appear with predictable regularity. In many cases the dynamics are such that
every sixth to eighth wave will be at the maximum height, but the precise beat
depends on the wave periods of the two swell systems and their resulting har-
monics. The variation of breaker heights produced by surf beat affects beach
processes because it changes the prevailing water level and notably alters the
velocity of nearshore currents.
In addition to these changes in wave mechanics, ocean swell entering
shallow water also changes direction. Because deflection of the wave crests is
a function of the water depth, the waves adjust to the contours of the bottom
topography and so bend according to the configuration of the shoreline. The
process, known as refraction, works because some sections of a wave crest
approaching a coast obliquely are moving in shallow water and at lower ve-
locities than other portions of the same wave, which are traveling in deeper
water (Silvester 1 966). As shown in figure 1 3.8, the waves converge on a land-
mass that juts into the ocean. When
convergence occurs, the energy per unit
length of wave is increased, causing higher waves and increased energy to
impinge on the headland region. Waves can also diverge over embayments or
submarine canyons, resulting in lower waves and a concomitant spreading of
the energy.
Figure 13.8.
Refraction of waves along an
irregular coastlineWave energy
converges on landmasses that
project oceanward and diverges
in recessed areas.
Tsunamis and Seiches Like the waves produced in the open ocean by winds,
other waves generated in different ways may have geomorphic significance.
Tsunamis are waves formed by sudden impulses beneath the ocean that cause
trains of waves to radiate in all directions from the point source. Tsunamis
are usually initiated by earthquakes of more than 6.5 magnitude on the Richter
scale with foci located less than 50 km beneath the ocean floor (Van Dorn
1966). Submarine landslides, volcanic eruptions, and slumping have also been
cited as causes. In any case, sudden movement displaces the overlying water
column, causing it to oscillate up and down as the water tries to reestablish
mean sea level.
The waves leaving the source zone of a tsunami have distinct character-
istics. They are extremely long (as much as 240 km), their period may be as
much as 1000 sec, and in open water they may be only a meter high. Tsunamis
obey the same laws that control shallow-water waves, and therefore their ve-
locity is proportional to the water depth. In 10,000 feet (3000 m) of water,
the wave velocity will be
= V32- 10,000
= 566 ft/sec or 386 mph (approximately 618 km/hr).
In the open ocean, such waves pass quickly and with little notice. As waves
approach the shore, however, they seem to trigger a harmonic oscillation that
does not follow the bottom configuration. In a typical event, a moderate rise
or recession of sea levelis followed by three to five major wave fronts that are
tens of meters high and capable of great destruction. For example, N. H. Heck
(as quoted by Bascom 1964) describes a remarkable case in which a U.S.
warship anchored in a Peruvian port in 1 868 was picked up by a tsunami wave,
carried over the top of the small port city (Iquique), and finally dropped 400
meters inland. After the major waves pass, the nearshore system gradually
returns to normal (Van Dorn 1965, 1966). Geomorphic effects of tsunamis are
dramatic but probably short-lived; tsunamis occur rarely, and normal waves
rework the coast according to more prevalent controls.
Chapter 13 492
A seiche is another wave type that is not directly related to a prevailing

open ocean wind. Seiches are free oscillations of water in enclosed or semien-
closed basins. Although originally observed in lakes, the seiche phenomenon
occurs also in harbors, where it is often called surging, and along open coasts
with a broad, shallow continental shelf. A seiche is recognized as a repeated
rise and fall of the water level; the oscillatory motion begins when some force
displaces the water from its equilibrium position. The driving impulse may be
heavy rainfall, flood discharge from nearby rivers, long-period waves such as
tsunamis or surf beat, or rapid pressure fluctuations associated with storms
(B. W. Wilson 1966). In any case, when the initiating force passes, the oscil-
Spring tide lations gradually decrease until the equilibrium level is once again attained.
Tides and Currents

Although waves are the dominant force influencing the coastal environment,
they are not the only significant water motion. Tides and currents each con-
stitute a type of movement that can modify coastal properties. Although our
treatment of these forces will be necessarily brief, a complete understanding
of nearshore geomorphic mechanics requires detailed consideration of these
factors.
M
Tides As any dedicated beachcomber knows, tides usually occur as a twice-
daily rise and fall of sea level. Away from coastal areas, this movement is of
Neap tide
little concern to most laypeople, but the tidal effect is of consequence to the
coastal geomorphologist for several reasons. First, because of the continuous
Figure 13.9.
change in water level, the position of wave attack migrates through a notable
Tidal distribution at various stages
ot the lunar cycle (S= sun; E = vertical range and a corresponding lateral shift, increasing the size of the beach
Earth; M = moon). During spring and complicating our understanding of beach processes. Second, tides initiate
tide the moon and sun are aligned
currents that flow into and out of constricted reaches of the shoreline such as
and cause larger tides During
neap tide the moon and sun are bays or lagoons. Many times this ebb and flow can keep drifting beach debris
not aligned, and tides are lower from closing the entrance to the embayment. and in fact, some tidal currents
than average (Not to scale)
are capable of eroding coastal rocks. This is especially true in narrow bays
such as the Bay of Fundy in eastern Canada, where the inland constriction
produces a maximum range of 15.6 m, the highest value on Earth. In some
constricted estuaries tides move inland with a pronounced wave front called
a tidal bore (see Lynch 1982). Often the bores actually break, providing ob-
servers with a spectacular show. For example, the tidal bore on the Amazon
River looks like an 8 m high waterfall advancing upstream for 480 km at a
rate of 1 2 knots.
The tides, of course, are driven by the gravitational effect exerted on the
Earth by the sun and the nonmathematical discussion see Bascom
moon (for a
1964; for an understandable mathematical discussion. Komar 1976). Here we
need only recognize that in most coastal regions the lunar influence results in
two high tides and two low tides daily. The gravitational attraction of the sun
complements or detracts from that of the moon. As figure 13.9 illustrates,
every two weeks the moon and sun are aligned, causing a higher tide than
normal, called the spring tide. Midway between spring tides, the moon and
sun reach positions that are 90° apart. The solar pull detracts from the lunar
effect, and the tide is lower than normal, the so-called neap tide. Spring tides
are about 20 percent greater than the average tidal range and neap tides about
20 percent less.
Not places on Earth experience the tidal motion that pure astronom-
all
icaland gravitational theory predicts. Some areas, for example, have only one
high and low tide, which occur at the "wrong" times. The tide at any locality
also varies in magnitude because of other factors, such as perturbation of the
lunar orbit, tilting of the Earth's axis, and ocean bottom topography. Davis
(1964) has classified tides according to their tidal range as microtidal (0-
6 ft), mesotidal (6-12 ft), and macrotidal (> 12 ft).
Normal tides are not usually destructive. However, some unique tidal
events can be devastating, especially if they coincide with storms that produce
strong onshore winds. For example, because the orbital path of the moon is
elliptical, the moon periodically reaches a position where it is closer to the

Earth than at any other time. This point is known as perigee. Occasionally
the location of perigee is perfectly aligned with the celestial orientation during
spring tide, producing what is known as a perigean spring tide. These unique
tides not only raise the normal tide levels but, more important, increase the
rate at which the tide rises (Wood 1978). In March 1962 the chance combi-
nation of a perigean spring tide and a large offshore storm generated enormous
flooding and erosion along the Atlantic coast from the Carolinas to Cape Cod
(fig. The event resulted in a loss of 40 lives and $500
13.10). million in prop-
erty damage; in some areas destruction was almost total.
Nearshore Currents Another type of water motion complicates the near-

shore system and our understanding of the mechanics in the surf and swash
zones. Other than wave action itself, two wave-induced currents control water
movement in the beach zone: (1) a cell circulation consisting of rip currents
and their associated longshore currents; and (2) longshore currents that usu-
ally are generated by waves striking at an angle to the prevailing direction of
the shoreline but may also be caused by tides or storms.
Rip Currents and Cell Circulation Rip currents are narrow zones of strong
flow that move seaward through the surf (fig. 13.1 1). They return to the off-
shore zone water that was moved toward the beach by waves of translation.
The rips are caused by small currents that move on the beach face parallel to
the shoreline; these originate about halfway between two adjacent rips. The
velocity of the feeder current reaches a maximum at the entrance to the rip
zone. Velocity in the rip current itself can be significant, with known speeds
reaching approximately 2 m/sec (Sonu 1972).
The development of the cell formed by a longshore current and a rip is
related to variations in the rise of mean water level above the level normally
attained under still water, a phenomenon called wave set-up. Circulation be-
gins because the height of incoming waves varies in a longshore direction, and
Figure 13.10.
Overwash flooding of the New
Jersey coast during the March
1962 offshore storm combined
with the pengean spring tide.
Figure 13.11.
Nearshore circulation pattern at
Scripps Scripps Beach, La Jolla, Cat,
Rip currents — •*-- -5^" pier showing rip currents and
longshore currents that feed the
rips.
Waves /
-No current
\ ^ / Beach
\ 500 1000
3
Feet
-No current
higher waves create greater wave set-up. Longshore currents therefore flow
from zones of the highest breakers and return oceanward at the position of
the lowest breakers. Waves of similar height also initiate cells, but the process
probably depends on the type of breakers involved (Sonu 1972).
Shepard and his coworkers (1941) realized long ago that the position of
rip currents is controlled by the bottom topography in the surf zone. We can
expect, therefore, that areas of wave convergence, where waves characteris-
tically are higher, are the starting points of the longshore component in the
cell. Rips occur away from these zones where breaker heights are smaller.
Although this analysis has been shown to be correct, rip currents also exist on
long, straight beaches with smooth bottoms. Therefore, bottom topography is
not the only cause of cell circulation. For example, it is now known that ocean
swell by itself produces secondary waves in the surf zone (Bowen and Inman
1969; Huntley and Bowen 1973). These waves, called edge waves, have crests
normal to the shoreline and wave lengths parallel to the shore; they oscillate
with an up-and-down motion. Some authors have suggested that edge waves
may be the triggering devices of rip currents (see Komar 1976), but this sug-
gestion by no means universally accepted (Hino 1975). Regardless of their
is
precise origin, rip currents tend to perpetuate themselves because they can
modify the bottom topography so that it enhances the circulation pattern.
Longshore Currents from Oblique Waves Longshore currents are also gen-
erated by waves that strike the beach obliquely. These currents are extremely
important in beach mechanics because they tend to move sediment parallel to
the shoreline for considerable distances and thereby present coastal engineers
Chapter 13 496
with innumerable problems. A number of attempts have been made to derive

equations that relate wave properties to the velocity of the longshore current
(for reviews, see Galvin 1967; Komar 1976). In general, most equations fail
in a predictive sense because they either (l) do not distinguish the longshore
current associated with cell circulation from that produced by obliquely striking
waves, or (2) employ invalid criteria as the fundamental theoretical base.
One approach, based on momentum analysis, has produced good results
and probably represents the best hope for a usable predictive model. Mo-
mentum, unlike energy, is preserved as waves break and is separated into two
components, one directed toward the shoreline and one directed parallel to it.
Thus, the flux in momentum directed along the shoreline should be propor-
tionately related to the velocity of the longshore current. This concept has
been developed thoroughly by many authors but most completely by Bowen
(1969), Longuet-Higgins (1970), Komar and Inman (1970), and Komar
(1975). Komar (1976) suggests that the best estimate of longshore currents
at the mid-surf position follows the equation
V =
t 2.7w„, sin a cos a,
where u m is the maximum orbital velocity at the breaker zone, V, is mean

velocity in cm/sec, and a is the breaker angle, i.e., the angle of incidence with
a line parallel to the shoreline (details of the derivation have been omitted).
It should be pointed out that tides and winds blowing in the longshore
direction may complicate the system. At low tide, for example, water can be
trapped in troughs that parallel the shoreline. Continued spilling of waves into
the troughs drives a circulation pattern in which water moves alongshore con-
it reaches a rip channel cut through an offshore
fined within the troughs until
bar. There the water turns seaward as part of the rip current. In these cases
the longshore current may have a velocity that obeys a mass continuity law
rather than a momentum law (see Inman and Bagnold 1963; Bruun 1963;
Galvin and Eagleson 1965).
Coastal Storms
In addition to the driving factors discussed above, the dynamics of coasts are
directly affected by storms that impinge on the shoreline. Coastal storms occur
as two types, tropical and extratropical. Both types are cyclones, meaning that
they involve a circular wind pattern, which in the Northern Hemisphere moves
in a counterclockwise direction. Extratropical storms derive their energy me-
chanically in a process associated with the interaction of air motion between
zones of high and low pressure. In contrast, tropical storms are fueled by latent
heat from the evaporation of water. They evolve as deep, atmospheric low-
pressure systems that originate and gain intensity over warm-water marine
areas. Tropical cyclones are given a variety of names depending on their geo-
graphical location and/or their wind velocities. In the Atlantic and Gulf coasts
of the United States, the term "hurricane" (fig. 1 3. 1 2) is employed when wind
velocity exceeds 74 mph (119 km/hr).
497 Coastal Zones— Processes and Land forms
Figure 13.12.
Hurricane Diana, September 1984
Chapter 13 498
Table 13.2 Characteristics of tropical and extratropical storms
Tropical Storms Extratropical Storms
Wind Speed Great, > 75 mph Less, usually < 50 mph

Duration 3 Short, few hours Longer, many hours to days
Size Small, 50-80 km Large. 100's km
Shape Circular Often elongate
Surge Large, > 15 ft Small, < 5 ft
Barometric Low central; greater pressure Higher central; lesser pressure

pressure differential differential
Fetch Small, 10's km Large. 100s km

Occurrence June-Sept Oct -Apr
a Duration refers to time of effect on coast
In the United States, hurricanes are predominant south of Cape Hatteras,

North Carolina, whereas extratropical storms are much more common north
of Cape Hatteras. This frequency distribution is important because the char-
A major difference
acteristics of the storms are quite different (see table 13.2).
between the storm types is the magnitude of storm surges that they generate.
A storm surge is an elevation of normal water level in response to a passing
storm. Storm surges are produced when strong onshore winds push and hold
water against the coast, thereby "setting up" the mean water level. This oc-
curs because any low pressure promotes a compensating upward bulge of the
ocean level (a 1-inch drop in pressure causes about a 13-inch rise in water
level). In hurricane-induced surges, the extremely low pressure in the storm
center compared to extratropical storms significantly enhances the surge mag-
nitude.
Extreme surges combined with high surface waves allow water to pene-
trate inland beyond the beach (a process known as overwash) and cover areas
normally immune from wave attack (see fig. 13.10). Historically, hurricane-
related surges are much more damaging than those associated with extra-
tropical storms. This does not imply that major flooding and destruction are
impossible in extratropical storm surges. On the contrary, the March 1962
storm discussed earlier is excellent proof that extratropical storms can be sig-
nificant events. Usually, however, they require involvement of some ancillary-
factor to reach their full destructive potential.
Beaches The difficulties of associating entire coasts with formative processes tend to
direct our attention to smallercomponents of the coast where processes are
more easily understood. The feature most amenable to process analysis is the
beach, which is most simply defined as the relatively narrow portion of a coast
that is directly affected by wave action. It usually terminates inland at a sea
cliff, a dune field, or at the boundaries of permanent vegetation. Oceanward.
under the constraints of our definition, part of the beach is continuously sub-
merged because it lies beneath low-tide sea level. It is still part of the beach.
Nearshore- Foreshore* -Backshore-

Figure 13.13.
Typical beach profile.
MHW
MLW ^^ |
\
\
\
Berm
k Ridge J
Beach face
Longshore bar I
Runnel
Breakers and swash- •+• Storm overwash
however, because the bottom is subjected to wave action. In that sense the
term "beach" is synonymous with the littoral zone, an expression used com-
monly in geological work.
The location of major beach forms as significant components of a coast-
line depends, to a large degree, on the availability of sand. The primary source
of such debris is the enormous detrital load delivered to the coastal domain
by major rivers. For example, the Mississippi River provides more than 3 bil-
lion tons of solid material to the Gulf Coast region annually. Sea cliff erosion
adds a significant amount of sediment, but normally it is less than 10 percent
of the total debris available for accumulation in beaches (Komar
1976). Small
additions are made to the detrital load by the wind or by slow onshore trans-
port of sand eroded from unconsolidated shelf regions.
The Beach Profile

Beaches represent dynamic systems where loose granular debris is moving
steadily under the attack of waves and currents. If water motion could be held
constant, this debris would be molded into a characteristic profile that would
reflect an equilibrium between the driving forces (waves and currents) and the
properties of the beach sediment. We can thus visualize an equilibrium beach
profile for any set of water and sediment conditions. Waves and currents, how-
ever, do not remain constant but change their properties on a daily or seasonal
basis, requiring some response in the process of sediment transportation and
ultimately in the beach profile. Although we can think of the equilibrium pro-
file as the ideal case, under the dynamics of the natural setting its properties
are constantly changing along with the driving forces.
The beach profile, shown in figure 13.13, consists of a number of com-
ponent parts, each of which develops its own diagnostic characteristics. In its
entirety the profile represents a topographic form that induces waves to dis-
sipate energy by breaking. The exact location of breakers is important because
it determines where the greatest amount of energy is expended and which part
of the beach will be subjected to the greatest change. The berm is a nearly
horizontal surface on the backshore portion of the beach. Some beaches have
more than one berm; others have none, especially if the sediment is coarser
than sand. Landward the highest berm terminates at the base of the sea cliff,
and oceanward it joins the beach face (fig. 13.13). Because the berm is formed
Chapter 13 500
by deposition of sediment during backwash, its elevation is determined by how

high the swash runs up the beach and by the grain size of the sediment. As
swash moves up the beach, it loses velocity because of friction and the loss of
water that permeates into the beach debris. Continued upward growth of the
berm surface would require ever higher waves and swash run-up. In this sense
berms are analogous to vertically accreted floodplains, and their rate of up-
ward growth must decrease with time, for only infrequent storm waves can
add sediment to the surface. As Bascom (1964) points out, however, storm
waves also erode the front edge of the berm, thereby reducing its horizontal
length while simultaneously building the remaining surface to a higher level.
The beach face is the sloping section of the beach profile immediately
seaward of the berm. The slope of the beach face is controlled by many factors
and so may range from nearly horizontal to gradients that approach the angle
of repose for unconsolidated sediment. Sediment moves up the beach face with
the swash and down the beach face during the less powerful backwash. The
beach face therefore represents a surface striving to attain some balance be-
tween onshore and offshore sediment transport. Its slope adjusts to provide the
balance.
Beach face slopes are directly proportional to the size of the particles being
moved in the swash zone. Large particle sizes tend to maintain steep slopes
and vice versa (Bascom 1951; Wiegel 1964; McLean and Kirk 1969; DuBois
1972; Wright et al. 1979). We would expect such a relationship because slope
provides the backwash velocity needed to transport sediment of any given size,
but the phenomenon is much more complicated, and beach slopes probably
relate to many factors involved in shoreline dynamics.
In the zone seaward from the beach face, a submerged longshore bar is
commonly, but not always, present. An associated trough develops between
the bar and the beach face. Longshore bars may be absent from the profile,
especially if the beach is steep. On the other hand, where the beach gradient
is low, several bars may be present. The creation of multiple bar systems may
be accomplished in a variety of ways: (1) each bar may reflect the breaker
position for waves of different sizes; (2) the bars may relate to shifting breaker
positions associated with high and low tide levels; or (3) oscillation waves that
break far offshore on low-gradient beaches may re-form over the trough and
break again closer to shore, each episode of breaking creating an associated
bar.
These hypotheses concerning bar formation are all based on the assump-
tion that breaking waves are intimately involved in the formative mechanics.
A number of early laboratory and field studies support that contention (Evans
1940; King and Williams 1949; Shepard 1950), and the relationship is now
generally accepted as being real. Usually breakers establish the size, position,
and depth of the bars and troughs. Larger breakers normally produce features
in deeper water. These generalizations, however, must be conditioned by evi-
dence showing that bar positions may migrate shoreward or seaward ac-
cording to variations in wave height, steepness, and breaker type.
Some bars are continuous parallel ridges that extend unbroken for tens
of kilometers. Bars do not usually display such regularity, however, because
shiftingwave directions realign parts of the bar system and break ridges into
segments with a chaotic distribution. Violent storms also produce drastic
changes in bar directions and patterns.
The maximum depth of bar formation depends on the depth to which wave
action is able to agitate the bottom sediment. Under average wave conditions,
this depth is approximately 10 m below low tide, but large storm waves are
known to move bottom sediment in water as deep as 25 m. The deepest bars,
molded in the fury of violent storms, maintain their position and shape for
long intervals during which normal waves do not influence the bottom. In con-
sidering bars as products of wave action and as controlling factors in the
breaking process, the shallow shoreward bars and troughs are therefore de-
cidedly more important in nearshore wave dynamics.
Beach Morphodynamics
The ideal beach profile is subject to pronounced changes resulting from de-
position of materials that are added to the profile or by eliminating compo-
nents during erosion. In addition, the position of the major parts of the profile
may be changed according to the dynamics of the modifying factors.
It has generally been assumed that the most pronounced changes in the
ideal profile are brought about by the dynamics associated with storm waves.
Where storms are common, the large and vigorous waves tend to destroy or
drastically limit the extent of the berm. The eroded material is simultaneously
shifted to an offshore positionwhere it collects in a series of longshore bars.
In contrast, where storm waves are unusual, the movement of beach sand is
shoreward. This tends to destroy the offshore bars and transport the sediment
to the upper part of the beach where it rebuilds a broad, flat berm. Because
the frequency of storms is a seasonal phenomenon, we can expect the beach
profile to change dramatically from one part of the year to another. Many
authors, therefore, refer to a summer profile, characterized by the absence of
bars and a wide berm, and a winter profile, with no berm and a series of long-
shore bars; an example is shown in figure 13.14. It is probably incorrect to
think of the summer and winter profiles as occurring specifically during those
seasons. What we are really talking about is a profile formed by storm waves
versus a profile generated by swell. Many field studies document the offshore
movement of sands during storm wave conditions and the landward transport
under swell (Shepard 1950; Bascom 1964; Strahler 1966; Gorsline 1966), and
the fact of the process cannot be questioned.
We now know, however, that the summer and winter profiles (fig. 13.14)
probably apply best in certain environmental conditions such as those along
the California coast. In other areas (eastern U.S. beaches and many Austra-
lian beaches), the morphodynamics are much more complex. For example,
numerous studies of Australian beaches (Short 1979; Wright et al. 1979; Short
Chapter 13 502
Figure 13.14.
Generalized summer and winter Summer profile (August-September
profiles along Scripps Pier, •Winter profile (March-April)
La Jolla, Cal (Adapted from
Shepard 1950)
Meters (Horizontal)
and Wright 1981; Wright et al. 1982a; Wright et al. 1982b; Wright and Short
1983) have shown how preexisting beach topography will affect transporting
mechanisms in the surf zone and how fluid motions other than storm waves,
such as rip currents and low-frequency standing oscillations, are extremely
important in determining the beach character.
The detailed studies of Australian beaches has led to the perception that
beaches and surf zones can be classified into six morphodynamic states, each
of which has a distinctly different association of morphology, water motions,
and sediment characteristics (Short 1979; Wright et al. 1979; Wright and Short
1983).
The end-member states in this classification are called reflective and dis-
sipative. Reflective beaches are characterized by steep, linear beach faces and
well-developed beach cusps (discussed later) and berms (fig. 13.15). They tend
to associate with surging breakers, high runup, and minimum setup of sea
level. Dissipative beaches have low-angle, concave-up beach faces attached to
a wide, flat surf zone. Topographically, they have one or more subtle offshore
bars (fig. 13.15). Beach cusps are absent, and berms, if present, are poorly
defined.
In simplest terms, the reflective and dissipative systems differ dynami-
cally because the accumulated wave energy is expended at different places.
most of the incident wave energy is expended at the
In the reflective system,
beach most wave energy is expended offshore,
face. In dissipative systems,
where energy is lost in turbulence as waves break over the bars. This pro-
nounced morphodynamic distinction can be demonstrated on the basis of the
surf-scaling parameter (Guza and Inman 1975), which is expressed as
( = ab u 2 /g tan 2 /8,
where a b is breaker amplitude, cl> is the incident wave radiant frequency

(2tt/T; T = period), is beach/surf zone gradient, and g is acceleration of
gravity.
Berm .Steep beach face

Figure 13.15.
Surging breakers Generalized profile characteristics
and wave properties for reflective
and dissipative morphodynamic
states. (A) Reflective
(B) Dissipative.
(A)
Beach face
Spilling breakers
(B)
Strongly reflective beaches occur when e < 2.5. In these instances, surging
waves prevail. Waves reach the beach face without breaking and surge up the
beach or collapse over a step at its base. Turbulence is therefore confined to
the zone of runup on the beach face. Wave energy that is reflected may be
trapped and initiate strong, standing-wave motion, commonly form of in the
edge waves. This motion promotes the formation of beach cusps. In extreme
dissipative systems, high e values (> 30) prevail. Spilling breakers occur
75-300 m seaward of the beach face, and the wave bores become smaller and
lose much of their energy before they reach the inshore zone.
In Australia, four intermediate states have been identified between the
two distinct end-member states. The intermediate states are much more dif-
ficult to characterize because each contains both dissipative and reflective ele-
ments. In some cases, the morphodynamic state may change from the offshore
to nearshore position or with the high and low tides. This complicates our
understanding enormously because different features may develop during the
tidal cycle or with distance from the shore. It is also apparent that the tidal
range may be significant. In general terms, however, dissipativeness increases
with increasing values of e, increasing swell-wave heights, and decreasing bed
slope. Greatest dissipativeness also prevails in areas with the largest amount
of inshore sediment and during and immediately following a severe storm. As
longshore bars migrate shoreward, transitional states develop, leading to a
steepened beach face and reflective conditions.
Chapter 13 504
Because morphodynamic states are transient from one to another, it is
probably not correct to think of an ideal or equilibrium beach profile. How-

ever, when considered over a long term, any beach and surf zone will assume
a most frequently occurring condition, which Wright and Short (1983) refer
to as the modal state. The modal state is dependent on the forcing wave climate
(the most frequently occurring wave conditions) and the input of the local
environmental setting. Thus, although there may be no all-encompassing equi-
librium profile, any given beach will attain its own particular modal state,
about which variations will occur as the water motions and bottom configu-
rations continuously change.
In addition to the morphodynamics just discussed, large beaches are usu-
ally spatially and temporally altered by currents moving in the longshore di-
rection. Although these currents can be generated in a variety of ways, the
majority of longshore flow derives from the momentum carried by waves
striking the beach in an oblique direction. This phenomenon takes place be-
cause waves are seldom completely refracted and, therefore, usually make their
final approach at some angle to the shoreline. The velocity of a longshore cur-
rent varies directly with wave height and the angle between the shoreline and
the approaching waves. Velocity also increases with distance from shore,
reaching maximum values near the mid-surf zone. Longshore currents trans-
port sediment that has been entrained by wave action, a process referred to
as littoral drift. Because obliquity of wave incidence tends to retard devel-
opment of rip currents, longshore currents are generally continuous and have
the potential to transport sediment for great distances.
Engineers and other scientists have attempted to construct predictive
models to estimate rates of littoral drift. In most cases, their predictions have
relied on empirical correlations between the rate of sand movement and some
estimate of the wave power expended in the longshore direction (see Komar
and Inman 1970; Komar 1976; 1983). In some cases empirical studies show
a good correlation between wave variables and transport rates, and parameters
used as indicators of the longshore components of wave power or energy are
compatible with wave theory (Komar and Inman 1970; Komar 1971a). A
complete understanding of the process, however, is complicated by the fact
that littoral drift occurs in two different modes of action. When wave steepness
is high {H/L >0.03; Bascom 1964) most transport occurs beneath the breaker
zone. But when the steepness is low, the drift functions along the beach face
in a process known as swash transport (beach drift). In this case, waves
breaking obliquely to the shoreline push sand grains up the beach face per-
pendicular to the direction of the wave crests. The backwash, however, af-
fected only by gravity, pulls the water and grains down the beach face in a
direction perpendicular to the shoreline. Thus, mobile grains in swash trans-
port move in a spasmodic, zig-zag fashion (fig. 13.16). Besides wave steepness,
the angle of wave incidence (a) also is important in the drift rate, maximizing
the rate at a 30° angle.
Figure 13.16.
Drift of sediment along a beach
face caused by waves striking at
an angle to the beach.
The littoral drift process is further complicated because sediment can be

moved in suspension or as bedload. Brenninkmeyer (1975) showed that sus-
pended load is important only in a narrow offshore zone. This analysis seems
to contrast with earlier studies, which suggested that suspended load was the
dominant mode of transport (Fairchild 1973; Thornton 1973). Perhaps some
of the difficulty in identifying load types is due to the fact that fine sediment
is commonly winnowed from beach debris and carried away from the near-
shore environment by rip currents and normal offshore diffusion. The rate and
volume of sediment movement is also dependent on the grain size of the trans-
ported load (Duane and James 1980).
Considerable doubt remains about the certainty of evaluation of littoral
drift rates based on empirical or theoretical analyses. For example, Komar
(1983a, p. 17) states, "From the methods described above, the final uncer-
tainty in the evaluation of the net drift could be much greater than the net
drift itself, and quite conceivably the direction of the net drift might be in-
correctly evaluated."
The best evaluation of littoral drift is based on direct measurements of
volumetric loss caused by erosion or beach growth brought on by deposition
of drifting sand. These measurements lend themselves to a budget of littoral
sediment in which all sediment contributions to a beach (sources) and all losses
from a beach (sinks) are taken into account. The main sediment sources are
beach nourishment, and incoming littoral drift.
rivers, sea cliff erosion, artificial
Common sinks are outgoing littoral drift, offshore transport, beach sand mining
by humans, and debris moved down submarine canyons. Over a period of years,
a particular beach will probably have a positive or negative budget. Obviously,
a negative budget indicates more losses than gains, and therefore the beach
is eroding. Positive budgets indicate net deposition.
In practice, making reasonable estimates of the components that make
up the littoral budget are very difficult. However, the positive or negative char-
acter of the budget can be determined by long-term monitoring of erosion and
deposition on the beach. This information is very important in predicting the
potential impacts that might be caused by humans.

Chapter 13 506
The rate of longshore drifting of sand can be significant when considered

on a time scale of human life, and the process is often shown in the accu-
mulation of beach debris against man-made structures such as groins (fig.
13.17). The net littoral drift is the sum

movements initiated by waves
total of
arriving at the shoreline from various directions. Waves may carry debris in
one direction for a short period and then, as conditions change, return the
material to its original position. Such drift reversals may be random or they
may be decidedly seasonal. In California, for example, sand usually drifts
southward in the winter and northward in the summer. Net littoral drift can
be determined only by long-term budget studies that can detail which direc-
tion is the prevailing one. Nonetheless, the process of longshore transport per-
haps causes more problems for coastal engineers than any other shoreline
Figure 13.17. phenomenon. It places sand where we do not want it (across harbors and bays)
Southern coast of Cape Cod,
Mass showing a series of groins
,
and removes sand from locations where we would like to keep it (resort
constructed to protect the beach beaches). Table 13.3 shows net littoral drift along a number of coastlines. The
from erosion Groins trap sand
volume of debris involved in the transport demonstrates the tremendous prob-
that is moving by littoral drift from
right to left on photo Mouth of
lems engineers face.
river is protected by jetties
Table 13.3 Representative rates of littoral drift along coasts of the United States
Location Predominant Rate of Method of Years of

Direction of Drift (cu yd Measure of Record
Drift per year) Rate of
Drift
Atlantic Coast
Suffolk Co NY., w 300,000 Accretion 1946-1955
Sandy Hook. N.J. N 493,000 Accretion 1885-1933
Sandy Hook, N.J N 436,000 Accretion 1933-1951
Asbury Park, N J N 200,000 Accretion 1922-1925
Shark River, N.J N 300,000 Accretion 1947-1953
Manasquan. N.J N 360,000 Accretion 1930-1931
Barneget Inlet, N J S 250,000 Accretion 1939-1941
Absecon Inlet, N.J. S 400,000 Erosion 1935-1946
Ocean City. N.J. s 400,000 Erosion 1935-1946
Cold Spring Inlet, N.J. s 200,000 Accretion —
Ocean Md.
City. s 150,000 Accretion 1934-1936
Atlantic Beach, N.C E 29,500 Accretion 1850-1908
Hillsboro Inlet, Fla. s 75,000 Accretion —
Palm Beach. Fla s 150,000 Accretion 1925-1939
to
225,000
Gulf of Mexico
Pinellas Co., Fla. s 50,000 Accretion 1922-1950
Perdido Pass. Ala w 200,000 Accretion 1934-1953
Galveston. Texas E 437,500 Accretion 1919-1934
Pacific Coast
Santa Barbara Calif E 280,000 Accretion 1932-1951
Oxnard Plamshore, Calif. S 1 ,000,000 Accretion 1938-1948
Port Hueneme, Calif S 500,000 Accretion 1938-1948
Santa Monica, Calif s 270,000 Accretion 1936-1940
El Segundo, Calif. s 162,000 Accretion 1936-1940
Redondo Beach, Calif s 30 000 Accretion —
Anaheim Bay, Calif. E 150.000 Erosion 1937-1948
Camp Pendleton. Calif S 100,000 Accretion 1950-1952
Great Lakes
Milwaukee Co., Wis. S 8,000 Accretion 1894-1912
Racine Co., Wis. s 40,000 Accretion 1912-1949
Kenosha, Wis s 15.000 Accretion 1872-1909
III. State Line to Waukegan s 90,000 Accretion —
Waukegan to Evanston, III. s 57,000 Accretion —
South of Evanston, III. s 40,000 Accretion
From J W Johnson 1956 Used with permission of the American Association of Petroleum Geologists
Chapter 13 508
Shoreline In addition to the modal profile that is oriented perpendicular to the shoreline,
Configurations many coastal geomorphologists now accept Tanner's (1958) premise that
and Landforms beaches develop a shoreline configuration revealing another type of balance
between water energy and sediment supply. This plan-view shape is best es-
tablished where no long-term unidirectional movement of sediment occurs
parallel to the shoreline. Like a graded river, the shoreline configuration is
developed "over a period of years" and is adjusted to the prevailing wave char-
acteristics.
It would seem that the most logical environment to preserve a modal con-
Figure 13.18. figuration would be protected bays where there is no dominant longshore cur-
Sand barrier off the coast of
northern Nantucket Island, Mass .
rent. In such a locale, waves might produce minor longshore transport of
separated from the mainland by a sediment where the wave crests are not completely refracted. The sediment
lagoon Large cuspate features will continue to drift until the shoreline is reoriented parallel to the attacking
have been formed on the lagoonal
side of the barrier.
wave crests at every segment. Offshore topography thus determines wave re-
fraction, and the waves in turn establish the shoreline configuration. The com-
plications of even this simple model are staggering. For example, any additional
sources of sand to the beach will prevent complete refraction because some
longshore transport away from the source will be necessary. In addition, bottom
topography is so variable and so susceptible to change that it seems too much
to expect that we can precisely describe the form of an equilibrium shoreline.
In light of the above, it is indeed remarkable that many shorelines, open
to all types of waves and currents, contain features that are similar in shape
and spaced with a regularity that can hardly be attributed to coincidence.
Usually crescentic, the features form as periodic seaward projections of the
shoreline itself or, on straight shorelines, as narrow pointed accumulations of
sediment piled perpendicular to the shore. In either case the seaward projec-
tions are separated by a curved embayment (fig. 1 3. 1 8). A complete hierarchy
of these features seems to exist (Dolan and Ferm 1968; Dolan et al. 1974),
ranging from minor forms with a wavelength of less than a meter to major
cuspate-like indentations of the coastline with spacing measured in hundreds
of kilometers. Table 13.4 shows this hierarchy.
Beach Cusps
The most common crescentic forms are beach cusps. Cusps develop at the
upper part of the beach face and along the outer fringe of the berm. They are
usually spaced less than 30 m apart and can form in beach sediment of any
size, including boulders and cobbles (Russell and Mclntire 1965). Some sorting
is produced in the formative mechanics because the cusp projections, or horns,
are usually more coarse-grained than the intervening embayments.
Table 13.4 H erarchy of crescentic landforms found on coasts
Form Cusplet Cusp Sand Waves Secondary Capes Primary Capes

Characteristic
Spacing to 3 m 3 to 30 m 1 00 to 3000 m 1 to 100 km 200 km
Material Fine sand-gravel Sand-boulders Sand Sand Sand-gravel
Topographic Step Berm, beach face Beach berm-offshore Coastal plains; shores Coastal plain
Association bar system with sufficient deltas
sediment
Rhythmicity Yes Yes Yes Often Not always
Motion Fixed Normal to beach Downdrift Probably downdrift Slow downdrift
Temporal Minutes to hours Hours to days Weeks to years Decades Centuries
Suggested Swash action on Berm deposition Wave action, Kinematic nature of Wave action,
Processes beach face, and erosion nearshore sediment transport, confluence of
Groove erosion circulation cells, circulation cells coastal currents,
back eddies of back-set eddies,
longshore and shoals
transport currents
From Dolan el al 1974 Used with permission of Zeitschnft fur Geomorphologie. published by Gebruder Borntraeger. Stuttgart
.
Chapter 13 510
Beach cusps seem to form most readily where wave crests strike parallel
to the shoreline. This perhaps explains why the features tend to remain fixed
in their positions, although laboratory studies indicate that some longshore
migration of the forms is possible if the drift is not excessive (Krumbein 1944).
There seems to be some agreement that spacing of cusps is related to the wave
height as well as wave direction; the higher the waves, the greater the spacing
Even this, however, cannot be pronounced as an inviolate rule; A. T
interval.
Williams (1973) found no correlation between wave height and spacing. The
spacing of the cusps he studied in a Hong Kong bay were most closely related
swash distance, i.e., the length between the breaker zone and the highest
to the
encroachment of the swash.
Because every study of beach cusps seems to reveal some contradiction of
earlier studies, it should come as no surprise that the origin of this feature has
been controversial since its earliest description. The simplest explanation of
the feature, proposed by D. W. Johnson (1910, 1919) and modified by Kuenen
(1948), is based on a process that causes irregular erosion of the beach face.
Swash erosion of the beach face initiates cusps, and backwash transports the
sediment away. Eroded materials are carried seaward until they deposit as
deltaic projections that stand opposite the excavated hollows. Continued swash
action progressively transforms the initial hollows into larger embayments until
the water crossing the depressions reaches a critical depth. Swash is then re-
fracted in such a way that coarse sediment is deposited on the cusp horns and
finer sediment is transported farther offshore. Bays and horns grow until the
central area attains a limiting depth, swash action is retarded, and the feature
assumes an equilibrium condition.
Actually, the process is more complicated, and it may be that cusps can
be generated in several ways. It is now clear, however, that many investigators
believe that most cusps are formed in coastal settings that produce a coupling
(Bowen and Inman
action between incident-surging waves and edge waves
1969, 1971; Bowen 1973; Komar 1973; Guza and Inman 1975; Dolan et al.
1979; Holman 1983). In these settings (primarily reflective state), edge waves
augment the breakers systematically in the longshore direction. At positions
where the breakers are relatively high, recesses are cut into the beach face or
berm by the swash-backwash sequence. The cusp development occurs within
minutes or hours after the generating event begins, and following the event
the shoreline displays the rhythmic cusp pattern on the beach surface.
It may be possible that cusps can form in other wave climate settings. For
example, DuBois (1978, 1981) and Sallenger (1979) suggest that cusps have
developed under plunging wave attack. Cusps in these cases form gradually
after swash extending over a berm is ponded in low areas on the berm. As the
ponded water is allowed to return seaward, it cuts channels through the berm
that are molded into cusps by action during tidal variations.
51 1 Coastal Zones —Processes and Landforms
Large-Scale Rhythmic Topography and Capes

Larger features in the hierarchy of crescentic forms have been called rhythmic
topography by Komar (1976). They are of two main types: (1) crescentic bars,
and (2) rhythmic variations that are controlled by rip currents associated with
cell circulation. The latter forms have been referred to variously as sand waves,
giant cusps, or shoreline rhythms. The features occur in widely divergent en-
vironments including large lakes (Evans 1938; Krumbein and Oshiek 1950),
enclosed seas (King and Williams 1949), and open ocean coastlines in many
parts of the world.
Rhythmic topography differs from beach cusps in several important ways.
(1) Rhythmic topographic features commonly migrate parallel to the shore-
line. The rate of the movement varies but can be a kilometer or more a year.
(2) Much of rhythmic topography is submerged. The emergent forms are usu-
ally observed as giant cusps or sand waves that have considerably greater
spacing than beach cusps. The wavelength of these features ranges from 100
to 3000 m, with horns extending oceanward as much as 25 m. The submerged
rhythmic topography is usually in the form of crescent-shaped sand bars or
longshore bars that are segmented by rip current channels. The crescentic bars
are concave toward the shoreline. They commonly stand opposite the horns of
large cusps exposed on the beach face itself, but they also exist off straight
beaches, especially in protected bays with a small tidal range (Shepard 1952,
1963; King and Williams 1949; Bowen and Inman 1971). (3) Rhythmic to-
pography seems to be more dependent on the bottom configuration in the off-
shore surf zone than normal beach cusps.
Uncertainty about the origin of rhythmic topography equals the mystery
surrounding beach cusps. There is little doubt that rhythmic topography is a
function of wave action, cell circulation with associated rip currents, and long-
shore transport of sediment driven by obliquely striking waves. Precisely how
these combine to construct the features and the details of the processes is simply
not clear.
In addition to the cuspate features already examined, many coasts of the
world display extremely large shoreline crenulations called capes. The map in
figure 13.19 shows that they are prominently developed on the southeast coast
of the United States, where their spacing is roughly one to two orders of mag-
nitude greater than rhythmic topographic features. Capes may be partially
relictand therefore not necessarily adjusted to modern conditions. For ex-
ample, some of the capes shown on the map are fringed by barrier islands that
are Holocene in age. Some also coincide with the position of major rivers (W. A.
White 1966; Hoyt and Henry 1971), which may indicate that the seaward
projections began as ancient deltas.
Many workers feel that these large-scale rhythms are related to rotational
cells that are set up as eddy currents along the western edge of the Gulf Stream.
Such eddies have not been proven to exist, however, and the direct cause of
the capes is unknown.
Chapter 13 512
Figure 13.19. o^
c
Large capes and crescentic Chowan ,S°'
recessions along the south

Atlantic coast of the United ^
States.
CAPE HATTERAS
5
"~ \
Pamlico
Sound
'
rnlic
pa
River
CAPE LOOKOUT
ATLANTIC OCEAN
V
CAPE FEAR
\
15 15 3 45
Approx. scale (miles)
CAPE ROMAIN
Erosion of shorelines constitutes one of the major problems facing scientists, Erosional Landforms
engineers, and land managers throughout the world. In the United States alone, and Rates
approximately 25 percent of our coastlines have been categorized as seriously
eroding (U.S. Army Corps of Engineers 1971). The annual cost of prevention
techniques is staggering and will probably increase in the future because of
the population strain placed on the coastal environment. Shoreline changes
along the Great Lakes and along the marine coastlines of the United States
have been estimated by May et al. ( 983). On a national scale, U.S. shorelines
1
are receeding at an average rate of 0.8 m/yr. Rates vary on a regional scale
from 0.0 m/yr along the Pacific coast to minus 0.8 m/yr along the Atlantic
and minus 1.8 m/yr (negative indicates erosion) along the Gulf coast region
(table 13.5). Great Lakes shorelines are retreating at 0.7 m/yr. Significantly,
rates vary drastically on a local basis where they depend on geology and wave
climate. In fact, net accretion occurs commonly on a local scale over the pe-
riods of record (fig. 13.20). Because of this, it is unrealistic to consider average
rates as suitable for land management evaluation. Kuhn and Shepard (1983)
have shown conclusively that coastal erosion (in this case sea cliff retreat) is
episodic, site-specific, strongly related to meteorological conditions, and influ-
enced by man-induced factors. Clearly, then, rates presented here are not meant
to be utilized in a predictive manner.
Figure 13.20.
Shoreline erosion and accretion
Shoreline trend
trends along the Atlantic and Gulf
| j
Erosion coasts of the United States
I Accretion
Chapter 13 514
Table 13.5 Rate of shoreline change along U.S. marine coasts and bays and
lakes, <x indicates within -state standard deviation of rates
Marine Coasts
Region X, m/yr3 a Total Range 3 A/

6
Atlantic Coast -0.8 22 25.5 -24.6 510

Maine -0.4 0.6 1.9 -0.5 16
New Hampshire -0.5 — -0.5 -05 4
Massachusetts -0.9 19 4.5 -4.5 48
Rhode Island -0.5 0.1 -03 -0.7 17
New York 0.1 3.2 188 -2.2 42
New Jersey -1.0 54 25.5 -15.0 39
Delaware 0.1 2.4 5.0 -2.3 7
Maryland -1.5 3.0 1.3 -88 9
Virginia -4.2 5.5 0.9 -24.6 34
North Carolina -0.6 2.1 94 -6.0 101
South Carolina -2.0 3.8 5.9 -177 57
Georgia 0.7 2.8 5.0 -4.0 31
Florida -0.1 1.2 5.0 -2.9 105
Gulf of Mexico -1.8 27 8.8 -153 358
Florida -0.4 1.6 8.8 -4.5 118
Alabama -1.1 0.6 -08 -3.1 16
Mississippi -06 2.0 0.6 -64 12
Louisiana -4.2 3.3 3.4 -15.3 106
Texas -12 1.4 0.8 -5.0 106
Pacific Coast -o.o 1.5 10.0 -5.0 305
California -0.1 1.3 10.0 -4.2 164
Oregon -0.1 1.4 5.0 -5.0 86
Washington 0.5 2.2 5.0 —3 9 46
Alaska -24 2.0 29 —60 69
Bays and Lakes
Region x, m/yr a a Total Range 3 N"
Delaware Bay
New Jersey -1.9 1.3 03 -3.0 13
Delaware -1.3 2.1 5.0 -3.0 12
Chesapeake Bay -0.7 0.7 1.5 -4.2 136
Western shore -0.7 0.5 1.5 -19 2"
Maryland -0.7 0.3 -0.1 -1.3 35
Virginia -08 0.7 1.5 -19 32
Eastern shore -0.7 0.8 01 -42 69
Maryland -0.8 0.9 -03 -4.2 47
Virginia -0.5 0.4 01 -12 22
Great Lakes -0.7 0.5 0.6 -2.7 327
Lake Erie -0.7 0.6 -0.2 -2.4 98
Ohio -0.6 0.6 -02 -2.2 68
Pennsylvania -0.3 0.1 -0.2 -04 14
New York -1.4 0.6 -05 -24 20
Lake Ontario -0.5 0.2 -02 -12 58
Lake Huron -0.4 0.3 -0.3 -1 3 28
Lake Michigan -0.6 0.8 0.6 -99 184
Western shore -0.6 0.4 06 -1 5 62
Eastern shore -07 0.9 0.3 -99 122
Wisconsin -0.7 0.3 -0.3 -15 46
Illinois -0.2 0.4 0.6 -09 16
Indiana -0.4 0.5 -0.3 -09 12
Michigan -07 09 -0.3 -99 110
Lake Superior -1 3 07 -0.3 -27 35
Minnesota -0.8 04 -03 -15 16
Wisconsin -1 8 0.6 -0.9 -27 19
From May el al issue 35 1983 p 522 Copyright by the American Geophysical Union Reprinted By permcssKX
3 Negative values indicate erosion, the positive values indicate accretion
"Total number ot 3 mm grid cells over which the statistics are calculated
51 5 Coastal Zones — Processes and Landforms
Table 13.6 Major erosional processes functioning along coasts
Process Description
Corrosion Solution of coastal rocks by chemical action of seawater

Attrition Diminution of rock particles as water rolls, bounces, or slides them
on a beach or wave-cut platform.
Corrasion Physical erosion of bedrock caused by the grinding action of rock
fragments that are carried in the ocean waves and currents.
Hydraulic Action Erosion caused by the force of the water itself Includes wave shock
pressure and pneumatic quarrying by air trapped in cracks of the
headland rocks
Much of the erosion included in the rates just provided involves removal
of sand from the beach itself. However, because of increased use of the entire
coastal zone, engineers and coastal zone managers have placed considerable
attention on the processes and rates of sea cliff erosion. When considering sea
cliff erosion, several facts become immediately apparent. First, retreat of a
sea cliff requireswave erosion at the cliff base. Second, erosion at the base
leads to increased mass movement of the sea cliff material because of the re-
sulting increase in slope angle and shear stress. Third, the debris of mass
movement collects at the cliff base. No base erosion can occur again until this
debris removed and the toe of the cliff is again exposed to wave attack. The
is
erosion is accomplished by a group of geomorphic processes, listed in table

13.6, that function in a complex of interactions to produce a variety of land-
forms. The final effect sometimes is controlled by ancillary processes such as
burrowing action of marine organisms (see Ahr and Stanton 1973), frost ac-
tion, and mass movement. As might be expected, the effectiveness of each
process varies with the properties of the shore material and with the particular
dynamics of the local ocean or lake system. Corrosion affects those rocks that
are most susceptible to solution. At normal temperatures seawater is saturated
with respect to calcium carbonate and does not dissolve limestone and CaC0 3
directly. Solution of these materials, however, may be aided by rainwater or
by organisms that create local acidic conditions. As chunks of coastal or lake-

shore rocks are released, attrition decreases the size of the particles and allows
subsequent waves to drive the sediment into the cliff face. Corrasion then as-
sumes a dominant role in the erosive mechanics. Hydraulic action is especially
important where the rocks are highly fractured. Not only does the force of the
wave exert pressure on the cliff face, but the advancing waves may compress
air in the rock cavities, producing a pneumatic effect in the cracks. As the
wave recedes, external pressure is instantaneously released, and the com-

pressed air within the rocks exerts an outward stress that may disaggregate
the outer zones of the cliff face.
have been documented by a variety of techniques in-

Cliff erosion rates
cluding comparison of sequential ground and aerial photography and maps,
exposure of pins inserted into the sea cliff (Hodgkin 1964), instruments de-
signed to measure microerosion (Trudgill 1976; Robinson 1977), and detailed
air-photogrammatic maps (Norrman 1980).
Chapter 13 516
Table 13.7 Representative rates of sea cliff retreat.
Location Material Rate m/100 yr) Reference
New England Crystalline rock 1
USSR. Volcanics 2
England Crystalline rock 3
(Cornish coast)
Northern France Chalk 25 3
England (Yorkshire) Sedimentary rocks 9 3
England (Yorkshire) Glacial drift 28 3
Louisiana Sands and clays 800- 3800 4
(coastal islands)
Southern California Alluvium 30 5
USSR. Clay 1200 2
New Jersey Sand, clay and gravel 180 6
Cape Cod, Mass. Glacial drift 30 7
1) Johnson 1925 5) Shepard and Granl 1947

2) in Zenkovich 1967 6) Rankin 1952
3) in King 1972 7) Zeigler et al 1959
4) Peyronnin 1962
Average erosion rates are shown in table 13.7 to demonstrate the vari-
ability of sea cliff erosion under different conditions. A much more extensive
table of cliff erosion rates is presented in Sunamura (1983). It is clear that
rates of sea cliff erosion are extremely variable. This variability derives pri-
marily from differences in geology, wave climate, and time. Lithology and co-
hesiveness of the coastal material clearly are of paramount importance in
determining the rate of cliff retreat (fig. 13.21). Unconsolidated debris is eroded
most rapidly and, in some cases, catastrophically. For example, short-term
rates resulting from storms can be tens of meters of erosion in one day. Rocks
that are nonresistant to wave attack are usually friable sandstones and shales.
Rocks that resist sea cliff retreat are massive igneous rocks, high-rank meta-
morphic rocks, and certain massive carbonates.
In a series of papers, Sunamura (1975, 1976, 1977, 1978, 1982) has ex-
amined cliff erosion in theoretical terms of driving and resisting forces. The
assailing wave force (/,.) is initially determined by energy derived in deep water,
but it is directly influenced by factors such as water level or tide, bottom and
beach topography, and beach sediment. These determine the wave type and
height, where the waves break, and at what level they strike the sea cliff. Un-
is no way to directly measure /„.; therefore, some related
fortunately, there
parameter such as wave height is used as an estimate of the driving force.
Although lithology is the primary determinant of resistance (/",)< it is also
influenced by contributing factors such as mechanical strength and geologic
structures. Most workers use compressive strength as a surrogate for resis-
tance.
51 7 Coastal Zones —Processes and Landforms
Figure 13.21.
Generalized orders of erosion
Granitic rocks - rates for sea cliffs composed of
different lithologies and
mechanical strength (Data from
Limestone - Sunamura 1983)
Flysch and shale -
Chalk and Tertiary _

sedimentary rocks
Quaternary deposits -
Unconsolidated _
volcanic ejecta
2
10 10 10 1
10° 10 1 10 2 10 3
Cliff erosion rates (m/yr)
Thus, the cliff erosion rate becomes a function of fw/f„ which Sunamura
(1977, 1983) has expressed as
where H is wave height at the cliff base, S t

is compressive strength, A: is a
dimensionless constant, and Ca constant with dimensions of LT~ }
. This equa-
tion suggests that a critical wave height (Hcnl ) exists that is needed to cause
sea cliff erosion and can be derived by setting dX/dt = 0. This yields
H = —s- e~
c ,
c
. In areas of relatively homogeneous parent material and distinct
wave climate, this approach is quite significant. For example, at Byobugaura,

on the Pacific coast of Japan, a homogeneous Pliocene mudstone forms the
lower half of a 10-60 m high cliff. Sunamura (1982) was able to determine
k and C and H^ nl by using data of cliff strength, wave climate, and
calculate
erosion rates of two time intervals. The plots shown in fig. 13.22 indicate that
erosion was caused only under attack of the rarer but larger waves. The smaller,
frequent waves produced no sea cliff erosion. In contrast to our earlier remark
about average rates, this approach led to the identification of a threshold con-
dition that has direct application in coastal land management. It also illus-
trates the problem of using average rates because erosion occurs in spasms
(fig. 13.23) of extremely rapid erosion rather than continual removal of ma-
terial at a constant rate from the cliff face. Short-term measurements may be
far from reality if measured during unusual times, and long-term analyses
r
Chapter 13 518
Figure 13.22.
•men
m,-
met
Wave occurrence frequency, wave
duration, erosion rate and
;
recession distance plotted against i =
wave height at Byobugaura, H - o
Japan
CD
t .05 -
"O
>
(0
m/y
50
X H„ - dX/dt
T3 40
30
20
10
i i i i
4 -
Ho" X
2 -
g
ifl
to
1
-
0)
u
d) i i i i
DC
.5 1.5 2.4 3.3 4.2 5.1 m 5 1.5 2.4 3.3 4.2 5.1m
Offshore wave height. H Offshore wave height, H
Figure 13.23.
Schematic illustration of cliff
retreat rates using a long-term

average versus the actual
continuum of erosion within the
long-term interval
Time (yrs.) -
may mask the important aspect of prevailing wave climate. Clearly, as Su-
namura (1983) suggests, the rate of sea cliff erosion should be determined by
the occurrence frequency of waves exceeding the threshold height value.
The base of any sea cliff is periodically covered by debris. This occurs
during intervals of rapid subaerial erosion when climate change or human ac-
tivities deliver more debris to the beach surface than can be removed by wave
action (Emery and Kuhn 1982; Kuhn and Shepard 1983). This material pre-
vents further undercutting of the cliff until stronger wave action removes it
and thereby exposes the cliff face to renewed supercritical wave attack. The
implication is that the sea cliff profile will vary with time (Emery and Kuhn
1982; Kuhn and Shepard 1983; Sunamura 1983). These variations tend to
offset one another such that, over a long term, the profile may appear to retreat
in a parallel manner. However, within that long term, the profile will change
according to the amount of debris stored on the beach and/or whether the
beach elevation causes supercritical waves to strike the cliff face at higher or
lower levels.
Emery and Kuhn (1982) have used profiles to classify sea cliffs as active,
inactive, orformer. These apparently indicate whether marine or subaerial
processes are dominant at any given time. A smooth curve at the base (in-
active) indicates that subaerial erosion is prevailing. In contrast, a sharp-an-
gled basal contact (active) with the beach suggests that marine erosion is
dominant. A former profile is one that has been removed from the influence
of marine processes. Because inactivity is associated with dry climates, the
incipient climate change in southern California to more humid conditions may
initiate a change to active sea cliff erosion. Areas presently near inactive cliffs
are probably overdeveloped, which will exaccerbate the effects of the on-
coming climate shift. Humans, it seems, have primed the pump for serious
environmental problems.
As the cliff retreats, it leaves behind a beveled surface, called the wave-
cut platform, that stands slightly below water level at high tide (fig. 13.24).
Although many processes are involved Wentworth 1938),

in its creation (see
corrasion at the base of the sea cliff is probably responsible for most of the
planating action. The platform is not flat but slopes gently oceanward with a
declivity that ranges between 0.02 and 0.01 Therefore, under a stationary sea
.
level the maximumwidth of the wave-cut platform depends largely on the

depth at which wave abrasion is still a viable process. Bradley (1958) suggests
that wave-cut platforms can be eroded in water no deeper than 10 m; under
Chapter 13 520
Figure 13.24. the common slope range, platforms wider than 500 m probably form only if
Wave-cut platform near MacLeod sea level is continuously rising.
Harbor, Alaska Platform has been
raised above wave level by recent In addition to the effect of water depth, on low-gradient platforms and
tectonism with constant sea level, the ultimate width is self-controlling because the rate
of platform expansion decreases as the sea cliff recedes. Platforms develop
rapidly in the early stages of cliff retreat. In time, however, incoming waves
lose much of their energy as they interact with the progressively widening
platform. Eventually, cliff retreat and platform expansion cease unless depo-
sition on the original platform surface raises the level of incoming waves and
changes the frictional component.
As the shoreline retreats and irregularities appear, a group of landforms
develop that are characteristic of coastal erosion. Stacks are isolated parts of
the headland formed when narrow oceanward extensions of the coastal rocks
are cut into isolated remnants by wave attack. The process is accentuated when
waves are refracted around the headland reach. Sometimes less resistant rock
zones are exposed to local corrasion or scouring, and notches or sea caves are
formed. As sea caves grow, they may extend completely through the headland
to produce a feature known as a sea arch (fig. 13.25). Any or all of these
features indicate a local erosional environment. As the headland rocks undergo
attrition, however, more sediment covers the wave-cut platform. Beach pro-
files develop, and wave energy dissipates farther offshore as waves break over
the longshore bars. This sequence may not occur in every situation. For ex-
ample, in large lakes such as Lake Erie or Lake Michigan, energy loss over
longshore bars is of little or no consequence.
521 Coastal Zones— Processes and
Landforms
Figure 13.25.
Sea arch or window along coast
of Puerto Rico at low tide. Hole
is
caused by combination of wave
action and solution of limestone
bedrock.
Chapter 13 522
Depositional In coastal areas where the supply of sand is abundant and local ocean forces
Shorelines are capable of transporting sediment, the shapes of some landforms are de-
termined primarily by depositional events, even though erosional processes are
very much involved in the entrainment of the sand before its final deposition.
Large depositional features usually occur in the form of spits, baymouth bars.
or barrier islands. The first two are definitely related to longshore transport
of sediment (littoral drift), with the site of deposition being in tranquil waters
of bays, estuaries, or the open ocean. Barrier islands are large elongate fea-
tures that parallel the shoreline but are not physically connected to it. Barrier
islands are veryabundant along the Atlantic and Gulf shorelines of the United
States and will be treated in a separate section. Other depositional features
may include some of the cuspate forms discussed earlier (especially cuspate
forelands associated with sand waves). Most of these differ in several ways
from the large depositional beaches. First, the cuspate forms show a regularity
that is not evident in the large beaches and probably relates to different hy-
drodynamics. It is known that some crescentic forms (rhythmic topography)
will migrate in the direction of the longshore transport. However, very strong
longshore drift will at first skew the sand waves and giant cusps
drastically
and then, with increased transport form completely. Second,
rates, eliminate the
cuspate features may be superimposed on the larger beach types, indicating

that in some cases they are secondary rather than primary features.
Spits and Baymouth Bars

Spits and baymouth bars develop when littoral drift plays a predominant role
in the system, provided the drifting sediment enters a zone of slack water where
deposition can occur. Possibly the erosional and depositional framework in the
littoral in terms of a littoral power gradient similar
system can be considered
to that proposed by May
and Tanner (1973). In their study, a wave power
model utilizing £,/\. and dqjdx {q being the quantity of sand transported and
x the distance along the beach) as the basic parameters was constructed to
predict zones of different littoral transport along a beach. Where dq/dx is
greatest, beach erosion is also greatest because a large amount of sediment is
being placed in transit. Where dq/dx is negative, deposition is occurring.

Spits and baymouth bars are essentially the same feature. They differ
only in that spits extend into the open ocean while baymouth bars cross the
gap between two headland reaches. Baymouth bars are also more likely to
initiate lagoonal areas shoreward of the bar deposit. These lagoons may grad-
ually change into tidal marshes or swamps as the embayments fill in with flu-
vially derived sediment.
Although and baymouth bars have been attributed to a variety of

spits
ocean processes, their origin is clearly and most prominently related to littoral
drift. The elongate extensions into open water represent continuations of the
beach that rests against the coast. They expand continuously in the direction
of littoral drift unless other water motions interfere with the growth process.
For example, it is not uncommon for wave refraction around the free end (Evans Figure 13.26.
Recurved spit on west coast of
1942) or wave trains approaching from different directions (King and Mc- Florida Series of beach ridges
Cullagh 1 97 1 ) to reorient the terminal end of the spit. In this case the feature show accretionary pattern of spH
may be called a recurved spit or a hook. Spits or baymouth bars may also
widen by progradation associated with the construction of a series of sandy
or pebbly ridges on the oceanward side of the feature (fig. 1 3.26). These ridges
or growth beach ridges, will be discussed in the next section. Fi-
lines, called
nally, spits may link the main coastal region to an offshore island, producing
the feature known as a totnbolo. Tombolos, however, also form on normal

coastal beaches when wave refraction around the island shapes the beach into
a cuspate foreland that eventually extends to the island itself. This process
even functions around breakwaters that have been constructed offshore to pre-
vent beach erosion (Inman and Frautschy 1966).
Chapter 13 524
Figure 13.27.
Schematic of idealized barrier Barrier Islands
island showing vegetal and
geomorphic facies. (Diagram by
R Craig Kochel) Distribution and Characteristics Barrier islands are elongate bodies of sand
that are not attached to the mainland but are separated from it by a lagoon
or bay. The islands normally range in width from 2km, in length from
to 5
10 to 100 km, and are usually less than 6 m in elevation. They are commonly
large enough to support major cities —
Atlantic City, Miami Beach, and Gal-
veston are examples. Barrier islands occur on 13 percent of the world's coasts
(King 1972), but they appear to be concentrated where tidal ranges, wave
energies, and offshore gradients are low. Along the boundaries of the United
States, 282 barrier islands exist adjacent to the Gulf coast and the Atlantic
coast between Florida and the middle Atlantic states (Dolan et al. 1980).
The origin of barrier islands has always been steeped in controversy, mainly
because most of the original interpretations were based on morphology alone.
More recent theories rely on subsurface exploration and paleoenvironmental
interpretation of facies both shoreward and seaward of the barrier. Although
differences of opinion exist, most workers seem to believe that barrier islands
have a long growth history associated with postglacial sea level rise. Re-
member that during the last full glacial episode, sea level was probably 120
m lower than it is today, and the shorelines possibly 60 to 150 km seaward of
their present positions (Curray 1965; Dolan et al. 1980). Beginning as shore-
line, depositional nuclei during the low sea level of the late Pleistocene, the
initial ridges migrated landward as sea level rose. Sea level approached its
present level about 5000 years ago. at which time our modern barrier
4000 to
islands developed their distributional and environmental character. The is-
lands arc still evolving, and because sea level is still slowly rising, they continue
-
Mainland—}- Lagoon }- Marsh L.Dunes-j-Beach-f- Ocean

(wind tidal flat)
Figure 13.28.
to migrate landward. This movement, however, requires the unique combi- idealized cross profile of a barrier
nation of environments, sand supply, and sand transport found in the barrier island
system.
Barrier islands are composed of distinct geomorphic and vegetative zones
(figs. 13.27, 13.28). On the ocean side, the islands are characterized by low-
gradient beaches that alternately change their configuration during storms
and during intervals in which a swell-wave climate prevails. Significantly, the
backshore environment of most beaches is characterized by sand dunes rather
than a distinct sea cliff. The dune line represents the "backbone" of the island,
usually standing up to 6 m in elevation but occasionally reaching much greater
heights (fig. 13.29). For example, in a study of southeast African coasts, Orme
(1973) reports Holocene dunes climbing to elevations greater than 100 m.
Dune elevation and beach properties are closely related to the prevailing wind
and wave direction. Where beach orientation is dunes
conducive to littoral drift,
may be stunted or even absent. These conditions result narrow islandsin low,
that are extremely vulnerable during storm surges. In contrast, where beaches
are perpendicular to the prevailing wind and wave direction, sand is pushed
shoreward and eventually is driven by the wind into pronounced dunes.
Behind the dune line is a low, flat zone that is covered with grasses and
shrub forests; stands of pine and oak are sometimes found in sheltered areas
(fig. 13.27). This zone grades into the salt marshes and tidal flats that abut
the lagoon itself. The lagoon, salt marshes, and tidal flats are fed and main-
tained by tidal inlets that link the open ocean to the sound-side environments.
The Geomorphic Processes In addition to the wind action associated with

dune formation and sand transport, two of the most important processes func-
tioning in barrier island systems are overwash and inlet formation (Dolan et
al. 1980). In severe storms, parts of every barrier island are inundated by high
water levels and affected by wave action. The process, known as overwash, can
cover large areas, especially where dune development is minor. Overwash may
also occur in narrow washover channels that breach the dune line (fig. 1 3.27).
In such cases, sand is transported from the beach and deposited as distinct
geomorphic features known as washover fans (Pierce 1970; fig. 13.30). The
total transport distance of sand during overwash depends on the tidal range,
severity of the storm, island elevation, and storm surge penetration. It is pos-
sible, however, for overwash to traverse the entire dune line and place fans on
top of older sound-side deposits.
Chapter 13 526
Figure 13.29.
Giant dune at Jockeys Ridge.
NC
Figure 13.30.
Outer Banks. NC Stabilized dune
breached by washover fan Old
road surface has been eroded by
ocean and buried by recent
overwash deposits as the barrier
island migrated landward
Figure 13.31.
Perhaps of equal or greater importance is the fact that major inlets are
Ebb tidaldelta and flood tidal
probably formed during storms. These essentially segment the islands and delta at Brown Inlet, N.C. Ebb
physically connect the lagoon with the open ocean. The positions of most inlets tidal delta marked by offshore
shoaling of waves
are ephemeral because they tend to migrate in the direction of littoral drift,
extending by spit formation on one side of the inlet and erosion on the other.
Some inlets may simply fill in with sediment, thereby destroying the free
channel. Nonetheless, when active, inlets serve as avenues for water and sand
movement into and out of the lagoonal zone. During high tide, material moves
landward through the inlet and deposits a flood tidal delta on the inside of the
barrier island. During ebb tide, a similar ebb tidal delta is formed in the ocean
(figs. 13.27, 13.31). During low-tide events, the sound-side shoals are exposed
and eventually become the substrate for new salt marshes (Godfrey 1976).
Enormous investments have been made in attempts to fix tidal inlets in a
single location. Usually jetties are constructed to prevent filling of the channel
and to stabilize its position. In many cases these efforts fail without continuous
and costly maintenance. A good example is the Ocean City Inlet, Maryland,
which breached a barrier island during a 1933 hurricane. Since construction,
the jetty system has interrupted the southward net littoral drift (Dean and
Perlin 1977). The effect has been a shoreline advance of about 240 m im-
mediately north of the inlet, where drifting sand has been trapped by the
Chapter 13 528
northern jetty. The shoreline immediately south of the inlet has eroded land-
ward about 515 m because the beach was starved when its supply of littoral
sediment was eliminated (fig. 1 3.32). The inshore portion of the south jetty is
so low and permeable that sand being eroded and transported from the beaches
is moving over and through the jetty back into the inlet itself.
south of the inlet
This is producing a shoal area within the inlet and will require major repair
of the south jetty (Dean and Perlin 1977).
There is irrefutable evidence that most mid-Atlantic barrier islands are
still migrating landward (Fisher and Simpson 1979). Evidence for this
transgression is that shells of lagoonal fauna and salt marsh peats are now
found on the beach side of the islands, indicating that the present beach is
resting on top of older lagoonal sediment. Clearly, the barrier island system
is dynamic. Islands are eroded on the ocean side, and the entrained sediment
is transported farther inshore (fig. 13.30). The lagoonal
side of the island system
grows by deposition of overwash and tidal inlet sediment, and windblown sand.
Thus, the landward march of the islands will proceed as long as sea level con-
Figure 13.32. tinues to rise and overwash and inlet processes along with wind action provide
Inlet and jetties at Ocean City,
Md ,
looking north Barrier island
sediment to cover the lagoonal facies. This covering simultaneously creates a
south of (Assateague Island)
inlet new substrate for continued growth of the vegetational zones located inshore
has migrated landward relative to from the island dunes.
island north of inlet (Fenwick
Island)
Coastal zones reflect the balance between the driving forces of ocean waters Summary
and the resistance offered by the rocks that form the shoreline. Large lakes
may experience the same phenomena as ocean coasts. Beaches are the com-
ponents of coasts that respond most obviously to the dynamics of the system.
Energy possessed by waves, currents, and tides is expended on the beach sur-
face. The response to this activity is the creation of a beach profile that is
adjusted to the mean values of the wave properties. Waves striking the beach
zone obliquely also cause longshore (littoral) drift of sediment.
Coasts and beaches display a wide variety of geomorphic forms that range
in size from minor modifications of the beach face to features that encompass
kilometers of the shoreline. Commonly the shoreline is indented with a hier-
archy of crescentic or rhythmic forms that are related in some way to cellular
circulation of the ocean water. Coasts may be typified as erosional or depo-
sitional depending on whether the dominant action is causing the shoreline to
recede landward or expand oceanward. Each coastal type possesses large-scale
features that reflect the prevailing action. In some cases (barrier islands), both
erosion and deposition are necessary to maintain the feature.
cussed in this chapter. Each reference cited has an extensive bibliography of
topics that may be of particular interest to the reader.
Dolan, R.; Hayden, B.; and Lins, H. 1980. Barrier islands. Am. Scientist
68:16-25.
Komar, P. D. 1976. Beach processes and sedimentation. Englewood Cliffs,
N.J.: Prentice-Hall.
, ed. 1983. CRC handbook of coastal processes and erosion. Boca
Raton, Fla.: CRC Press.
May, S.; Dolan, R.; and Hayden, B. 1983. Erosion of U.S. shorelines. EOS
64:521-23.
Sunamura, T. 1982. A predictive model for wave-induced erosion, with
application to Pacific coasts of Japan. Jour. Geology 90:167-78.
Wright, L.; Chappell, J.; Thorn, B.; Bradshaw, M.; and Cowell, P. 1979.
Morphodynamics of reflective and dissipative beach and inshore
systems: Southeastern Australia. Marine Geol. 32:105-40.
.
Bibliography
Abrahams, A. D. 1972. Ahnert, F. 1970. Functional Anderson, R. S.; Hallet, B.; Walder,
Environmental constraints on the relationships between denudation, J.; and Aubry, B. F. 1982.
substitution of space for time in relief, and uplift in large mid- Observations in a cavity beneath
the study of natural channel latitude drainage basins. Am. Grinnell Glacier. Earth Surf.
networks. Geol. Soc. America Jour. Sci. 268:243-63. Proc. and Landforms 7:63-70.
Bull. 83:1523-30. Ahr, W. M., and Stanton, R. J. 1973. Andrews, D. E. 1980. Glacially
. 1980. A multivariate The sedimentologic and thrust bed — An indication of late
analysis of chain lengths in paleoecologic significance of Wisconsin climate in western New
natural channel networks. Jour. Lithotyra, a rock-boring barnacle. York State. Geology 8:97-101.
Geology 88:681-96. Jour. Sed. Petrology 43:20-23. Andrews, E. D. 1979. Scour and fill
Abrahams, A. D., and Miller, A. J. Akagi, Y. 1980. Relations between in a stream channel, East Fork
1982. The mixed gamma model rock type and the slope form in River, western Wyoming. U.S.
for channel link lengths. Water the Sonora Desert, Arizona. Zeit. Geol. Survey Prof. Paper 1117.
Resour. Res. 18:1126-36. f. Geomorph. 24:129-140.
.
1981. Measurement and
Ackers, P., and Charlton, F. G. 1971. Alter, A. J. 1966. Sanitary computations of bed-material in a
The slope and resistance of small engineering in Alaska. In Proc. shallow sand-bed stream, Muddy
meandering channels. Inst. Civil Permafrost Internal. Conf. Creek, Wyoming. Water Resour.
Engrs. Proc, Supp. 15, 1970, (Lafayette, Ind., 1963). Natl. Res. 17:131-41.
paper 73625. Acad. Sci. Natl. Research Council 1983. Entrainment of gravel
Addington, A. R. 1927. Porter's Cave Pub. 1287, pp. 407-8. from naturally sorted riverbed
and recent drainage adjustments American Geological Institute. 1972. material. Geol. Soc. America
Indiana Acad. Sci.
in its vicinity. Glossary of geological terms. Bull.94:1225-31.
Proc. 36:107-16. Anderson, D. M. 1968. Andrews, J. T. 1963. Cross-valley
Ahlbrandt, T. S., and Fryberger, Undercooling, freezing, point moraines of the Rimrock and
S. G. 1980. Eolian deposits in the depression, and ice nucleation of Isotoq river valleys, Baffin Island.
Nebraska Sand Hills. U.S. Geol. soil water. Israel J. Chem. A descriptive analysis. Geogr.
Survey Prof. Paper 1 120A. 6:349-55. Bull. 19:49-77.
. 1982. Introduction to eolian . 1970. Phase boundary water ed. 1974. Glacial isostasy.
.
deposits. In Sandstone in frozen U.S. Army Corps

soils. Stroudsburg, Pa.: Dowden,
depositional environments, edited Engrs., Cold Regions Res. and Hutchinson and Ross.
by P. Scholle and D. Spearing, Eng. Lab. Research Rept. 274. Andrews, J. T, and Smithson, B. B.
1 1-48. Tulsa, Okla: Am. Assoc. Anderson, L. W. 1978. Cirque glacier 1966. Till fabrics of the cross-
Petroleum Geologists. erosion rates and characteristics valley moraines of north-central
Ahlmann, H. W. 1948. Glaciological of neoglacial Pangnirtung
tills, Baffin Island. Geol. Soc. America
research on the North Atlantic. fjord area, Baffin Island, N.W.T., Bull. 77:271-90.
Royal Geog. Soc. Res. Ser. 1 Canada. Arc. Alp. Res. Arkley, R. 1963. Calculation of
10:749-60. carbonate and water movement in
soil from climatic data. Soil Sci.
96:239-48.
531
—
Bibliography 532
Arvidson, R. E., and Guinness, E. A. 1976a. Hydrogeology of a

. Barnes, H. A. 1968. Roughness
1982. Clues to tectonic styles in cavernous limestone terrane and characteristics of natural
the global topography of Earth, the hydrochemical mechanisms of channels. U.S. Geol. Surve\
Venus, Mars. Jour. Geol. Educ. its formation, Mohawk River Water Supply Paper 1849.'
30:86-92. basin. New York. Empire State Barnes. and Tabor, D. 1966.
P.,
Atkinson, H., and Wright, J. 1957. Geogram 12, no. 2:2-65. Plastic flow and pressure melting
Chelation and the vertical 1976b. Hydrogeomorphic in the deformation of ice. Nature
movement of soil constituents. methods for the regional 210:878-82.

SoilSci. 84:1-11. evaluation of flood hazards. Barnett. D. M.. and Holdsworth. G.
Atkinson, T., and Smith, D. 1976. Environ. Geology 1:261-81. 1974. Origin, morphology, and
The erosion of limestones. In The 1977. Stream-channel
. chronology of sublacustrine
science of speleology, edited by response to floods, with examples moraines. Generator Lake, Baffin
T. Ford and C. Cullingford, from central Texas. Geol. Soc. Island, Northwest Territories.
pp. 1 5 1 —77. New York: Academic America Bull. 88:1057-71. Canada. Can. J. Earth Sci.
Press. 1978. Adjustment of fluvial 11:380-408.
Atterberg, A. 1911. Die Plastizitat systems to climate and source Barsch, D. 1971. Rock glaciers and
der Tone. Intern. Mitt. Boden terrain in tropical and subtropical ice-cored moraines. Geogr. Annlr.
1:4-37. environments. In Fluvial 53A:203-6.
Baas-Becking, L.; Kaplan, I.; and sedimentology, edited by Barsch. D.. and Updike, R. G. 1971.
Moore, D. 1960. Limits of the A. Miall, pp. 21 1-30. Can. Soc. Periglaziale Formung am
natural environment in terms of Petrol. Geol. Mem. 5. Kendrick Peak in Nord-Arizona
pH and oxidation-reduction 1981. The geomorphology of wahrend der letzten kaltzeit.
potentials. Jour. Geology Mars. Progress in Phys. Geog. Geog. Helvetica. 26:99-114.
68:243-84. 5:475-513. Barshad. I. 1964. Chemistry of soil
Bagnold, R. A. 1941. The physics of Baker, V. R.: Kochel, R. C; Patton, development. In Chemistry of the
blown sand and desert dunes. P. C; and Pickup, G. 1983. soil, edited by F. Bear, pp. 1-70.
London: Methuen and Co. Palaeohydrologic analysis of New York: Reinhold Book Corp.
. 1960. Some aspects of river Holocene flood slack-water Bascom, W. N. 1951. The
meanders. U.S. Geol. Survey Prof. sediments. Spec. Pubs. Intl. relationship between sand size and
Paper 282-E. Assoc. Sediment. 6:229-39. beach face slope. Am. Geophvs.
1973. The nature of
.
Baker, V. R.; Pickup, G.; and Polach, Union Trans. 32:866-74.
saltation and of bedload transport H. A. 1983. Desert palaeo floods 1 964.
. Waves and beaches.
in water. Proc Royal Soc. . in central Australia. Nature Garden City. N.Y.: Doubleday.
London, ser. A:332:473-504. 301:502-4. Bates. C. C. 1953. Rational theory of
1977. Bed-load transport by Baker, V. R.. and Ritter. D. F. 1975. delta formation. Am. Assoc.
natural rivers. Water Resour. Res. Competence of rivers to transport Petroleum Geologists Bull.
13:303-12. coarse bedload material. Geol. 37:2119-62.
Baker, V. R. 1973a. Geomorphology Soc. America Bull. 86:975-78. Bathurst. J. C; Thorne, C. R.: and
and hydrology of karst drainage Bakker, J. P. 1965. A forgotten Hey. R. D. 1979. Secondary flow
basins and cave channel networks factor in the interpretation of and shear stress at river bends.
in east-central New York. Water glacial stairways. Zeit. f Jour. Hydraulics Div., ASCE
Resour. Res. 9:695-706. Geomorph. 9:18-34. 105:1277-95.
1973b. Paleohydrology and
. Baldwin, M.: Kellogg, C; and Battle, W. R. B. 1960. Temperature
sedimentology of Lake Missoula Thorp, J. 1938. Soil classification. observation in bergschrunds and
flooding in eastern Washington. In Soils and Men, U.S. Dept. their relationship to frost
Geol. Soc. America Spec. Paper Agri. Yearbook, pp. 979-1001. shattering. In Norwegian cirque
144. Ballard, T. M. 1973. Soil physical glaciers, edited by W. V. Lewis,
1974. Paleohydraulic properties in a sorted stripe field. pp. 83-96. Royal Geog. Soc. Res.
interpretation of Quaternary Arc. Alp. Res. 5:127-31. Ser. 4.
alluvium near Golden, Colorado. Bandy, O., and Marincovich, L. Bauer, B. 1980. Drainage density
Quat. Res. 4:94-112. 1973. Rates of late Cenozoic An integrative measure of the
. 1975. Urban geology of uplift, Baldwin Hills. Los dynamics and the quality of
Boulder, Colorado: A progress Angeles, California. Science watersheds. Zeit. f. Geomorph.
report. Environmental Geol. 181:653-55. 24:263-72.
1:75-88.
533 Bibliography
Baulig, H. 1957. Peneplains and Benedict, J. B. 1970. Downslope soil Birkeland, P. 1 974. Pedology,
pediplains. Geol. Soc. America movement Colorado alpine
in a weathering, and
Bull. 68:913-30. region. Rates, processes, and geomorphological research.
Beaty, C. B. 1963. Origin of alluvial climatic significance. Arc. Alp. London: Oxford Univ. Press.
fans, White Mountains, California Res. 2:165-226. . 1984. So/75 and
and Nevada. Ann. Assoc. Am. . 1976. Frost creep and geomorphology. New York:
Geog. 53:516-35. gelifluction features. A review. Oxford Univ. Press.
.1970. Age and estimated Quat. Res. 6:55-77. Black, R. F. 1954. Permafrost, a
rate of accumulation of an alluvial Benson, M. A. 1950. Use of review. Geol. Soc. America Bull.
fan, White Mountains, California, historical data in flood-frequency 65:839-55.
U.S.A. Am. Jour. Sci. 268:50-77. analysis. Am. Geophys. Union . 1963. Les coins de glace et
1975a. Sublimation or Trans. 31:419-24. le gelpermanent dans le Nord de
—
.
melting observations from the . 1971. Uniform flood- TAlaska. Annates Geog.
White Mountains, California and frequency estimating methods for 72:257-71.
Nevada, U.S.A. Jour. Glaciol. federal agencies. Water Resour. 1969. Climatically
14:275-86. Res. 4:891-908. significant fossil periglacial
1975b. Coulee alignment Berner, R. A., and Holdren, G. R., phenomena in north-central
and the wind in southern Alberta, Jr. 1977. Mechanism of feldspar United States. Biuletyn Perygl.
Canada. Geol. Soc. America Bull. weathering: Some observational 20:225-38.
86:119-28. evidence. Geology 5:369-72. 1974. Ice-wedge polygons of
Beaumont, P. 1972. Alluvial fans . 1979. Mechanisms of northern Alaska. In Glacial
along the foothills of the Elburz feldspar weathering — II. geomorphology, edited by D. R.
Mountain, Iran. Paleogeogr. Observations of feldspars from Coates, pp. 247-75. S.U.N.Y,
Paleoclimatol. Paleoecol. soils. Geochim et Cosmochim Binghamton: Pubs, in
12:251-73. Acta 43:1 173-86. Geomorphology, 5th Ann.
Begin, Z. B. 1981. Stream curvature Berner, R. A., and Morse, J. 1974. Symposium.
and bank erosion: A model based Dissolution kinetics of calcium 1976. Periglacial features
on the momentum equation. Jour. carbonate in seawater — IV: indicative of permafrost: Ice and
Geology 89:497-504. Theory of calcite dissolution. Am. soil wedges. Quat. Res. 6:3-26.
Begin, Z. B.; Meyer, D. F.; and Jour. Sci. 274:108-34. Blackwelder, E. 1927. Fire as an
Schumm, S. A. 1980. Knickpoint Beskow, G. 1947. Soil freezing and agent in rock weathering. Jour.
migration due to base-level frost heaving with specific Geology 35:135-40.
lowering. Am. Soc. Civil application to roads and railroads. . 1931. The lowering of playas
Engineers, Jour. Water, Port, Evanston, 111.: Northwestern by deflation. Am. Jour. Sci.
Coastal and Ocean Div. University Tech. Inst. 21:140-44.
106:369-87. Bhowmik, N. G. 1982. Shear stress -. 1934. Yardangs. Geol. Soc.
. 1981. Development of distribution and secondary America Bull. 45:159-66.

longitudinal profiles of alluvial currents in straight open channels. Blagborough, J. W., and Farkas,
channels response to base-level
in In Gravel-bed rivers, edited by R. Rock glaciers in the
S. E. 1968.
lowering. Earth Surf. Proc. and Hey, J. Bathurst, and C. Thorne, San Mateo Mountains, south-
Landforms 6:49-68. pp. 31-62. New York: John Wiley central New Mexico. Am. Jour.
Behrendt, C. 1965. Densification
J. & Sons. Sci. 266:812-23.
of snow on the ice sheet of Bhowmik, N. G., and Demissie, M. Blatt, H., and Jones, R. 1975.
Ellsworth Land and South 1982. Carrying capacity of flood Proportions of exposed igneous,
Antarctic Peninsula. Jour. plains. Am. Soc. Civil Engineers metamorphic, and sedimentary
Glaciol. 5:451-60. Proc, Jour. Hydraulics Div. 108, rocks. Geol. Soc. America Bull.
Belly, P. V. 1964. Sand movement by HY3:443-52. 86:1085-88.
wind. U.S. Army Corps Engrs., Bindschadler, R. 1983. The Blissenbach, E. 1954. Geology of
Coastal Eng. Res. Center Tech. importance of pressurized alluvial fans in semiarid regions.
Memo 1. subglacial water in separation and Geol. Soc. America Bull.
Bender, M. L.; Fairbanks, R. G.; sliding at the glacier bed. Jour. 65:175-90.
Taylor, F. W.; Matthews, R. K.; Glaciol. 29:3-19. Bloom, A. L. 1965. The explanatory
Goddard, J. G.; and Broecker, Bird, E. C. F. 1969. Coasts. description of coasts. Zeit. f.
W. S. 1979. Uranium-series Cambridge, Mass.: M.I.T. Press. Geomorph. 9(4):422-36.
dating of Pleistocene reef tracts of . 1967. Pleistocene shorelines:
Barbados, West Indies. Geol. Soc. A new test of isostasy. Geol. Soc.
America Bull. 90:577-94. America Bull. 78:1477-94.
—
Bibliography 534
. 1980. Late Quaternary sea . 1968. Flow tills and related Bowen, A. J., and Inman, D. L. 1969.
level change on South Pacific deposits on some Vestspitsbergen Rip currents, 2. Laboratory and
coasts: A study in the tectonic glaciers. Jour. Glaciol. field observations. Jour. Geophys.
diversity. In Earth rheology, 7:391-412. Research 74:5479-90.
isostasy, and eustasy, edited by 1970a. On the origin and .1971. Edge waves and
N. A. Morner, pp. 505-16. New transport of englacial debris in crescentic bars. Jour. Geophys.
York: John Wiley & Sons. Svalbard glaciers. Jour. Glaciol. Research 76:8662-71.
Bloom, A. L.; Broecker, W. S.; 9:213-29. Bowen, N. 1928. The evolution of
Chappell, J. M. A.; Matthews, 1970b. On the deposition of the igneous rocks. Princeton,
R. K.; and Mesolella, K. J. 1974. subglacial and melt-out tills at the N.J.: Princeton Univ. Press.
Quaternary sea level fluctuations margins of certain Svalbard Bradley, W. C. 1957. Origin of
on a tectonic coast: New 230Th/ glaciers. Jour. Glaciol. 9:231-46. marine-terrace deposits in the
234
U dates from the Huron -. 1971. Till genesis and fabric Santa Cruz area, California. Geol.
Peninsula, New Guinea. Quat. in Svalbard, Spitsbergen. In Till: Soc. America Bull. 68:421-44.
Res. 4:185-205. A symposium, edited by R. P. 1958. Submarine abrasion
.
Bluck, B. J. 1964. Sedimentation on Goldthwait, pp. 41-72. Columbus: and wave-cut platforms. Geol.
an alluvial fan in southern Ohio State Univ. Press. Soc. America Bull. 69:961-1 A.
Nevada. Jour. Sed. Petrology 1972a. The role of thermal
.
1970. Effect of weathering
34:395-400. regime in glacial sedimentation on abrasion of granitic gravel,
Bogli, A. 1964. Mischungskor- a general theory. In Polar Colorado River (Texas). Geol.
rosion —
ein Beitrag zur geomorphology, edited by R. J. Soc. America Bull. 81:61-80.
Verkarstungsproblem. Erdkunde Price and D. E. Sugden, pp. 1-19. Bradley, W. C, and Griggs, G. B.
18:83-92. Inst. Brit. Geog. Spec. Paper 4. 1976. Form, genesis, and
. 1980. Karst hydrology and 1972b. Modern Arctic
. deformation of central California
physical speleology. Berlin: glaciers as depositional models for wave-cut platforms. Geol. Soc.
Springer- Verlag. former ice sheets. Jour. Geol. Soc. America Bull. 87:433-49.
Boothroyd, J. C, and Ashley, G. M. Lond. 128:361-93. Bradley, W. C; Hutton, J. T; and
1975. Processes, bar morphology, -. 1974. Processes and patterns Twidale, C. R. 1978. Role of salts
and sedimentary structures on of glacial erosion. In Glacial in development of granitic tafoni,
braided outwash fans, geomorphology, edited by D. R. south Australia. Jour. Geology
northeastern Gulf of Alaska. In Coates, pp. 41-87. S.U.N.Y., 86:647-54.
Glaciofluvial and glaciolacustrine Binghamton: Pubs, in Bradley, W. C, and Mears, A. I.
sedimentation, edited by Geomorphology, 5th Ann. 1980. Calculations of floods
B. McDonald and A. Jopling, Symposium. needed to transport coarse
pp. 193-222. Tulsa, Okla.: SEPM 1976. The origin of glacially fraction of Boulder Creek
Spec. Pub. 23. fluted surfaces — observation and alluvium at Boulder, Colorado.
Born, S. M. 1972. Late Quaternary theory. Jour. Glaciol. 17:287-309. Geol. Soc. America Bull. 91 (pt.
history, deltaic sedimentation, 1978. Boulder shapes and l):135-38.
and mud-lump formation at grain-size distribution of debris as Brakenridge, G. R. 1981. Late
Pyramid Lake, Nevada. Univ. indication of transport paths Quaternary floodplain
Nevada, Reno, Desert Research through a glacier and till genesis. sedimentation along the Pomme
Inst. Sedimentology 25:773-99. de Terre River, southern Missouri.
Born, S. M., and Ritter, D. F. 1970. 1979. Processes of glacier Quat. Res. 15:62-76.
Modern terrace development near erosion on different substrata. Brenninkmeyer, B. M. 1975.
Pyramid Lake, Nevada, and its Jour. Glaciol. 23:15-36. Frequency of sand movement in
geologic implications. Geol. Soc. Boulton, G. and Jones, A. S.
S., the surf zone. Proc. Nth Conf. on
America Bull. 81:1233-42. 1979. Stability of temperate ice Coast. Eng., 812-27.
Bottomley, J. T. 1872. Melting and caps and ice sheets resting on beds Bretz, J. H. 1942. Vadose and
regelation of ice. Nature 5:185. of deformable sediment. Jour. phreatic features of limestone
Boulton, G.S. 1967. The Glaciol. 24:29-43. caverns. Jour. Geologv
development of a complex Bowen, A. J. 1969. The generation of 50:675-811.
supraglacial moraine at the longshore currents on a plane Brice, J. C. 1964. Channel patterns
margin of Sorbreen, Ny Friesland, beach. J. Marine Res. 37:206-15. and terraces of the Loup River in
Vestspitsbergen. Jour. Glaciol. . 1973. Edge waves and the Nebraska. U.S. Geol. Survey
6:717-36. litoral environment. Proc. 13th Prof. Paper 422-D.
Conf. on Coast. Eng., . meander
1974. Evolution of
pp. 1313-20. loops. Geol. Soc. America Bull.
85:581-86.
—
535 Bibliography
Brook, G.; Folkoff, M.; and Box, E. Bryan, R. B. 1979. The influence of 1975b. Landforms that do
1983. A world model of soil slope angle on soil entertainment not tend toward a steady state. In
carbon dioxide. Earth Surf. Proc. by sheetwash and rainsplash. Theories of I andform
and Landforms 8:79-88. Earth Surf. Proc. and Landforms development, edited by W. N.
Brook, G., and Ford, D. 1982. 4:43-58. Melhorn and R. C. Flemal,
Hydrologic and geologic control of Buckman, H. O., and Brady, N. C. pp. 111-28. S.U.N. Y,
carbonate water in the subarctic 1960. The nature and properties Binghamton: Pubs, in
Nahanni karst, Canada. Earth of soils. New York: Macmillan. Geomorphology, 6th Ann. Mtg.
Surf. Proc. and Landforms Buckman, R. C, and Anderson, . 1979. Threshold of critical
7:1-16. R. E. 1979. Estimation of fault- power in streams. Geol. Soc.
Brookfield, M. 1970. Dune trend and scarp ages from a scarp-height- America Bull. 90 (pt. l):453-64.
wind regime in central Australia. slope-angle relationship. Geology -. 1980. Geomorphic
Zeit.f. Geomorph., Suppl. 7:11-14. thresholds as defined by ratios. In

10:121-58. Budd, W. F. 1975. A first simple Thresholds in geomorphology,
Brown, R. J. E. 1967. Permafrost in model for periodically self-surging edited by D. Coates and J. Vitek,
Canada. Canada Geol. Survey glaciers. Jour. Glaciol. 14:3-21. pp. 259-63. London: Allen and
Map 1246A, Natl. Res. Council Geomorphology
Biidel, J. 1968. Unwin Ltd.
Publ. NRC 9769. Encyclopedia of
principles. In . 1984. Tectonic
. 1969. Permafrost in Canada. geomorphology, edited by R. W. geomorphology. Jour. Geol. Educ.
Canada Geol. Survey Map Fairbridge, pp. 416-22. New 32:310-24.
1246A. York: Reinhold Book Corp. Bull, W. B., and McFadden, L. D.
1 970. Permafrost in 1982. Climatic
. 1977. Tectonic geomorphology
Canada. Toronto: Univ. Toronto geomorphology. Princeton, N.J.: north and south of the Garlock
Press. Princeton Univ. Press. Fault, California. In
Browning, M. 1973. Catastrophic
J. Budyko, M. I. 1977. Climatic Geomorphology in arid regions,
rock slides, Mount Huascaran, changes. Baltimore: Waverly edited by D. O. Doehring,
north-central Peru, May 31, 1970. Press, p. 261. pp. 115-38.S.U.N. Y.
Am. Assoc. Petroleum Geologists Bull, W. B. 1963. Alluvial-fan Binghamton: Proc. 8th Ann.
Bull. 57:1335-41. deposits in western Fresno Geomorph. Symposium.
Brucker, R. W.; Hess, J. W.; and County, California. Jour. Geology Bullard, F. 1962. Volcanoes. Austin:
White, W. B. 1972. Role of 71:243-51. Univ. Texas Press.
vertical shafts in the movement of . 1964a. Alluvial fans and Buol, S. W.; Hole, F. D.; and
ground water in carbonate near-surface subsidence in McCracken, R. J. 1973. Soil
aquifers. Ground Water 10(6):5. western Fresno County, genesisand classification. Ames,
Brunsden, D., and Kesel, R. H. 1973. California. U.S. Geol. Survey Iowa: Iowa State Univ. Press.
Slope development on a Prof. Paper 437-A. Burdon, D. J., and Safadi, C. 1963.
Mississippi River bluff in historic 964b. Geomorphology of
1 Ras-el-ain: The great karst spring
time. Jour. Geology 81:576-97. segmented alluvial fans in western of Mesopotamia. J. Hydrol.
Brush, L. M. 1961. Drainage basins, Fresno County, California. U.S. 1:58-95.
channels, and flow characteristics Geol. Survey Prof. Paper 552-F. Burke, R. M., and Birkeland, P. W.
of selected streams in central . 1968. Alluvial fans. Jour. 1979. Reevaluation of
Pennsylvania. U.S. Geol. Survey Geol. Educ. 16:101-6. multiparameter relative dating
Prof. Paper 282-F. -. 1974. Geomorphic tectonic techniques and their application
Brush, L. M., and Wolman, M. G. analysis of the Vidal region. In to the glacial sequence along the
1960. Knickpoint behavior in Woodward-McNeill and Assoc, eastern escarpment of the Sierra
noncohesive material: a laboratory Vidal nuclear gen. station, Units Nevada, California. Quat. Res.
study. Geol. Soc. America Bull. 1 and 2, appendix 2.5B (Geology 11:21-51.
71:57-76. and seismology). Rosemead, Burnett, A. W., and Schumm, S. A.
Bruun, P. 1963. Longshore currents Calif.: So. Calif.Co. 1983. Alluvial river response to
and longshore troughs. Jour. -. 1975a. Allometric change of neotectonic deformation in
Geophys. Research 68:1065-78. landforms. Geol. Soc. America Louisiana and Mississippi. Science
Bryan, K. 1922. Erosion and Bull. 86:1489-98. 222:49-50.
sedimentation in the Papago Butzer, K.W. 1964. Environment
County, Arizona with a sketch of and archeology. Chicago: Aldine
the geology. U.S. Geol. Survey Publishing.
Bull. 730-B: 19-90.
.
Bibliography 536
Cailleaux, A., and Tricart, J. 1956. Carter, L. D. 1983. Fossil sand . 1959. The shape of
Le probleme de la classification wedges on the Alaskan arctic drumlins. Jour. Glaciol.
des faits geomorphologiques. coastal plain and their 3:339-44.
Annates Geog. 65:162-86. paleoenvironmental significance. 1962. Geomorphology and
Caine, N. 1968. The block fields of In Proc. Permafrost 4th Internal. the general systems theory. U.S.
northeastern Tasmania. Confi, pp. 109-14. Natl. Acad. Geol. Survey Prof. Paper 500-B.
Australian Natl. Univ., Dept. Sci. . 1969a. Introduction to
Geog. Publ. G/6. Casagrande, A. 1948. Classification fluvial processes. London:
. 1981. A source of bias in and identification of soils. Am. Methuen and Co.
movement as
rates of surface soil Soc. Civil Engineers Trans. ,ed. 1969b. Water, earth, and
estimated from marked particles. 113:901-91. man. London: Methuen and Co.
Earth Surf. Proc. and Landforms CERC (Coastal Engineering . 1978. The hillslope
6:69-75. Research Center). 1973. Shore hydrological cycle. In Hillslope
1982. Toppling failures from protection manual. 3 vols. Hydrology, edited by M. J.
Alpine cliffs on Ben Lomond, Washington, D.C.: U.S. Army Kirkby, pp. 1-42. New York: John
Tasmania. Earth Surf. Proc. and Corps Engrs. Wiley & Sons.
Landforms 7:133-52. Chadbourne, B. D.; Cole, R. M.; Chorley, R. J., and Kennedy, B. A.
Calkin, P., and Cailleaux, A. 1962. A Tootill, S.; and Walford, M. E. R. 1971. Physical geography.
quantitative study of cavernous 1975. The movement of melting London: Prentice-Hall
weathering (tafonis) and its ice over rough surfaces. Jour. International.
application to glacial chronology Glaciol. 14:287-92. Chorley, R. J., and Morley, L. S. D.
in Victoria Valley, Antarctica. Chamberlain, R. T 1928. 1959. A simplified approximation
Zeit.f. Geomorph. 6:317-24. Instrumental work on the nature for the hypsometric integral. Jour.
Calkin, P. E., and Rutford, R. H. of glacial motion. Jour. Geology Geology 67:566-71.
1974. The sand dunes of Victoria 36:1-30. Chose, B.; Pandy, S.; and Lai, G.
Valley, Antarctica. Geogr. Rev. Chambers, M. J. G. 1970. 1967. Quantitative geomorphology
64:189-216. Investigations of patterned ground of the drainage basins in the
Callander, R. A. 1969. Instability at Signy Island, South Orkney central Lumi basin in western
and river channels. Jour. Fluid Islands, IV. Longterm Rajasthan. Zeit.f. Geomorph.
Mech. 36:465-80. experiments. British Antarctic 11:146-60.
Campbell, W. J., and Rasmussen, Surv. Bull. 23:93-100. Church, M. 1972. Baffin Island
L. A. 1969. Three-dimensional Chappell, J. M. A. 1974. Geology of sandurs: A study of arctic fluvial
surges and recoveries in a coral terraces, Huon Peninsula, processes. Canada Geol. Survey
numerical glacier model. Can. J. New Guinea: A study of Bull. 216.
Earth Sci. 6:979-86. Quaternary tectonic movements . 1978. Paleohydrological
Carlston, C. W. 1963. Drainage and sea level changes. Geol. Soc. reconstructions from a Holocene
density and streamflow. U.S. Geol. America Bull. 85:553-70. valley. In Fluvial sedimentology,
Survey Prof. Paper 422-C. Cheel, R. 1982. The depositional edited by A. Miall, pp. 743-72.
. 1969. Downstream history of an esker near Ottawa, Can. Soc. Petrol. Geol. Mem. 5.
variations in the hydraulic Canada. Can. J. Earth Sci. Church, M., and Jones, D. 1982.
geometry of streams: Special 19:1417-27. Channel bars in gravel-bed rivers.
emphasis on mean velocity. Am. Chepil, W. S. 1945. Dynamics of In Gravel-bed rivers, edited by
Jour. Sci. 267:499-509. wind erosion, II. Initiation of soil R. Hey, J. Bathurst, and
Carroll, D. 1958. Role of clay movement. Soil Sci. 60:397-41 1 C. Thorne, pp. 291-338. New
minerals in the transportation of — . 1959. Equilibrium of soil York: John Wiley & Sons.
iron. Geochim. et Cosmochim. grains at the threshold of Clapperton, C. M. 1975. The debris
Acta 14:1-27. movement by wind. Soil. Sci. Soc. content of surging glaciers in
. 1970. Rock weathering. Am. Proc. 23:422-28. Svalbard and Iceland. Jour.
New York: Plenum Press. Chepil, W. S., and Woodruff, N. P. Glaciol. 14:395-406.
Carson, M. A. 1969. Models of hill- 1963. The physics of wind erosion Clark, J. A., and Bloom. A. L. 1979.
slope development under mass and its control. Advances in Hvdroisostasy and Holocene
failure. Geographical Analysis Agron. 15:211-302. emergence of South America. In
1:76-100. Chorley, R. J. 1957. Illustrating the Proc. Internal. Symposium on
Carson, M. A., and Kirkby, M. 1972. laws of morphometry. Geol. Mag. Coastal Evolution in the
Hillslope form and process. 94:140-50. Quaternary, Sao Paulo. 1978,
London: Cambridge Univ. Press. edited by K. Suguio and others,
pp. 41-60.
537 Bibliography
Clark. S.P., and Jager, E. 1969. Colman, S. M., and Pierce, K. L. Corte, A. E. 1969. Geocryology and
Denudation rate in the Alps from 1981. Weathering rinds on engineering. In Reviews in
geochronological and heat flow andesitic and basaltic stones as a engineering geology 2, edited by
data. Am. Jour. Sci. Quaternary age indicator, western D. Varnes and G. Kiersch,
267:1143-60. United States. U.S. Geol. Survey pp. 19-85. Boulder, Colo.: Geol.
1
Clayton, L. 1967. Stagnant-glacial Prof. Paper 1210. Soc. America.

features of the Missouri Coteau in Colman, S. M., and Watson, K. Costa, J. E. 1974a. Stratigraphic,
North Dakota. North Dakota 1983. Ages estimated from a morphologic, and pedologic
Geol. Survey Misc. Series diffusion model for scarp evidence of large floods in humid
30:25-46. degradation. Science 221:263-65. environments. Geology 2:301-3.
Cleaves, E.; Fisher. D.: and Bricker, Connell, W., and Patrick, W. 1968. 1974b. Response and
.
O. 1974. Chemical weathering of Sulfate reduction in soil. Effects recovery of a piedmont watershed
serpentinite in the eastern of redox potential and pH. from tropical storm Agnes, June
piedmont of Maryland. Geol. Soc. Science 159:86-87. 1972. Water Resour. Res.
America Bull. 85:437-44. Cook, J. H. 1946. Kame complexes 10:106-12.
Coates, D. R. 1976. Geomorphology and perforation deposits. Am. . 1978. Holocene stratigraphy
and engineering. Stroudsburg, Jour. Sci. 244:573-83. in floodfrequency analysis. Water
Pa.: Dowden, Hutchinson and Cooke, H. J. 1973. Tropical karst in Resour. Res. 14:626-32.
Ross. northeast Tanzania. Zeit. f.
-. 1983. Paleohydraulic
Coates. D. R.. and Vitek, J. D., eds. Geomorph. 17:443-59. reconstruction of flash-flood peaks
1980. Thresholds in Cooke. R. U. 1970. Morphometric from boulder deposits in the
geomorphology. London: Allen analysis of pediments and Colorado Front Range. Geol. Soc.
and Lnwin Ltd. associated landforms in the America Bull. 94:986-1004.
Colbeck, S. C, ed. 1980. Dynamics western Mojave Desert, Costa, J. E., and Baker, V. R. 1981.
of snow and ice masses. New California. Am. Jour. Sci. Surficial geology —
Building with
York: Academic Press. 269:26-38. the Earth. New York: John Wiley
Colbeck, S. C, and Evans, R. J. Cooke, R. U., and Doornkamp, J. & Sons.
1973. A flow law for temperate 1974. Geomorphology in Cox, E. T. 1874. Fifth annual report
glacier ice. Jour. Glaciol. environmental management. of the Geological Survey of
12:71-86. London: Clarendon Press. Indiana, pp. 280-305.
Colby, B. 1963. Fluvial sediments: A Cooke, R. U., and Reeves, R. W. Indianapolis: Indiana Geol.
summary of source, 1972. Relations between debris Survey.
transportation, deposition, and size and the slope of mountain Cramer, H. 1941. Die Systematik der
measurement of sediment fronts and pediments in the Karstdolinen. Neues Jb. Miner.
discharge. U.S. Geol. Survey Bull. Mojave Desert, California. Zeit.f Geol. Paldont. 85:293-382.
1181-A:21. Geomorph. 16:76-82. Crary, A. P. 1966. Mechanisms for
. 1964. Scour and sandfill in Cooke, R. U., and Warren, A. 1973. fiord formation indicated by
bed streams. U.S. Geol. Survey Geomorphology in deserts. studies of an ice-covered inlet.
Prof. Paper 462-D. London: Batsford Ltd. Geol. Soc. America Bull.
Coleman, J. M. 1968. Deltaic Cooley, R.; Fiero, G.; Lattman, L.; 77:911-30.
evolution. In Encyclopedia of and Mindling, A. 1973. Influence Crittenden, M. 1963. New data on
geomorphology, edited by R. W. of surface and near-surface the isostatic deformation of Lake
Fairbridge. pp. 255-60. New caliche distribution on infiltration Bonneville. U.S. Geol. Survey-
York: Reinhold Book Corp. characteristics and flooding, Las Prof. Paper 454-E.
Colman. S. M. 1982. Clay Vegas area, Nevada. Univ. Cronin, E. 1983. Design and
J.
mineralogy of weathering rinds Nevada, Reno, Desert Research performance of a liquid natural
and possible implications Inst. Proj. Rept. 21. convection subgrade cooling
concerning the sources of clay Cooper, W. S. 1958. Coastal sand system for construction on ice-rich
minerals in soils. Geology dunes of Oregon and Washington. permafrost. In Proc. Permafrost
10:370-75. Geol. Soc. America Mem. 72. 4th Internat. Conf, pp. 198-203.
.1983. Progressive changes in Corbel, J. 1957. Karsts hauts-Alpins. Natl. Acad. Sci.
the morphology of fluvial terraces Rev. Geogr. Lyon 32:135-58. Cronin, T. M. 1981. Rates and
and scarps along the . 1959. Vitesse de l'erosion. possible causes of neotectonic
Rappahannock River, Virginia. Zeit.f. Geomorph. 3:1-28. movements of the
vertical crustal
Earth Surf. Proc. and Landforms 1 964. L'erosion terrestre, emerged southeastern United
8:201-12. etude quantitative. Annates Ge'og. States Atlantic coastal plain,
73:385-412. Geol. Soc. America Bull.
92:812-33.
Bibliography 538
Cronin, T. M.; Szabo. B.: Ager, Dalrymple. G.; Silver. E.; and Dean. R. G.. and Perlin. M. 1977.
T. A.; Hazel, and Owens,
J. B.; Jackson, E. 1973. Origin of the Coastal engineering study of
J. P. 1981. Quaternary climates Hawaiian Islands. Am. Scientist Ocean City Inlet, Maryland. In
and sea level: US Atlantic coastal 61:294-308. Coastal sediments 77. Am. Soc.
plain.Science 21 1:233-40. Dalrymple. J. B.; Blong, R. J.; and Civil Engineers. Symposium
Crosby. W. O. 1902. Origin of eskers. Conacher. A.J. 1968. A Water, Port, Coastal and Ocean
Am. Geologist 30:1-38. hypothetical nine-unit landsurface Div. 5:520-42.
Crowley, K. D. 1983. Large-scale bed model. Zeit.f. Geomorph. Deere. D. U.. and Peck. R. B. 1959.
configurations (macroforms), 12:60-76. Stability of cuts in fine sands and
Platte River basin. Colorado and Dalrymple, T. 1960. Flood frequency varved clays. Northern Pacific
Nebraska: Primary structures and analysis. Manual of hydrology. Railroad. Noxon Rapids line
formative processes. Geol. Soc. part 3. Flood flow techniques. change. Montana. Proc. AREA
America Bull. 94:117-33. U.S. Geol. Survey Water Supply 59. pp. 807-15.
Crozier, M. J. 1973. Techniques for Paper 1543-A. Denny. C. S. 1956. Surficial geology
the morphometric analysis of Dalrymple. T., and Benson. M. A. and geomorphology of Potter
landslips. Zeit. f. Geomorph. 1967. Measurement of peak County. Pennsylvania. U.S. Geol.
17:78-101. discharge by the slope-area Survey Prof. Paper 288.
Cummans, J.1981. Mudflows method. U.S. Geol. Suney . 1965. Alluvial fans in the
resulting from the May 18. 1980, Techniques Water Resour. Res. Death Valley region. California
eruption of Mount St. Helens, Div.. Bk. 3, chap. A-2. and Nevada. U.S. Geol. Sune>
Washington. U.S. Geol. Survey Daly, R. 1933. Igneous rock and the Prof. Paper 466.
Circ. 850-B. depths of the Earth. New York: 1967. Fans and pediments.
Cunningham, F., and Griba, W. McGraw-Hill. Am. Jour.Sci. 265:81-105.
1973. A model of slope Daniel, J. R. K. 1981. Drainage Derbyshire. E. 1976. Geomorphology
development, and its applications density as an index of climatic and climate. New York: John
to the Grand Canyon. Arizona. geomorphology. J. Hydro/. Wiley & Sons.
Zeit.f. Geomorph. 17:43-77. 50:147-54. Dietrich. W. E.. and Dunne. T. 1978.
Curray, J. R. 1960. Sediments and Davies. J. L. 1973. Geographical Sediment budget for a small
history of Holocene transgression, variation in coastal development. catchment in mountainous terrain.
continental shelf, northwest Gulf New York: Hafner. Zeit.f. Geomorph.. Suppl. Bd.
of Mexico. In Recent sediments, Davies, W. E. 1960. Origin of caves 29:191-206.
northwest gulf of Mexico, edited in folded limestone. Natl. Speleol. Dingman. S. L. 1978. Drainage
by Shepard, F. P., et al. Amer. Soc. Bull. 22:5-18. density and streamflow: A closer
Assoc. Petroleum Geologists, Davis. D. G. 1980. Cave development look. Water Resour. Res.
pp. 221-66. in the Guadalupe Mountains, a 14:1183-87.
1961. Late Quaternary sea
. critical review of recent Dionne. J. C. 1974. Polished and
level; a discussion. Geol. Soc. hypotheses. Natl. Speleol. Soc. striated mud surfaces in the St.
America Bull. 72:1707-12. Bull. 42:42-48. Lawrence tidal flats. Quebec.
. 1965. Late Quaternary Davis, J. A morphogenetic
L. 1964. Can. J. Earth Sci. 1:860-66. 1
history, continental shelves of the approach world shorelines.

to Dolan. R.. and Ferm. J. C. 1968.
United States. In The Quaternary Zeit.f. Geomorph. 8:127-42. Crescentic landforms along the
of the United States, edited by Davis, R. A., ed. 1978. Coastal mid-Atlantic coast. Science
H. Wright and D. Frey, sedimentary environments. New 159:627-29.
pp. 723-35. Princeton^ N.J : York: Springer- Verlag. Dolan. R.; Hayden. B.: and Felder.
Princeton Univ. Press. Davis. W. M. 1902. River terraces in W. 1979. Shoreline periodicities
Cvijic, J. 1893. Das Karstphanomen. New England. Mus. Comp. and edge waves. Jour. Geology
Geogr. Abh. 5:217-329. Zoology Bull. 38:77-111. 87:175-85.
Czudek. T.. and Demek. J. 1970. . 1930. Origin of limestone Dolan. R.; Hayden. B.; and Lins. H
Thermokarst in Siberia and its caverns. Geol. Soc. America Bull. 1980. Barrier islands. Am.
influence on the development of 41:475-628. Scientist 68:16-25
lowland relief. Quat. Res. Day. M. 1978. Morphology and Dolan. R.: Vincent. L.: and Hayden.
1:103-20. distribution of residual limestone B. 1974. Crescentic coastal
Dahl. R. 1965. Plastically sculptured hills(mogotes) in the karst of landforms. Zeit. f. Geomorph.
detail forms on rock surfaces in northern Puerto Rico. Geol. Soc. 18:1-12.
northern Nordland. Geogr. Annlr America Bull. 89:426-32.
47:83-140.
539 Bibliography
Doornkamp, J. C, and King, Drever, J. I., and Smith, C. L. 1978. -. 1965. Theoretical
C. A. M. 1971. Numerical Cyclic wetting and drying of the implications of underfit streams.
analysis in geomorphology: An soil zone as an influence on the U.S. Geol. Survey Prof. Paper
Introduction. London: Edward chemistry of groundwater in arid 452-C.
Arnold Ltd. terrains. Am. Jour. Sci. 1966a. Duricrusted residuals
.
Douglas, G. R.; McGreevey, J. P.; 278:1448-54. on the Barrier and Cobar

and Whalley, W. B. 1983. Rock Duane, D. B., and James, W. R. pediplains of New South Wales.
weathering by frost shattering 1980. Littoral transport in the Jour. Geol. Soc. Australia
processes. In Proc. Permafrost 4th surf zone elucidated by an 13:299-307.
Internal. Conf., pp. 244-48. Natl. Eulerian sediment tracer 1966b. Pediment slope and
.
Acad. Sci. experiment. Jour. Sed. Petrology particle size at Middle Pinnacle,
Douglas, I. 1964. Intensity and 50:929-42. near Broken Hill, New South
periodicity in denudation process DuBois, R. N. 1972. Inverse relation Wales. Austr. Geog. Studies
with special reference to the between foreshore slope and mean 4:1-17.
removal of material in solution by grain size as a function of the . 1969. Relation of
rivers. Zeit. f. Geomorph. heavy mineral content. Geol. Soc. morphometry to runoff frequency.
8:453-73. America Bull. 83:871-76. In Introduction to fluvial
1967. Man, vegetation and
. 1978. Beach topography and
. processes, edited by R. J. Chorley,
sediment yields of rivers. Nature beach cusps. Geol. Soc. America pp. 177-88. London: Methuen
215:925-28. Bull. 89:1133-39. and Co.
Drake, J. 1980. The effect of soil 1981. Foreshore topography, -. 1973. Magnitude-frequency
activity on the chemistry of tides and beach cusps, Delaware. analysis and channel morphology.
carbonate groundwater. Water Geol. Soc. America Bull. 92 (pt. In Fluvial geomorphology, edited
Resour. Res. 16:381-86. l):132-38. by M. Morisawa, pp. 91-121.
Drake, J. J., and Harmon, R. S. Dunkerley, D. L. 1980. The study of S.U.N.Y., Binghamton: Pubs, in
1973. Hydrochemical the evolution of slope form over Geomorphology.
environments of carbonate long periods of time: A review of Duval, P. 1981. Creep and fabrics of
terrains. Water Resour. Res. methodologies and some new polycrystalline ice under shear
9:949-57. observational data from Papua, and compression. Jour. Glaciol.
Drake, J. J., and Wigley, T. M. L. New Guinea. Zeit.f. Geomorph. 27:129-40.
1975. The effect of climate on the 24:52-67. Duval, P., and Hughes, L. G. 1980.
chemistry of carbonate Dunne, T. 1978. Field studies of Does the permanent creep-rate of
groundwater. Water Resour. Res. hillslope flow processes. In polycrystalline ice increase with
11:958-62. Hillslope hydrology, edited by crystal size? Jour. Glaciol.
Drake, L. D. 1968. Till studies in M. J. Kirby, pp. 227-94. New 25:151-57.
New Hampshire. Ph.D. York: John Wiley & Sons. Dylik, J. 1964. The essentials of the
dissertation, Ohio State . 1979. Sediment yield and meaning of the term "Periglacial."
University. land use in tropical catchments. Soc. Sci. et Lettres Ltfdz Bull.
. 1974. Till fabric control by J. Hydrol. 42:281-300. 15:2:1-19.
clast shape. Geol. Soc. America Dunne, T, and Black, R. D. 1970a. Eakin, H.M. 1916. The Yukon-
Bull. 85:247-50. An experimental investigation of Koyukuk region. Alaska. U.S.
Drake, L. D., and Shreve, R. L. runoff production in permeable Geol. Survey Bull. 631.
1973. Pressure melting and soils. Water Resour. Res. Eaton, J., and Murata, K. 1960. How
regelation of ice by round wires. 6:478-90. volcanoes grow. Science
Proc. Royal Soc. London, ser. 1970b. Partial area
. 132:925-38.
A:332:51-83. contribution to storm runoff in a Ede, D. P. 1975. Limestone drainage
Dreimanis, A., and Vagners, U. J. small New England watershed. systems. /. Hydrol. 27:297-318.
1971. Bimodal distribution of rock Water Resour. Res. 6:1 296- 1311. Egemeir, S. J. 1981. Cavern
and mineral fragments in basal Dunne, T, and Leopold, L. B. 1978. development by thermal waters.
till. In Till: A symposium, edited Water in environmental planning. Natl. Speleol. Soc. Bull.
by R. P. Goldthwait. Columbus: San Francisco: W. H. Freeman. 43:31-52.
Ohio State Univ. Press. Dury, G. H. 1964. Principles of Eggler, D. H.; Larson, E. E.; and
underfit streams. U.S. Geol. Bradley, W. C. 1969. Granites,
Survey Prof. Paper 452-A. grusses, and the Sherman erosion
surface, southern Laramie Range,
Colorado- Wyoming. Am. Jour.
Sci.267:510-22.
—
Bibliography 540
Einstein, H. A. 1950. The bedload Evans, 1. S. 1969. Salt crystallization . 1969. Morphology of the
function for sediment and rock weathering: A review. Slims River. In Icefield Ranges
transportation in open channel Rev. geomorph. dynamique Research Project, Scientific
flows. U.S. Dept. Agri. Tech. Bull. 19:157-77. Results §1, edited by V. C.
1026. Evans, O. F. 1938. Classification and Bushell and R. H. Ragle,
El-Baz. and Maxwell, T. A., eds.
F.. origin of beach cusps. Jour. pp. 161-72. Am. Geog. Soc. and
1982. Desert landforms of Geology 46:615-27. Arctic Inst. N. America.
southwest Egypt: A basis for — . 1940. The low and ball of Fairbridge, R. W., and Newman. W.
comparison with Mars. NASA the east shore of Lake Michigan. 1968. Postglacial crustal
Sci. and Tech. Information Jour. Geology 48:476-51 1. subsidence of the New York area.
Branch, Washington, D.C., -. 1942. The origin of spits, Zeit.fi Geomorph. 12:296-317.
CR-3611. bars,and related structures. Jour. Fairchild, J.C. 1973. Longshore
Elliston,G. R. 1963. Catastrophic Geology 50:846-63. transport of suspended sediment.
glacier advances. Int. Assoc. Sci. Evans, R. 1976. Observations on a Proc. 1 3th Confi on Coast. Eng.,
Hvdrol. Bull. 8:65-66. stripe pattern. Biuletyn Pervgl. 1069-88.
Elson, J. A. 1968. Washboard 25:9-22. Ferguson, R. I. 1975. Meander
moraines and other minor Everett, D. H. 1961. The irregularity and wavelength
moraine types. In Encyclopedia of thermodynamics of frost damage estimation. J. Hydrol. 26:315-33.
geomorphology. edited by R. W. to porous solids. Trans. Faraday Ferrians, O. J. 1965. Permafrost map
Fairbridge, pp. 1213-19. New Soc. 57:1541-51. of Alaska. U.S. Geol. Survey
York: Reinhold Book Corp. Everitt, B. L. 1968. Use of Misc. Geol. Lav. Map 1-445.
Embleton, C, and King, C. A. M. cottonwood in an investigation of Ferrians, O. Kachadoorian, R.;
J.:
1968. Glacial and periglacial recent history of a flood plain. and Greene. G. W. 1969.
geomorphology. Edinburgh: Am. Jour. Sci. 266:417-39. Permafrost and related
Edward Arnold Ltd. Eyles, N.,and Rogerson, R. J. 1978. engineering problems in Alaska.
.
1975a. Glacial Sedimentology of medial moraines U.S. Geol. Survey Prof. Paper
geomorphology. New York: in Berendon Glacier, British 678.
Halsted Press. Columbia. Canada: Implications Feth, J.; Robertson, C: and Polzer.
.
1975b. Periglacial for debris transport in a W. 1964. Sources of mineral
geomorphology. New York: glacierized basin. Geol. Soc. constituents in water from
Halsted Press. America Bull. 89:1688-93. granitic rocks. Sierra Nevada,
Emery, K. O., and Kuhn, G. G. 1982. Eynon, G., and Walker, R. G. 1974. California and Nevada. U.S.
Sea cliffs: Their processes, profiles Facies relationships in Pleistocene Geol. Survev Water Supplv Paper
and classifications. Geol. Soc. outwash gravels, southern 1535-1.
America Bull. 93:644-54. Ontario: A model for bar growth Finkel. H. J. 1959. The barchans of
Emmett, W. W. 1980. Afield in braided rivers. Sedimentology southern Peru. Jour. Geology
calibration of the sediment- 21:43-70. 67:614-47.
trapping characteristics of the Fahey, B. D. 1975. Nonsorted circle Finlayson, B. 1981. Field
Helley-Smith bedload samples. development in a Colorado alpine measurements of soil creep. Earth
U.S. Geol. Survey Prof. Paper location. Geogr. Annlr. and Landforms
Surf. Proc.
1139. 57A:1 53-64. 6:35-48.
Engelhardt, H. F.; Harrison, W. D.; . 1983. Frost action and Fisher, J. E. 1963. Two tunnels in
and Kamb, B. 1978. Basal sliding hydration as rock weathering cold ice at 4000 m on the
and conditions at the glacier bed mechanism on schist: A Breithorn. Jour. Glaciol.
as revealed by bore-hole laboratory study. Earth Surf. 4:513-20.
photography. Jour. Glaciol. Proc. and Landforms. 8:535-45. Fisher, J. J., and Simpson, E. J.
20:469-508. Fahnestock, R. K. 1961. Competence 1979. Washover and tidal
Ericksen, G. E.; Pflacker, G.; and of a glacial stream. U.S. Geol. sedimentation rates as
Fernandez, J. V. 1970. Survey Prof. Paper 424-B:21 1-13. environmental factors in
Preliminary report on the . 1963. Morphology and development of a transgressive
geological events associated with hydrology of a glacial stream barrier shoreline. In Barrier
the May 31, 1970, Peru White River, Mount Rainier, islands, edited by S. Leatherman.
earthquake. U.S. Geol. Survey Washington. U.S. Geol. Sur\e\ New York: Academic Press
Circ. 639. Prof. Paper 422-A.
541 Bibliography
Fisk, H. N. 1944. Geological Folk, R. L. 1971a. Longitudinal Friedkin, J. F. 1945. A laboratory

investigation of the alluvial valley dunes of the northwestern edge of study of the meandering of
of the lower Mississippi River. the Simpson Desert, Northern alluvial rivers. U.S. Army Corps
Vicksburg, Miss.: Mississippi Territory, Australia, 1. Engrs., U.S. Waterways Eng.
River Comm. Geomorphology and grain size Exp. Sta.
. 1951. Loess and Quaternary relationships. Sedimentology Friese-Greene, T. W., and Pert. G. J.
geology of the lower Mississippi 16:5-54. 1965. Velocity fluctuations of

Valley. Jour. Geology 59:333-56. . 1971b. Genesis of Berksackerbrae, east Greenland.
Fiske, R. S.;Hopson, C. A.; and longitudinal and oghurd dunes Jour. Glaciol. 5:739-47.
Waters, A. C. 1963. Geology of elucidated by rolling upon grease. Fryberger, S. G. 1979. Dune forms
Mount Ranier National Park. Geol. Soc. America Bull. and wind regime. In A study of
U.S. Geol. Survey Prof. Paper 82:3461-68. global sand seas, edited by
444. 1975. Glacial deposits E. McKee. U.S. Geol. Survey
Messungen von
Fitze, P. 1971. identified by chattermark trails in Prof. Paper 1052:137-70.
Bodenbewegungen auf West- detrital garnets. Geology Fryberger, S. G., and Ahlbrandt,
Spitzbergen. Geog. Helvetica 3:473-75. T. S. 1979.Mechanisms for the
26:148-52. Folk, R. L., and Patton, E. B. 1982. formation of eolian sand seas.
Flemal, R. C. 1976. Pingos and pingo Buttressed expansion of granite Zeit.f. Geomorph. 23:440-60.
scars: Their characteristics, and development of grus in Frye,J. C, and Leonard, A. R. 1954.
distribution, and utility in central Texas. Zeit.f geomorph. Some problems of alluvial terrace
reconstructing former permafrost 26:17-32. mapping. Am. Jour. Sci.
environments. Quat. Res. Foose, R. M. 1967. Sinkhole 252:242-51.
6:37-53. formation by groundwater M. L. 1914. The geology of
Fuller,
Flemal, R. C; Hinkley, K. C; and withdrawal: Far West Rand, Long Island, New York. U.S.
Hesler, J. L. 1976. DeKalb South Africa. Science Geol. Survey Prof. Paper 82.
mounds: A possible Pleistocene 157:3792:1045-48. Galvin, C. J. 1967. Longshore
(Woodfordian) pingo field in Forbes, J. D. 1843. Travels through current velocity: A review of
north-central Illinois. Geol. Soc. the Alps of Savoy. Edinburgh: theory and data. Revs, in
America Mem. 136:229-50. Oliver and Boyd. Geophys. 5:3:287-304.
Fleming, R. W., and Johnson, A. M. Ford, D., and Ewers, R. 1978. The . 1968. Breaker type
1975. Rates of seasonal creep of development of limestone cave classification on three laboratory
silty clay soil. Quart. Jour. Engr. systems in length and depth. Can. beaches. Jour. Geophys. Research
Geol. 8:1-29. J. Earth Sci. 15:1783-98. 73:3651-59.
Flint, R. F. 1963. Altitude, lithology, Fournier, M. F. 1960. Climat et Galvin, C. J., and Eagleson, P. S.
and the Fall Zone in Connecticut. erosion. Paris: Presses Univ. 1965. Experimental study of
Jour. Geology 71:683-97. France. longshore currents on a plane
. 1971. Glacial and Francis, J. R. D. 1973. Experiments beach. U.S. Army Corps Engrs.,
Quaternary geology. New York: on the motion of solitary grains Coastal Eng. Res. Center Tech.
John Wiley & Sons. along the bed of a water stream. Memo 10.
Flint, R. F., and Bond, G. 1968. Proc. Royal Soc. London, ser. Gamper, M. W. 1983. Control and
Pleistocene sand ridges and pans A:332:443-71. rates of movement of solifluction
inwestern Rhodesia. Geol. Soc. Frazee, C. J.; Fehrenbacher, J. B.; lobes in the eastern Swiss Alps. In
America Bull. 79:299-314. and Krumbein, W. C. 1970. Loess Proc. Permafrost 4th Internat.
Foley, M. G. 1978. Scour and fill in distribution from a source. Soil Conf, pp. 328-33. Natl. Acad.
steep, sand-bed ephemeral Sci. Soc. Am. Proc. 34:296-301. Sci.
streams. Geol. Soc. America Bull. Free, G. R. 1960. Erosion Gams, I. 1969. Some morphological
89:559-70. characteristics of rainfall. Agri. characteristics of the Dinaric
1980a. Bed-rock incision by
. Engineering 41:447-49, 455. karst. Geogr. Jour. 135:563-72.
streams. Geol. Soc. America Bull. Freeze, R. A. 1980. A stochastic- Gardner, J. H. 1935. Origin and
91 (pt. 2):2189-213. conceptual analysis of rainfall- development of limestone caverns.
1980b. Quaternary diversion runoff processes on a hillslope. Geol. Soc. America Bull.
and incision, Dearborn River, Water Resour. Res. 16:391-408. 46:1255-74.
Montana. Geol. Soc. America French, H. M. 1974. Mass-wasting Gardner, J. S. 1979. The movement
Bull. 91 (pt. 2):2152-88. at Sachs Harbour, Barks Island, of material on debris slopes in the
N.W.T., Canada. Arc. Alp. Res. Canadian Rocky Mountains. Zeit.
6:77-78. f Geomorph. 23:45-57.
—
Bibliography 542
Gardner, T. W. 1973. A model study and

Gile, L. FL; Peterson, F.; Gold, L. W., and Lachenbruch. A. H.
of river meander incision.M.S. Grossman, R. 1965. The K 1973. Thermal conditions in
thesis, Colorado State Univ., Fort horizon. A
master soil horizon of permafrost — A review of North
Collins. carbonate accumulation. Soil Sci. American literature. In Proc.
. 1983. Experimental study of 99:74-82. Permafrost 2nd Internal. Conf.
knickpoint and longitudinal profile . 1966. Morphological and pp. 3-23. Natl. Acad. Sci.-Natl.
evolution in cohesive, genetic sequences of carbonate Res. Council, Yakutsk, U.S.S.R.,
homogeneous material. Geol. Soc. accumulation in desert soils. Soil 1973.
America Bull. 94:664-72. Sci. 101:347-60. Goldich. S. 1938. A study of rock
Garrels, R. M., and Mackenzie. F. T. Gilluly, J. 1937. Physiography of the weathering. Jour. Geology
1971. Evolution of sedimentary- Ajo region, Arizona. Geol. Soc. 46:17-58.
rocks. New York: Norton and Co. America Bull. 43:323-48. Goldthwait, R. P. 1951. Development
Garwood, N. C; Janos, D. P.; and . 1949. The distribution of of end moraines in east central
Brokaw, N. 1979. Earthquake- mountain-building in geologic Baffin Island. Jour. Geology
caused landslide: A major time. Geol. Soc. America Bull. 59:567-77.
disturbance to tropical forests. 60:561-90. 1969. Patterned soils and
.
Science 205:997-99. . 1955. Geologic contrasts permafrost on the Presidential

Gentilli, J. 1968. Exfoliation. In between continents and ocean Range (abs). Paris. 8th INQUA
Encyclopedia of geomorphology, basins. Geol. Soc. America Spec. Cong. Resumes des
edited by R. W. Fairbridge, Paper 62:7-18. Communications. 1 50.
pp. 336-39. New York: Reinhold . 1964. Atlantic sediments. , ed. 1971. Till: A
Book Corp. erosion rates, and the evolution of symposium. Columbus: Ohio
Gerrard, A. J. 1981. Soils and the Continental Shelf —Some State Univ. Press.
landforms. London: Allen and speculations. Geol. Soc. America -. 1973. Jerky glacier motion
Unwin Ltd. Bull. 75:483-92. and meltwater. Int. Assoc. Sci.

Gibbs, R. J. 1967. The geochemistry -. 1969. Geological Hydrol. Bull. 95:183-88.
of the Amazon River system, part perspectives and the completeness 1976. Frost sorted patterned
.
I. Geol. Soc. America Bull. of the geologic record. Geol. Soc. ground: A review. Quat. Res.
78:1203-32. America Bull. 80:2303-12. 6:27-35.
Gilbert, G. K. 1877. Geology of the Gilpin, R. R. 1980. A model for the 1979. Giant grooves made
.
Henry Mountains (Utah). U.S. prediction of ice lensing and frost by concentrated basal ice streams.
Geog. and Geol. Survey of the heave in soils. Water Resour. Res. Jour. Glaciol. 23:297-307.
Rocky Mtn. Region. Washington, 16:918-30. Goodman. D. J.: King. G. C. P.:
DC.:' U.S. Govt. Printing Office. Gjessing, J. 1967. On plastic Millar. D. H. M.; and Robin. G.
1917. Hydraulic-mining
.
scouring and subglacial erosion. DeQ. 1979. Pressure-melting
debris in the Sierra Nevada. U.S. Norsk. Geogr. Tidsskr. 20:1-37. effects in basal ice of temperate
Geol. Survey Prof. Paper 105. Glen. J. W. 1952. Experiments on the glaciers: Laboratory studies and
Gile, L. H. 1966. Cambic and certain deformation of ice. Jour. Glaciol. field observations under Glaciere
non-cambic horizons in desert 2:111-14. D'Argentiere. Jour. Glaciol.
soils of southern New Mexico. . 1955. The creep of 23:259-70.
Soil Sci. Soc. Am. Proc. polycrystalline ice. Proc. Royal Gordon. J. E. 1977. Morphometry of
30:773-81. Soc. London, sen A: 228:519-38. cirques in the Kintail Affric-
— . 1975. Holocene soils and .
1958. Mechanical properties Cannich area of northwest
soil-geomorphic relations in an of ice. 1. The plastic properties of Scotland. Geogr. Annlr.
arid region of southern New ice. Philos. Mag.. Suppl. 59A: 177-94.
Mexico. Quat. Res. 5:321-60. 7:254-65. . 1981. Ice-scoured
Gile, L., and Grossman, R. 1979. Glen. J. W.. and Lewis, W. V. 1961. topography and its relationship to
The Desert Project soil Measurements of side-slip at bedrock structure and ice
monograph. U.S. Dept. Agri., Soil Austerdalsbreen, 1959. Jour. movements in parts of northern
Conserv. Serv. Glaciol. 3:1121. Scotland and West Greenland.
Gile, L.; Hawley, J.; and Grossman, Glennie. K. W. 1970. Desert Geogr. Annlr. 63A:55 55.
R. 1981. Soils and geomorphology sedimentary environments. Gordon, M.; Tracey, J.: and Ellis. \1
in Range area of
the Basin and Amsterdam: Elsevier. 1958. Geology of the Arkansas
Southern New Mexico Godfrey, P. J. 1976. Barrier beaches bauxite region. U.S. Geol. Survey
Guidebook to the Desert Project. of the East Coast. Oceanus Prof. Pape"r 299.
New Mexico Bur. Mines and Min. 19:27-40.
Res. Memo 39.
543 Bibliography
Gorsline, D. S. 1966. Dynamic Grim, R. 1962. Applied clay . 1966. Circular patterns and
characteristics of west Florida mineralogy. New York: McGraw- exfoliation in crystalline terrane,
Gulf Coast beaches. Marine Geol. Hill. Grandfather Mountain area,
4:187-206. Grove, J. M. 1960. The bands and North Carolina. Geol. Soc.
Gow, A. J., and Williamson, T. 1976. layers of Vesl-Skautbreen. In America Bull. 77:975-86.
Rheological implications of the Norwegian cirque glaciers, edited 1973. Stream-profile
internal structure and crystal by W. V. Lewis, pp. 1-23. Royal1 analysis and stream-gradient
fabrics of the West Antarctic ice Geog. Soc. Res. Ser. 4. index. U.S. Geol. Survey Jour.
sheet as revealed by deep core Guilcher, A. 1958. Coastal and Research 1:421-29.
drilling at Byrd Station. Geol. submarine morphology. London: Hack, J. T, and Goodlett, J. C. 1960.
Soc. America Bull. 87:1665-77. Methuen and Co. Geomorphology and forest ecology
Graf, W. H. 1971. Hydraulics of Gunn, J. 1981. Hydrological of a mountain region in the
sediment transport. New York: processes in karst depressions. central Appalachians. U.S. Geol.
McGraw-Hill. Zeit.f. Geomorph. 25:313-31. Survey Prof. Paper 347.
Graf, W. L. 1970. The .
1983. Point recharge of Hadley, R. F. 1961. Influence of
geomorphology of the glacial limestone aquifers — A model riparian vegetation on channel
valley cross section. Arc. Alp. from New Zealand karst. shape, northeastern Arizona. U.S.
Res. 2:303-12. J. Hydrol. 61:19-29. Geol. Survey Prof. Paper 424-
. 1976. Cirques as glacier Gupta, A., and Fox, H. 1974. Effects C:30-31.
locations. Arc. Alp. Res. 8:79-90. of high-magnitude floods on 1967. Pediments and
.
Gravenor, C. P. 1953. The origin of channel form: A case study in pediment-forming processes. Jour.
drumlins. Am. Jour. Sci. Maryland Piedmont. Water Geol. Educ. 15:83-89.
251:674-81. Resour. Res. 10:499-509. Hadley, R. F., and Schumm, S. A.
. 1982. Chattermarked Gutenberg, B. 1941. Changes in sea 1961. Sediment sources and
garnets in Pleistocene glacial level, postglacial uplift, and drainage basin characteristics in
sediments. Geol. Soc. America mobility of the earth's interior. upper Cheyenne River basin. U.S.
Bull. 93:751-58. Geol. Soc. America Bull. Geol. Survey Water Supply Paper
Gravenor, C. P., and Kupsch, W. O. 52:721-72. 1531-B:137-96.
1959. Ice disintegration features Guza, R. T, and Inman, D. L. 1975. Haefeli, R., and Brentani, F. 1955.
inwestern Canada. Jour. Geology Edge waves and beach cusps. Observations in a cold ice cap.
67:48-64. Jour. Geophys. Research Jour. Glaciol. 2:571-80.
Gravenor, C. P., and Meneley, W. A. 80:21:2997-3012. Hagerty, D. J. 1980. Multifactor
1958. Glacial flutings in central Haan, C. T, and Johnson, H. P. analysis of bank caving along a
and northern Alberta. Am. Jour. 1966. Rapid determination of navigable stream. In Natl.
Sci. 256:715-28. hypsometric curves. Geol. Soc. Waterways Roundtable Proc.
Gray, J. M. 1982. Unweathered, America Bull. 77:123-25. U.S.Army Engr. Water Res.
glaciated bedrock on an exposed Hack, J. T 1941. Dunes of the Support Ctr., Inst, for Water
lake bed in Wales. Jour. Glaciol. western Navajo country. Geogr. Resources, IWR-80-L463-92.
28:483-97. Rev. 31:240-63. Haig, M. 1979. Ground retreat and
Gray, W. M. 1965. Surface spalling . 1957. Studies of longitudinal slope evolution on regraded
by thermal stresses in rocks. In stream profiles in Virginia and surface-mine dumps, Waunafon,
Rock mechanics symposium. Maryland. U.S. Geol. Survey Gwent. Earth Surf. Proc. and
Toronto: Proc. Ottawa, Can. Dept. Prof. Paper 294-B:45-97. Landforms 4:183-89.
Mines and Tech. Surveys. .
1960a. Relation of solution Haigh, M. J., and Wallace, W. L.
Green, and Short, N. 1971.
J., features to chemical character of 1982. Erosion of strip-mine dumps
Volcanic landforms and surface water in the Shenandoah Valley, in LaSalle County, Illinois:
features. New York: Springer- Virginia. U.S. Geol. Survey Prof. Preliminary results. Earth Surf.
Verlag. Paper 400-B:387-90. Proc. and Landforms 7:79-84.
Gregory, K. and Walling, D. E.
J., -. 1960b. Interpretation of Hallberg, G. R. 1979. Wind-aligned
1973. Drainage basin form and erosional topography in humid drainage in loess in Iowa. Iowa
process. New York: Halsted Press. temperate regions. Am. Jour. Sci. Acad. Sci. Proc. 86:4-9.
Griggs, D. 1936a. The factor of (Bradley Vol.) 258-A:80-97. Hallet, B. 1976a. Deposits formed by
fatigue in rock exfoliation. Jour. 1965. Postglacial drainage subglacial precipitation of CaCO,.
Geology 44:783-96. evolution in the Ontonagan area, Geol. Soc. America Bull.
1936b. Deformation of rocks
. Michigan. U.S. Geol. Survey Prof. 87:1003-15.
under high confining pressures. Paper 504-B:l-40.
Jour. Geology 44:541-77.
Bibliography 544
. 1976b. The effect of Harvey, A. M.; Hitchcock, D. H.: Hennion, F. B., and Lobacz, E. F.
subglacial chemical processes on and Hughes, D. J. 1979. Event 1973. Corps of engineers
glacier sliding. Jour. Glaciol. frequency and morphological technology related to design of
17:209-21. adjustment of fluvial systems. In pavements in areas of permafrost.
. 1979. A theoretical model of Adjustments of the fluvial In Proc. Permafrost 2nd Internal.
glacial abrasion. Jour. Glaciol. system, edited by D. D. Rhodes Conf, pp. 426-29. Natl. Acad.
23:39-50. and G. P. Williams, pp. 139-67. Sci. -Natl. Res. Council, Yakutsk,
Hammad, H. Y. 1972. River-bed Dubuque, Iowa: Kendall Hunt. U.S.S.R., 1973.
degradation after closure of dams. Harwood, T A. 1969. Some possible Herak, M., and Stringfield, V. T.
Am. Soc. Civil Engineers, Jour. problems with pipelines in 1972. Karst. Important karst
Hydraulics Div. 98:591-607. permafrost regions. Proc. 3rd regions of the northern
Handy, R. L. 1976. Loess distribution Canadian Conf. on Permafrost, hemisphere. Amsterdam: Elsevier.
by variable winds. Geol. Soc. pp. 79-84. Natl. Res. Council of Hey, R. D.; Bathurst. J. C; and
America Bull. 87:915-27. Canada. Tech. Memo 96. Thome, C. R. 1982. Gravel-bed
Hanshaw, B., and Back, W. 1980. Hastenrath, S. L. 1967. The rivers. New York: John Wiley &
Chemical mass-wasting of the barchans of the Arequipa region, Sons.
northern Yucatan Peninsula by southern Peru. Zeit.f. Geomorph. Hey. R. D.. and Thorne, C. R. 1975.
groundwater dissolution. Geology 11:300-331. Secondary flows in river channels.
8:222-24. Hastenrath, S., and Kruss, P. 1982. Area 7:191-95.
Happ, S. C; Rittenhouse, G.; and On the secular variation of ice Hickin. E. J. 1974. The development
Dobson, G. C. 1940. Some flow velocity at Lewis Glacier, of meanders in natural river
principles of accelerated stream Mount Kenya, Kenya. Jour. channels. Am. Jour. Sci.
and valley sedimentation. U.S. Glaciol 28:333-39. 274:414-42.
Dept. Agri. Tech. Bull. 695. Hausenbuiller, R. 1972. Soil science: Hickin, E. J., and Nanson, G. C.
Harpstead, M., and Hole, R. 1980. and practices.
Principles 1975. The character of channel
Soil science simplified. Ames: Dubuque, Iowa: Wm. C. Brown migration on the Beatton River,
Iowa State Univ. Press. Company Publishers. northeast British Columbia,
Harris, C. 1973. Some factors Haynes, V. C. 1982. The Darb El- Canada. Geol. Soc. America Bull.
affecting the rates and processes Arba'in Desert: A product of 86:487-94.
of periglacial mass movement. Quaternary climatic change. In Hill, A. R. 1971. The internal
Geogr. Annlr. 55A:24-58. Desert landforms of southwest composition and structure of
Harris, S. A. 1983. Comparison of Egypt: A basis for comparison drumlins in north Down and south
the climatic and geomorphic with Mars, edited by F. El-Baz Antrim, northern Ireland. Geogr.
methods of predicting permafrost andT. Maxwell, pp. 91-1 18. Annlr. 53:14-31.
distribution in western Yukon NASA, CR-3611. . 1973. Erosion of river banks
Territory. In Proc. Permafrost 4th Heim, A. 1932. Bergsturz und composed of glacial till near
Internal. Conf.. pp. 450-55. Natl. Menschenleben. Zurich: Fretz and Belfast, Northern Ireland. Zeit.f.
Acad. Sci. Wasmuth Verlag. Geomorph. 17:428-42.
Harris, S. E. 1943. Friction cracks Heller, P. L. 1981. Small landslide Hill,C. A. 1981. Speleogenesis of
and the direction of glacial types and controls in glacial Carlsbad Caverns and other caves
movement. Jour. Geology deposits: Lower Skagit River of the Guadalupe Mountains.
51:244-58. drainage, northern Cascade Proc. 8th Intl. Cong. Speleol.
Harrison, A. E. 1964. Ice surges on Range, Washington. Environ. Bowling Green, Ky., pp. 143-44.
the Muldrow Glacier, Alaska. Geol. 3:221-28. Hillaire-Marcel, C, and Fairbridge.
Jour. Glaciol. 5:365-68. Helley. E. J., and Smith, W. 1971. R. W. F. 1978. Isostasy and
Harrison, W. D. 1975. Temperature Development and calibration of a eustasy of Hudson Bay. Geology
measurements in a temperate pressure difference bedload 6:117-22.
glacier. Jour. Glaciol. 14:23-30. sampler. U.S. Geol. Survey, Water Hino, M. 1975. Theory on formation
Hartshorn, J. H. 1958. Flowtill in Resources Div., Open-File Rpt. of rip current and cuspidal coast.
southeastern Massachusetts. Geol. Hembree, C, and Rainwater, F. Proc. Nth Conf. on Coast. Eng..
Soc. America Bull. 69:477-82. 1961. Chemical degradation on pp. 901-19.
opposite flanks of the Wind River Hjulstrbm. F. 1939. Transportation
Range, Wyoming. U.S. Geol. of detritus by moving water. In
Survey Water Supply Paper Recent marine sediments: A
1535-E. symposium, edited by P. Trask.
Tulsa. Okla.: Am. Assoc.
Petroleum Geologists.
545 Bibliography
Hodge, S. M. 1974. Variations in the Hooke, R. LeB.; Alexander, E. C, Howard, Arthur D. 1959. Numerical
sliding of a temperate glacier. Jr.: and Gustafson, R. J. 1980. systems of terrace nomenclature:
Jour. Glaciol. 13:349-69. Temperature profiles in the A critique. Jour. Geology
. 1976. Direct measurement Barnes Ice Cap, Baffin Island, 67:239-43.
of basal water pressures: A pilot Canada, and heat flux from the . 1967. Drainage analysis in
study. Jour. Glaciol. 16:205-17. subglacial terrane. Can. J. Earth geologic interpretation: A
Hodgkin, E. P. 1964. Rate of erosion Sci. 17:1174-88. summation. Am. Assoc.
of intertidal limestone. Zeit. f. Hooke. R. LeB., and Hudleston, P. J. Petroleum Geologists Bull.
Geomorph. 8:385-92. 1980. Ice fabrics in a vertical flow- 51:2246-59.
Holdsworth, G. 1973. Ice calving into plane, Barnes Ice Cap, Canada. Howard, Arthur D.; Fairbridge,
the proglacial Generator Lake Jour. Glaciol. 25:195-214. R. W.; and Quinn, J. H. 1968.
Baffin Island. N.W.T.. Canada. .1981. Ice fabrics from a Terraces, fluvial — Introduction. In
Jour. Glaciol. 12:235-50. borehole at the top of the south Encyclopedia of geomorphology,
Holland. H. D.: Kirsipu. T. V.; dome, Barnes Ice Cap, Baffin edited by R. W. Fairbridge,
Huebner, J. S.; and Oxburgh, Island. Geol. Soc. America Bull. pp. 1 17-23. New York: Reinhold
1
V. M. 1964. On some aspects of 92 pt. 1:274-81. Book Corp.

the chemical evolution of cave Hooke, R. LeB., and Rohrer, W. L. Hoyt, J. H. 1966. Air and sand
waters. Jour. Geology 72:36-67. 1977. Relative erodibility of movements in the lee of dunes.
Holman. R. A. 1983. Edge waves and source-area rock types, as Sedimentology 7:137-44.
the configuration of the shoreline. determined from second-order . 1967. Barrier island
In CRC handbook of coastal variations in alluvial-fan size. formation. Geol. Soc. America
processes and erosion, edited by Geol. Soc. America Bull. Bull. 78:1125-36.
P. Komar, pp. 21-33. Boca Raton, 88:1177-82. Hoyt, J. H., and Henry, V. J. 1971.
Fla.: CRC Press. . 1979. Geometry of alluvial Origin of capes and shoals along
Holmes, C. D. 1947. Karnes. Am. fans: Effect of discharge and the southeastern coast of the
Jour. Sci. 245:240-49. sediment size. Earth Surf. Proc. United States. Geol. Soc. America
. 1960. Evolution of till-stone and Landforms 4:147-66. Bull. 82:59-66.
shapes, central New York. Geol. Hopkins, D. M., and Sigafoos, R. S. Hsu, K. J. 1965. Isostasy, crustal
Soc. America Bull. 71:1645-60. 1951. Frost action and vegetation thinning, mantle changes and the
Holmes, G. W.; Hopkins, D. M.: and patterns on Seward Peninsula, disappearance of ancient land
Foster, H. L. 1968. Pingos in Alaska. U.S. Geol. Survey Bull. masses. Am. Jour. Sci.
central Alaska. U.S. Geol. Survey- 974-C51-100. 263:97-109.
Bull. 1241-H. Hoppe, G., and Schytt, V
1953. 1975. Catastrophic debris
.
Holtedahl, H. 1967. Notes on the Some observations on fluted streams (sturzstroms) generated
formation of fjords and fjord moraine surfaces. Geogr. Annlr. by rockfalls. Geol. Soc. America
valleys. Geogr. Annlr. 35:105-15. Bull. 86:129-40.
49:188-203. Horton, R. E. 1933. The role of Hubbert, M. K. 1940. The theory of
Hooke, R. LeB. 1967. Processes on infiltration in the hydrological groundwater motion. Jour.
arid-region alluvial fans. Jour. cycle. Am. Geophys. Union Trans. Geology 48:785-944.
Geology 75:438-60. 14:446-60. Huddart, D., and Lister, H. 1981.
. 1968. Steady-state . 1945. Erosional development The origin of ice marginal
relationships on arid-region of streams and their drainage terraces and contact ridges of
alluvial fans in closed basins. Am. basins: Hydrophysical approach to East Kangerdluarssuk Glacier,
Jour. Sci. 266:609-29. quantitative morphology. Geol. SW Greenland. Geogr. Annlr.
-. 1972. Geomorphic evidence Soc. America Bull. 56:275-370. 63A:31-39.
for Late Wisconsin and Holocene Howard, Alan D. 1971. Simulation Hunt, C. B.; Averitt, P.: and Miller,
tectonic deformation. Death model of stream capture. Geol. R. L. 1953. Geology and
Valley, California. Geol. Soc. Soc. America Bull. 82:1355-76. geography of the Henry
America Bull. 83:2073-97. . 1977. Effect of slope on the Mountains region, Utah. U.S.
. 1977. Basal temperatures in threshold of motion and its Geol. Survey Prof. Paper 228.
polar ice sheets: A qualitative application to orientation of wind Hunt, C. B„ and Mabey, D. R. 1966.
review. Quat. Res. 7:1-13. ripples. Geol. Soc. America Bull. Stratigraphy and structure. Death
88:853-56. Valley, California. U.S. Geol.
Howard, Alan Morton, J. B.;
D.; Survey Prof. Paper 494-A.
Gal-el-Hak, M.; and Pierce, D. Huntley, D. A., and Bowen, A. J.
1977. Simulation model of erosion 1973. Field observations of edge

and deposition on a barchan dune. waves. Nature 243:160-61.
NASA, CR-2838.
A
Bibliography 546
Huntoon, P. W. 1974. The karstic Ippen. A. T, ed. 1966. Estuary and Jennings. J. N. 1967. Some karst
groundwater basins of the Kaibab coastline hydrodynamics. New areas of Australia. In Landform
Plateau, Arizona. Water Resour. York: McGraw-Hill. studies from Australia and New
Res. 10:579-90. Isherwood, D., and Street, A. 1976. Guinea, edited by J. N. Jennings
Hursh, C. R. 1936. Storm-water and Biotite-induced grussification of and J. A. Mabbutt, pp. 256-92.
absorption. Am. Geophys. Union the Boulder Creek Granodiorite, Canberra.
Trans. 17:301-2. Boulder County, Colorado. Geol. . 1971. Karst. Cambridge.
Hursh, C. R., and Brater. E. F. 1941. Soc. America Bull. 87:366-70. Mass.: M.I.T. Press.
Separating storm hydrographs Ives, J. D.. and Fahey. B. D. 1971. . 1983. Karst landforms. Am.
from small drainage areas into Permafrost occurrence in the Scientist 71:578-86.
surface and subsurface flow. Am. Front Range, Colorado Rocky Jennings.J. N.. and Bik, M. J. 1962.
Geophvs. Union Trans. Mountains, U.S.A. Jour. Glaciol. Karst morphology in Australian
22:863-70. 10:105-11. New Guinea. Nature
Hutchinson, N. 1968. Mass
J. Jackson, J. A.; Gagnepain, J.; 194:1036-38.
movement. In Encyclopedia of Houseman, G.; King, G. C. P.; Jenny. H. 1941. Factors of soil
geomorphology, edited by R. W. Papadimitriou, P.; Soufleris, C; formation. New York: McGraw-
Fairbridge. pp. 688-96. New and Virieux, J. 1982. Seismicity, Hill.
York: Reinhold Book Corp. normal faulting, and the . 1950. Origin of soils. In
Hutter. K. 1982. Glacier flow. Am. geomorphological development of Applied sedimentation, edited by
Scientist 70:26-34. the Gulf of Corinth (Greece): The P. Trask, pp. 41-61. New York:
Hutton. C. E. 1947. Studies of loess- Corinth earthquakes of February John Wiley & Sons.
derived soils in southwestern and March 1981. Earth and Jenny. H., and Leonard. C. 1939.
Iowa. Soil Sci. Soc. Am. Proc. Planat. Sci. Letters 57:377-97. Functional relationships between
12:424-31. Jackson, M.; Hseung, Y.; Corey, R.: soil properties and rainfall. So/7
Iken, A.; Rothlisberger, H.; Flotron, Evans, E.; and Heuval, R. 1952. Sci. 38:363-81.
A.; and Haeberli, W. 1983. The Weathering sequence of clay size Johnson, A. 1970. Physical process
Unteraargletscher at the
uplift of minerals in soils and sediments. in geology.San Francisco:
beginning of the melt season — Soil Sci. Soc. Am. Proc. 16:3-6. Freeman, Cooper and Co.
consequence of water storage at Jackson, T, and Keller, W. 1970. A Johnson. A. M.. and Rahn, P. H.
the bed? Jour. Glaciol. 19:28-47. comparative study of the role of 1970. Mobilization of debris flows.
Inglis, C. C. 1949. The behavior and lichens and "inorganic" processes Zeit.f. Geomorph.. Suppl.
control of rivers and canals. in the chemical weathering of 9:168-86.
Research Pub. Poona, India, no. recent Hawaiian lava flows. Am. Johnson, D. W. 1910. Beach cusps.
13, 2 vols. Jour. Sci. 269:446-66. Geol. Soc. America Bull.
Inman, D. L., and Bagnold, R. A. Jacobel, R. W. 1982. Short-term 21:604-21.
1963. Littoral processes. In The variations in velocity of South 1919. Shore processes and
.
sea. edited by M. N. Hill, Cascade Glacier, Washington, shoreline development. New

3:529-53. New York: U.S.A. Jour. Glaciol 28:325-32. York: John Wiley & Sons.
Interscience. Jahn, A. 1960. Some remarks on Facsimile edition: Hafner. New
Inman, D. L., and Brush, B. M. evolution of slopes on Spitsbergen. York, 1965.
1973. The coastal challenge. Zeit.f. Geomorph., Suppl. 1925. New England-
Science 181:20-32. 1:49-58. Acadian shoreline. New York:
Inman, D. L.; Ewing, G. C; and Jansen, J. M. L., and Painter, R. B. John Wiley & Sons.
Corliss, J. B. 1966. Coastal sand 1974. Predicting sediment yield 1932. Rock fans of arid
.
dunes of Guerrero Negro, Baja, from climate and topography. regions. Am. Jour. Sci. 23 (5th
California, Mexico. Geol. Soc. J. Hydrol. 21:371-80. ser.):389-416.
America Bull. 77:787-802. Jarvis, G. T, and Clarke. G. K. C. . 1944. Problems of terrace
Inman. D. L., and Frautschy, J. D. 1975. The thermal regime of correlation. Geol. Soc. America
1966. Littoral processes and the Trapridge Glacier and its Bull. 55:793-818.
development of shoreline. Proc. relevance to glacier surging. Jour. Johnson. J. W. 1956. Dynamics of
Coast. Eng. Speciality Conf., Am. Glaciol. 14:235-49. nearshore sediment movement.
Soc. Civil Engineers (Santa and Sham. C. H. 1981.
Jarvis. R. S.. Am. Assoc. Petroleum Geologists
Barbara, Calif.), pp. 511-36. Drainage network structure and Bull. 40:2211-32.
Inman. D. L., and Nordstrom, C. E. the diameter-magnitude relation. Johnson. W. D. 1904. The profile of
1971. On the tectonic and Water Resour. Res. 17:1019-27. maturity in alpine glacial erosion.
morphologic classification of Jour. Geology 12:7:569 78
Geology 79:1-21.
coasts. Jour.
547 Bibliography
Johnston, G. H. 1963. Pile A. 1971. Areal sorting of

Keller, E. Kellerhals, R. 1967. Stable channels
construction in permafrost. Proc. bedload material. Geol. Soc. with gravel-paved beds. Am Soc.
Permafrost Internat. Con/., America Bull. 82:753-56. Civil Engineers Proc, Jour.
(Lafayette, Ind., 1963). Natl. . 1972. Development of Waterways and Harbors
Acad. Sci.-Natl. Res. Council alluvial stream channels. A five- 93:63-84.
Pub. 1287, pp. 477-81. stage model. Geol. Soc. America Kelsey, H. M. 1980. A sediment
1983. Performance of an
.
Bull. 83:1531-36. budget and analysis of
insulated roadway on permafrost, Keller, E. A.; Bonkowski, M. S.; geomorphic process in the Van
Inuvik, N.W.T. In Proc. Korsch, R. J.; and Shlemon, R. J. Duzen River basin, north coastal
Permafrost 4th Internat. Conf., 1982. Tectonic geomorphology of California, 1941-1975. Geol. Soc.
pp. 548-51. Natl. Acad. Sci. the San Andreas fault zone in the America Bull. 91:190-95.
Jopling, A. V. 1966. Some southern Indio Hills, Coachella Kemmerly, P. R. 1976. Definitive
application of theory and Valley, California. Geol. Soc. doline characteristics in the
experiment to the study of America Bull. 93:46-56. Clarksville quadrangle, Tennessee.
bedding genesis. Sedimentology Keller, E. A.,and Melhorn, W. 1973. Geol. Soc. America Bull.
7:71-102. Bedforms and fluvial processes on 87:42-46.
Judson, S. 1968a. Erosion rates near alluvial stream channels: selected . 1980a. Sinkhole collapse in
Rome, Italy. Science observations. In Fluvial Montgomery County, Tennessee.
160:1444-46. geomorphology, edited by Tenn. Div. Geol., Environ. Geol.
1968b. Erosion of the land.
. M. Morisawa, pp. 253-83. Ser. no. 6.
Am. Scientist 56:356-74. S.U.N.Y., Binghamton: Pubs, in . 1980b. A time-distribution
Judson, S., and Ritter, D. F. 1964. Geomorphology. study of doline collapse:
Rates of regional denudation in . Rhythmic spacing and
1978. Framework for prediction.
the United States. Jour. Geophys. origin of pools and riffles. Geol. Environ. Geol. 3:123-30.
Research 69:3395-401. Soc. America Bull. 89:723-30. 1982. Spatial analysis of a
Kachadoorian, R., and Ferrians, Keller, E. A., and Swanson, F. J. karst depression population: Clues
O. J., Jr. 1973. Permafrost-related 1979. Effects of large organic to genesis. Geol. Soc. America
engineering problems posed by the material on channel form and Bull. 93:1078-86.
Trans-Alaskan Pipeline. fluvial processes. Earth Surf. Kemmerly, P., and Towe, S. 1978.
Permafrost 2nd Internat. Conf., Proc. and Landforms 4:361-80. Karst depressions in a time
pp. 684-87. Natl. Acad. Sci.- Keller, E. A., and Tally, T. 1979. context. Earth Surf. Proc. and
Natl. Res. Council, Yakutsk, Effects of large organic debris on Landforms 3:355-61.
U.S.S.R., 1973. channel form and fluvial processes Kesel, R. H. 1973. Inselberg
Kamb, B. 1964. Glacier mechanics. in the coastal redwood landform elements: Definition and
Science 146:353-65. environment. In Adjustments of geomorph.
synthesis. Rev.
1970. Sliding motion of
. the fluvial system, edited by dynamique 22:97-108.
glaciers: Theory and observation. D. Rhodes and G. Williams, . 1977. Some aspects of the
Rev. Geophys. and Space Phys. pp. 169-97. Dubuque, Iowa: geomorphology of inselbergs in
8:673-728. Kendall/Hunt Publishing central Arizona, U.S.A. Zeit. f.
Kamb, B., and LaChapelle, E. 1964. Company. Geomorph. 21:119-46.
Direct observation of the W. 1954. Bonding energies of
Keller, Kesel, R. H.; Dunne, K.: McDonald,
mechanism of glacier sliding over some silicate minerals. Am. R.; Allison, K.; and Spicer, B.
bedrock. Jour. Glaciol. 5:159-72. Mineralogist 39:783-93. 1974. Lateral erosion and
Kaye, C. A. 1957. The effect of 1976. Scan electron
. overbank deposition on the
solvent motion on limestone micrographs of kaolins collected Mississippi River in Louisiana
solutions. Jour. Geology from diverse environments of caused by 1973 flooding. Geology
65:34-47. origin. Pt. 1. Clays and Clay 2:461-64.
1964a. Outline of
. Minerals 24:107-13. Kesseli, J. E. 1941. Rock streams in
Pleistocene geology of Martha's . 1978. Kaolinization of the Sierra Nevada, California.
Vineyard, Massachusetts. U.S. feldspar as displayed in scanning Geogr. Rev. 31:203-27.
Geol. Survey Prof. Paper 501- electron micrographs. Geology Keyes, C. R. 1912. Deflative scheme
C134-39. 6:184-88. of the geographic cycle in an arid
1964b. Illinoian and early
-. . 1982. Kaolin —A most climate. Geol. Soc. America Bull.
Wisconsin moraines of Martha's diverse rock in genesis, texture, 23:537-62.
Vineyard, Massachusetts. U.S. physical properties and uses. Geol. Kiersch, G. A. 1964. Vaiont reservoir
Geol. Survey Prof. Paper 501- Soc. America Bull. 93:27-36. disaster. Civil Engineering
C: 140-43. 34:32-39.
Bibliography 548
Kilpatrick, F. A., and Barnes, H. H. Kneale, W. R. 1982. Field 1973. Observations of beach
1964. Channel geometry of measurements of rainfall drop- cusps at Mono Lake, California.
piedmont streams as related to size distribution,and the Geol. Soc. America Bull.
frequency of floods. U.S. Geol. relationship between rainfall 84:3593-3600.
Survey Prof. Paper 422E:1 — 10. parameters and soil movement by .1975. Nearshore currents:
King, C. A. M. 1972. Beaches and rainsplash. Earth Surf. Proc. and Generation by obliquely incident
coasts. New York: St. Martin's Landforms 7:499-502. waves and longshore variations in
Press. Knighton, A. D. 1974. Variation in breaker height. In Proc.
King, C. A. M., and Buckley, J. T. width-discharge relation and some symposium on nearshore
1968. The analysis of stone size implications for hydraulic sediment dynamics, edited by
and shape in Arctic environments. geometry. Geol. Soc. America J. R. Hails and A. Carr.
Jour. Sed. Petrology 38:200-214. Bull. 85:1069-76. pp. 17-45. London: John Wiley &
King, C. A. M., and Lewis, W. V. . 1977. Alternative derivation Sons.
1961. A tentative theory of ogive of the minimum variance 1976. Beach processes and
formation. Jour. Glaciol. hypothesis. Geol. Soc. America sedimentation. Englewood Cliffs,
3:913-39. Bull. 88:364-66. N.J.: Prentice-Hall.
King, C. A. M., and McCullagh, Knox, J. C. 1972. Valley alluviation -. 1979. Comparisons of the
M. J. 1971. A simulation model of in southwestern Wisconsin. Ann. hydraulics of water flows in
a complex recurved spit. Jour. Assoc. Am. Geog. 62:401-10. Martian outflow channels with
Geology 79:22-37. . 1976. Concept of the graded flows of similar scale on Earth.
King, C. A. M., and Williams, W. W. stream. In Theories of landform Icarus 37:156-81.
1949. The formation and development, edited by 1983a. Beach processes and
movement of sandbars by wave W. Melhorn and R. Flemal, erosion —an introduction. In CRC
action. Geogr. Jour. 107:70-84. pp. 168-98. S.U.N.Y., handbook of coastal processes
King, D. 1956. The Quaternary Binghamton: Pubs, in and erosion, edited by P. Komar,
stratigraphic record at Lake Eyre Geomorphology. pp. 1-20. Boca Raton, Fla.: CRC
North and the evolution of Kochel, R. C; Baker, V. R.; and Press.
existing topographic forms. Trans. Patton, P. C. 1982. -, ed. 1983b. CRC handbook
Royal Soc. Australia 79:93-103. Paleohydrology of Southwestern of coastal processes and erosion.
King, L. C. 1953. Canons of Texas. Water Resour. Res. Boca Raton, Fla.: CRC Press.
landscape evolution. Geol. Soc. 18:1165-83. Komar, P. D., and Inman, D. L.
America Bull. 64:751-52. Kochel, R. C, and Johnson, R. A. 1970. Longshore sand transport
Kirkby, M. J. 1967. Measurement 1984. Geomorphology and on beaches. Jour. Geophys.
and theory of soil creep. Jour. sedimentology of humid- Research 75:30:5914-27.
Geology 75:359-78. temperate alluvial fans, central Komar, P. D., and Reimers, C. E.
. 1969. Infiltration, Virginia. In Gravels and 1978. Grain shape effects on
throughflow, and overland flow; conglomerates, edited by settling rates. Jour. Geologv
and erosion by water on hillslopes. E. Koster and R. Steel, 86:193-209.
In Water, earth, and man, edited pp. 109-22. Can. Soc. Petrol. Konrad, J. M., and Morgenstern.
by R.J. Chorley, pp. 215-38. Geol. Mem. 10. N. R. 1983. Frost susceptibility of
London: Methuen and Co. Kolb, C. R., and Van Lopik, J. R. soils in terms of their segregation
Kirkby, M. J., and Chorley, R. J. 1958. Geology of the Mississippi potential. In Proc. Permafrost 4th
1967. Throughflow, overland flow River deltaic plain, southeastern Internal. Conf. pp. 660-65. Natl.
and erosion. Int. Assoc. Sci. Louisiana. U.S. Army Corps Acad. Sci.
Hydrol. Bull. 12:5-21. Engrs., Waterway Exp. Sta. Rept. Kottlowski, F.: Cooley, M.; and
Kirkby, M. J., and Kirkby, A. V. 3-483, Vicksburg. Ruhe, R. 1965. Quaternary
1969. Erosion and deposition on a Komar, P. D. 1971a. The mechanics geology of the southwest. In The
beach raised by the 1964 of sand transport on beaches. Quaternary of the United Slates.
earthquake, Montague Island, Jour. Geophys. Research edited by H. Wright and D. Frey.
Alaska. U.S. Geol. Survey Prof.
Paper 543-H:l-41.
Kirkby, R. 1969. Variation
—
76:3:713-21.
. 1971b. Nearshore cell
Princeton, N.J.: Princeton Univ.
Press.
P. in circulation and the formation of Krigstrom. A. 1962.
glacial deposition in a subglacial giant cusps. Geol. Soc. America Geomorphological studies of
environment: An example from Bull. 82:2643-50. sandar plains and their braided
Midlothian. Scott. J. Geol. rivers in Iceland. Geogr. Annlr.
5:49-53. 44:328-46.
—
549 Bibliography
Krinsley, D. H., and Donahue, J. M. 1934. Mechanik und

Lagally, Lattman, L. H. 1960. Cross section
1968. Environmental Thermodynamik des stationdren of a flood plain in a moist region
interpretation of sand grain Gletschers. Leipzig. of moderate relief. Jour. Sed.
surface textures of electron Lambe, T. 1953. The structure of Petrology 30:275-82.
microscopy. Geol. Soc. America inorganic soils. Am. Soc. Civil . 1968. Structural control in
Bull. 79:743-48. Engineers Proc. 79: Separate 315. geomorphology. In Encylopedia of
Kroethe, N., and Libra, R. 1983. Lancaster, N. 1980. The formation of geomorphology, edited by R. W.
Sulfur isotopes and hydrochemical seif dunes from barchans Fairbridge, pp. 1074-79. New
variations in spring waters of Supporting evidence for Bagnold's York: Reinhold Book Corp.
southern Indiana. U.S.A. model from the Namib Desert. 1973. Calcium carbonate
.
J. Hydrol. 61:267-83. Zeit.f. Geomorph. 24:160-67. cementation of alluvial fans in

Krumbein. W. C. 1944. Shore 1982. Dunes on the Skeleton
. Southern Nevada. Geol. Soc.
currents and sand movement on a Coast, Namibia (South West America Bull. 84:3013-28.
model beach. U.S. Army Corps Africa): Geomorphology and Laury, R. L. 1971. Stream bank
Engrs., Beach Erosion Board grain size relationships. Earth failure and rotational slumping.
Tech. Memo 7. Surf. Proc. and Landforms Preservation and significance in
Krumbein, W. C, and Oshiek, L. E. 7:575-87. the geologic record. Geol. Soc.
1950. Pulsation transport of sand Lane, E. W. 1937. Stable channels in America Bull. 82:1251-66.
by shore agents. Am. Geophys erodible materials. Am. Soc. Civil LaValle, P. 1967. Some aspects of
Union Trans. 31:216-20. Engineers Trans. 102:123-94. linear karst depression
Kuenen, P. H. 1948. The formation 1955. Design of stable
. development in south central
of beach cusps. Jour. Geology channels. Am. Soc. Civil Kentucky. Ann. Assoc. Am. Geog.
56:34-40. Engineers Trans. 120:1234-79. 57:49-71.
.1960. Experiment abrasion, 1957. A study of the shape
. . 1968. Karst depression
4. Eolian action. Jour. Geology of channels formed by natural morphology in south-central
68:427-49. streams in erodible material. Kentucky. Geogr. Annlr.
Kuhn, G. G., and Shepard, F. P. M.R.D. Sediments Series no. 9, 50A:94-108.
1983. Beach processes and sea U.S. Army Corps Engrs., Eng. Lawson, A. C. 1915. The epigene
cliff erosion in San Diego County, Div., Missouri River, Omaha, profiles of the desert. Univ. of
California. In CRC handbook of Neb. Calif. Dept. Geol. Bull. 9:23-48.
coastal processes and erosion, Langbein, W. B. 1964. Geometry of Lee, W., and Uyeda, S. 1965. Review
edited by P. Komar, pp. 267-84. river channels. Am. Soc. Civil of heat flow data. In Terrestrial
Boca Raton, Fla.- CRC Press. Engineers, Jour. Hydraulics Div. heat flow, edited by W. Lee. Am.
Kuno, H. 1969. Plateau basalts. In 90:301-13. Geophys. Union, Geophys.
The Earth 's crust and upper Langbein, W. B., and Leopold, L. B. Monograph 8.
mantle, edited by P. Hart, 1964. Quasi-equilibrium states in Legget, R. 1967. Soil: Its geology
pp.495-500. Am. Geophys. channel morphology. Am. Jour. and use. Geol. Soc. America Bull.
Union. Geophys. Monograph 13. Sci. 262:782-94. 78:1433-60.
Kupsch, W. O. 1955. Drumlins with 1966. River meanders:
. LeGrand, H. 1983. Perspective on
jointed boulders near Dollard, Theory of minimum variance. karst hydrology. J. Hydrol.
Saskatchewan. Geol. Soc. U.S. Geol. Survey Prof. Paper 61:343-55.
America Bull. 66:327-38. 422-H. Lehman, D. 1963. Some principles of
Lachenbruch, A. H. 1966. Langbein, W. B., and Others. 1949. chelation chemistry. Soil Sci. Soc.
Contraction theory of ice wedge Annual runoff in the United Am. Proc. 27:167-70.
polygons: A qualitative discussion. States. U.S. Geol. Survey Lehmann, H. 1936. Morphologische
Proc. Permafrost Inlernat. Conf. Circular 52. Studien auf Java. Stuttgart:
(Lafayette, Ind. 1963). Natl. Langbein, W. B., and Schumm, S. A. Geogr. Abh. 3:9.
Acad. Sci.-Natl. Res. Council 1958. Yield of sediment in Lehre, A. K. 1982. Sediment budget
Pub. 1287, pp. 63-71. relation to mean annual of a small coast range drainage
. 1970. Some estimates of the precipitation. Am. Geophys. basin in north-central California.
thermal effects of a heated Union Trans. 39:1076-84. In Sediment budgets and routing
pipeline in permafrost. U.S. Geol. Langford-Smith, T, and Dury, G. H. in forested drainage basins,
Survey Circ. 632. 1964. A pediment at Middle edited by F. J. Swanson et al.,
Lachenbruch, A. H., and Marshall, Pinnacle, near Broken Hill, New pp. 67-77. U.S.D.A., Forest Serv.
B. V. 1969. Heat flow in the South Wales. Jour. Geol. Soc. Genl. Tech. Rpt. PNW-141.
Arctic. Arctic 22:300-311. Australia 11:79-88.
—
Bibliography 550
Leigh, C. 1982. Sediment transport Leopold, L. B., and Miller, J. P. . 1968. General theory of
by surface wash and throughflow 1956. Ephemeral streams subglacial cavitation and sliding
at the Pasoh Forest Reserve, hydraulic factors and their of temperate glaciers. Jour.
Negri Sembilan, Peninsular relation to the drainage net. U.S. Glaciol. 7:21-58.
Malaysia. Geogr. Annlr. Geol. Survey Prof. Paper 282-A. Lliboutry, L., andReynaud, L. 1981.
64A:171-80. Leopold, L. B., and Wolman, M. G. "Global dynamics" of a
Leighton, M. W., and Pendexter, C. 1957. River channel patterns; temperature valley glacier. Mer
1962. Carbonate rock types. In braided, meandering and straight. de Glace and past velocities
of carbonate rocks,
Classification U.S. Geol. Survey Prof. Paper deduced from Forbes bands. Jour.
edited by Ham, pp. 33-61.
W. E. 282-B. Glaciol. 27:207-26.
Am. Soc. Petroleum Geologists . 1960. River meanders. Geol. Lobacz, E. F., and Quinn, W. F.
Mem. 1. Soc. America Bull. 71:769-794. 1963. Thermal regime beneath

Leliavsky, S. L. 1966. An Leopold, L. B.; Wolman, M. G.; and buildings constructed on
introduction to fluvial hydraulics. Miller, J. P. 1964. Fluvial permafrost. Proc. Permafrost
New York: Dover Publications. processes in geomorphology. San Internat. Conf, (Lafayette, Ind.,
Lemke, R. W. 1958. Narrow linear Francisco: W. H. Freeman. 1963). Natl. Acad. Sci.-Natl. Res.
drumlins near Velva, North Lewin, J. 1976. Initiation of bed Council Pub. 1287, pp. 159-64.
Dakota. Am. Jour. Sci. forms and meanders in coarse- Lockwood, J. G. 1979. Causes of
256:270-83. grained sediment. Geol. Soc. climate. London: Edward Arnold

Leopold, L. B. 1953. Downstream America Bull. 87:281-85. Ltd.
change of velocity in rivers. Am. Lewis, W. V. 1947. Valley steps and Lohnes, R. A., and Handy. R. L.
Jour. Sci. 251:606-24. glacial valley erosion. Inst. Brit. 1968. Slope angles in friable loess.
. 1982. Water surface Geog. Trans. 14:19-44. Jour. Geology 76:247-58.
topography in river channels and .1954. Pressure release and Lombard. R. E.; Miles, M. B.:
implications for meander glacial erosion. Jour. Glaciol. Nelson, L. M.; Kresch, D. L.; and
development. In Gravel-bed 2:417-22. Carpenter, P. J. 1981. Channel
rivers, edited by R. Hey, -, ed. 1960. Norwegian cirque conditions in the lower Toutle and
J. Bathurst, and C. Thorne, glaciers. Royal Geog. Soc. Res. Cowlitz Rivers resulting from the
pp. 359-88. New York: John Ser. 4. mudflows of May 18. 1980. U.S.
Wiley & Sons. Li,Y H. 1976. Denudation of Geol. Survey Circ. 850-C.
Leopold, L. B., and Bull, W. B. 1979. Taiwan Island since the Pliocene Long. J. T, and Sharp, R. P. 1964.
Base aggradation and grade.
level Epoch. Geology 4:105-7. Barchan-dune movement in the
Proc. Am.
Phil. Soc. Likens, G. E.; Bormann, F. H.: Imperial Valley, California. Geol.
123:168-202. Pierce, R. S.; Eaton, J. S.; and Soc. America Bull. 75:149-56.
Leopold, L. B., and Emmett, W. W. Johnson, N. M. 1977. Longuet-Higgins, M. S. 1970.
1976. Bedload measurements, Biogeochemistry of a forested Longshore currents generated by
East Fork River, Wyoming. Natl. ecosystem. New York: Springer- obliquely incident sea waves.
Acad. Sci. Proc. 73:1000-1004. Verlag. Jour. Geophvs. Research
— 1977. 1976 bedload
. Linell, K. A., and Johnston, G. H. 75:6778-6801.
measurements, East Fork River, 1973. Engineering design and Longwell, C. R. 1960. Interpretation
Wyoming. Natl. Acad. Sci. Proc. construction in permafrost of the leveling data. U.S. Geol.
74:2644-48. regions. Proc. Permafrost 2nd Survey Prof. Paper 295:33-38.
Leopold, L. B., and Langbein, W. B. Internat. Confi, pp. 553-75. Natl. Loughnan. F. 1969. Chemical
1962. The concept of entropy in Acad. Sci. -Natl. Res. Council, weathering of the silicate
landscape evolution. U.S. Geol. Yakutsk, U.S.S.R., 1973. minerals. New York: American
Survey Prof. Paper 500-A:20. Livesey, R. H. 1965. Channel Elsevier.
1963. Association and
. armoring below Fort Randall Loughnan, F., and Bayliss. P. 1961.
indeterminancy in geomorphology. dam. Proc. Fed. Inter-Agency The mineralogy of the bauxite
In The fabric of geology, edited Sedimentation Conf. U.S. Dept. deposits near Weipa. Queensland.
by C. C. Albritton, pp. 184-92. of Agri. Pub. 970:461-70. Am. Mineralogist 46:209-17.
Reading, Mass.: Addison-Wesley. Livingstone, D. A. 1963. Chemical Lozinski, W. 1912. Die periglaziale
Leopold, L. B., and Maddock, T., Jr. composition of rivers and lakes. Fazies der mechanischen
1953. The hydraulic geometry of U.S. Geol. Survey Prof. Paper Verwitterung. Internat. Geol.
stream channels and some 440-G. Cong., 11th. Stockholm. 1910,
physiographic implications. U.S. Lliboutry. L. 1964. Subglacial Compte rendu, pp. 1039 53
Geol. Survey Prof. Paper 252. "supercavitation" as a cause of
the rapid advances of glaciers.
Nature 202:77.
551 Bibliography
Lugn, A. L. 1 962. The origin and 1975a. The closing of ice-

.
Marrs, R. W., and Kolm, K. E., eds.
sources of loess. Lincoln: Univ. wedge cracks in permafrost, 1982. Interpretation of windflow
Nebraska Studies, new series, no. Garry Island, Northwest characteristics from eolian
26. Territories. Can. J. Earth Sci. landforms. Geol. Soc. America
. 1968. The origin of loesses 12:1668-74. Spec. Paper 192.
and Great
their relation to the -. 1975b. The stability of Martel, E. A. 1921. Nouveau traite
Plains in North America. In permafrost and recent climatic des eaux souterraines. Paris:
Loess and related eolian deposits change in the Mackenzie Valley. Delagrave.
of the world, edited by C. B. NWT. Canada Geol. Survey Martini, I. P. 1978. Tafoni
Schultz and J. C. Frye. p. 139. Paper 75-1 B:173-76. weathering, with examples from
Lincoln: Univ. Nebraska Press. Mackay, J. R., and Matthews, W. H. Tuscany. Italy. Zeit.f. Geomorph.
Luk, S. H. 1979. Effect of soil 1974. Movement of sorted stripes, 22:44-67.
properties on erosion by wash and the Cinder Cone, Garibaldi Park, Matschinski, M. 1968. Alignment of
splash. Earth Surf. Proc. and B.C.. Canada. Arc. Alp. Res. dolines northwest of Lake
Landforms 4:241-55. 6:347-59. Constance, Germany. Geol. Mag.
Lumley. J. L.. and Panofski. H. A. Mackin, J. H. 1936. The capture of 105(1):56-61.
1964. The structure of the Grevbull River. Am. Jour. Matsumota, T 1967. Fundamental
atmospheric turbulence. New Sci. 31:373-85. problems in the circum-Pacific
York: John Wiley & Sons. . 1937. Erosional history of orogenesis. Tectonophysics
Lustig, L. K. 1965. Clastic the Big Horn Basin, Wyoming. 4:595-613.
sedimentation in Deep Springs Geol. Soc. America Bull. Matthes, F. E. 1900. Glacial
Valley, California. U.S. Geol. 48:813-93. sculpture of the Bighorn
Survey Prof. Paper 352-F. -. 1948. Concept of the graded Mountains, Wyoming. U.S. Geol.
1969. Trend surface analysis
. river. Geol. Soc. America Bull. Survey 21st Ann. Rept.,
of the Basin and Range province 59:463-512. 1899-1900, pt. 2, pp. 167-90.
and some geomorphic . 1963. Rational and 1930. Geologic history of the
.
implications. U.S. Geol. Survey empirical methods of investigation Yosemite Valley. U.S. Geol.
Prof. Paper 500-D. in geology. In The fabric of Survey Prof. Paper 160.
Lynch, D. K. 1982. Tidal bores. Sci. geology, edited by C. Albritton. Matthews, R. K. 1973. Relative
Amer. 247:146-56. Reading. Mass.: Addison- Wesley. elevation of late Pleistocene high
Mabbutt, J. A. 1966. Mantle- Maddock. T, Jr. 1969. The behavior sea level stands; Barbados uplift
controlled planation of pediments. of straight open channels with rates and their implications. Quat.
Am. Jour. Sci. 264:78-91. movable beds. U.S. Geol. Survey Res. 3:147-53.
1971. The Australian arid
. Prof. Paper 622-A:70. May, and Tanner, W. F. 1973.
J. P.,
zone as a prehistoric environment. Malott, C. A. 1921. A subterranean The power gradient and
littoral
In Aboriginal man and cut-off and other subterranean shoreline changes. In Coastal
environment in Australia, edited phenomena along Indian Creek, geomorphology, edited by D. R.
by D. Mulvaney and J. Golson, Laurence Co., Indiana. Indiana Coates, pp. 43-60. S.U.N.Y,
pp. 66-79. Canberra: Australian Acad. Sci. Proc. 31:203-10. Binghamton: 3rd Ann. Geomorph.
Natl. Univ. Press. .
1938. Invasion theory of Symposium.
Macdonald, G. 1972. Volcanoes. cavern development (abs.). Geol. May, S. K.; Dolan, R.; and Hayden,
Englewood Cliffs, N.J.: Prentice- Soc. America Proc. 1937, p. 323. B. P. 1983. Erosion of U.S.
Hall. -. 1939. Karst valleys. Geol. shorelines. EOS 64:521-23.
Mackay, J. R. 1970. Disturbances to Soc. America Bull. 50:1984. McCall, J. G. 1952. The internal
the tundra and forest tundra Mammerickx, J. 1964. Quantitative structure of a cirque glacier:
environment of the western observations on pediments in the Report on studies of the englacial
Arctic. Can. Geotechnical Jour. Mojave and Sonoran deserts movements and temperatures.
7:420-32. (southwestern United States). Jour. Glaciol. 2:\l2-l\.
1973. The growth of pingos,
. Am. Jour. Sci. 262:417-35. .
1960. The flow
western Arctic coast, Canada. Manley, S., and Manley, R. 1968. characteristics of a cirque glacier
Can. J. Earth Sci. 10:979-1004. Beaches; their lives, legends and and their effect on glacial
-. 1974. Ice-wedge cracks, lore. Philadelphia: Chilton. structure and cirque formation. In
Garry Island, Northwest Mark, D. M. 1974. Line intersection Norwegian cirque glaciers, edited
Territories. Can. J. Earth Sci. method for estimating drainage by W. V. Lewis, pp. 39-62. Royal
11:1366-83. density. Geology 2:235-36. Geog. Soc. Res. Ser. 4.
Bibliography 552
McComas, M.; Hinkley, K.; and McKee, E. D., and Tibbitts, G. C, Mellor, M. 1970. Phase composition
Kempton, J. 1969. Coordinated Jr. 1964. Primary structures of a of pore water in cold rocks. U.S.
mapping of geology and soils for seif dune and associated deposits Army Corps Engrs.. Cold Regions
land-use planning. Illinois Geol. in Libya. Jour. Sed. Petrology Res. and Eng. Lab. Research
Survey Environ. Geol. Note 29. 34:5-17. Rept. 292.
McCoy, R. M. 1971. Rapid McKenzie, G. D. 1969. Observations Melton. F. A. 1940. A tentative
measurement of drainage density. on a collapsing kame terrace in classification of sand dunes. Jour.
Geol. Soc. America Bull. Glacier Bay National Monument, Geology 48:113-7 3.
82:757-62. S.E. Alaska. Jour. Glaciol. Melton, M. A. 1958. Correlation
McDonald, R. C. 1975. Observations 8:413-25. structure of morphometric
on hillslope erosion in tower karst McLean, R. F., and Kirk, R. M. properties of drainage systems
topography of Belize. Geol. Soc. 1969. Relationship between grain and their controlling agents. Jour.
America Bull. 86:255-56. size, size-sorting and foreshore Geology 66:442-60.
. Tower karst
1979. slope on mixed sand-shingle 1965a. The geomorphic and
.
geomorphology in Belize. Zeit f. beaches. New Zealand J. Geol. paleoclimatic significance of

Geomorph., Suppl. 32:35-45. and Geophys. 12:138-55. alluvial deposits in southern
McDowall, I. C. 1960. Particle size McPherson, H. J., and Rannie, W. F. Arizona. Jour. Geology 73:1-38.
reduction of clay minerals by 1967. Geomorphic effects of the 1965b. Debris-covered
.
freezing and thawing. New May, 1967, flood in Graburn hillslopes of the southern Arizona
Zealand J. Geol. and Geophys. watershed, Cypress Hills, Alberta, desert —consideration of their
3:337-43. Canada. /. Hydro!. 9:307-21. stability and sediment
McGee, W. J. 1897. Sheetflood McPherson. M. B. 1974. contribution. Jour. Geology
erosion. Geol. Soc. America Bull. Hydrological effects of 73:715-29.
8:87-112. urbanization. Paris: UNESCO Menard, H. W. 1961. Some rates of
McGinnis, L. D. 1966. Crustal Press. regional erosion. Jour. Geology
tectonics and Precambrian Meade, R. H. 1969. Errors in using 69:155-61.
basement in northeastern Illinois. modern stream-load data to Menzies. J. 1979. The mechanics of
Illinois Geol. Survey Rept. of Inv. estimate natural rates of drumlin formation with particular
219. denudation. Geol. Soc. America reference to the change in pore-
McGowen, J. H., and Garner,
H. J. Bull. 80:1265-74. water content of the till. Jour.
1970. Physiographic features and . 1982. Sources, sinks and Glaciol. 22:373-84.
stratification types of coarse- storage of river sediment in the 1981. Temperatures within
.
grained point bars. Modern and Atlantic drainage of the United subglacial debris —
A gap in our
ancient examples. Sedimentology States. Jour. Geology 90:235-52. knowledge. Geology 9:271-73.
14:77-111. Meier, M. F. 1960. Mode of flow of Meyer-Peter. E.. and Muller. R.
McGreevey, J. P. 1981. Some Saskatchewan glacier. Alberta, 1948. Formulas for bed-load
perspectives on frost shattering. Canada. U.S. Geol. Survey Prof. transport. Intnal. Assoc, for Hydr.
Proc. in Phys. Geog. 5:56-75. Paper 351. Structures Res. Proc. 2nd
McHattie, R. L., and Esch, D. C. Meier, M. F., and Johnson, A. 1962. Meeting. Stockholm, pp. 39-65.
1983. Benefits of a peat underlay The kinematic wave on Nisqually Mickelson, D. M. 1973. Nature and
used in road construction on Glacier, Washington. Jour. rate of basal till deposition in a
permafrost. In Proc. Permafrost Geophys. Research 67:886. stagnating ice mass. Burroughs
4th Internal. Conf., pp. 826-31. Meier, M. F.; Kamb, W. B.; Allen, Glacier, Alaska. Arc. Alp. Res.
Natl. Acad. Sci. C. R.; and Sharp. R. P. 1974. 5:17-27.
McKee, E. D. 1966. Structures of Flow of Blue Glacier, Olympic Mickelson. D. M.. and Berkson.
dunes at White Sands National Mountains, Washington, U.S.A. J. M. 1974. Till ridges presenth
Monument, New Mexico, and a Jour. Glaciol. 13:187-212. forming above and below sea level
comparison with structures of Meier,M. F., and Post, A. S. 1969. in Wachusett Inlet, Glacier Ba\.
dunes from other selected areas. What are glacier surges? Can. J. Alaska. Geogr. Annlr.
Sedimentology 7:1-69. Earth 6:807-17.
Sci. 56A;1 11-19.
. 1979. Introduction to a Meier, M. and Tangborn, W. V.
F.. Mielenz. R.. and King. M. 1955.
study of global sand seas. In A 1965. Net budget and flow of Physical-chemical properties and
study of global sand seas, edited South Cascade Glacier, engineering performance of clays,
by E. McKee. U.S. Geol. Survey Washington. Jour. Glaciol. California Div. of Mines Bull.
Prof. Paper 1052:1-20. 5:547-66. 169:196-254
a
553 Bibliography
Milanovic, P. 1981. Karst hydrology. . 1970. A glossary of karst Morner, N. A. 1980. Earth rheology,
Littleton, Colo.: Water Resources terminology. U.S. Geol. Survey isostasy, and eustasy. New York:
Publications. Water Supply Paper 1899 K. John Wiley & Sons.
Miller, J. P. 1958. High mountain . 1976. The karst landforms Morris, E. M. 1976. An experimental
streams; effects of geology on of Puerto Rico. U.S. Geol. Survey study of the motion of ice past
channel characteristics and bed Prof. Paper 899. obstacles by the process of
material. New Mexico State Bur. Moon, W. 1980. On the expected
J. regelation. Jour. Glaciol.
Mines and Min. Res. Memo 4. diameter of random channel 17:79-98.
Miller, R. D. 1966. Phase equilibria networks. Water Resour. Res. Morton, R. A., and Donaldson, A. C.
and soil freezing. Proc. 16:1119-20. 1978. Hydrology, morphology, and
Permafrost Internal. Conf. Moore, G., ed. 1960. Origin of sedimentology of the Guadalupe
(Lafayette, Ind., 1963). Natl. limestone caves: A symposium fluvial-deltaic system. Geol. Soc.
Acad. Sci.-Natl. Res. Council with discussion. Natl. Speleol. America Bull. 89:1030-36.
Pub. 1287, pp. 193-97. Soc. Bull. 22. Mosley, M. P. 1979. Streamflow
Millette, J. F. G., and Higbee, H. W. Moore, J. 1970. Relationship generation in a forested
1958. Periglacial loess, I. between subsidence and volcanic watershed, New Zealand. Water
Morphological properties. Am. load, Hawaii. Bull. Volcanol. Resour. Res. 15:795.
Jour. Sci. 256:284-93. 34:562-76. . 1982. The effect of a New
Milliman, J. and Meade, R. H.
D., Moore, T. R. 1979. Land use and Zealand beech forest canopy on
1983. World-wide delivery of river erosion in the Machakos Hills. the kinetic energy of water drops
sediment to the oceans. Jour. Ann. Assoc. Am. Geographers, and on surface erosion. Earth
Geology 91:1-22. 69:419-31. Surf. Proc. and Landforms
Mills, H. C, and Wells, P. D. 1974. Moran, S.; Clayton, L.; Hooke, 7:103-7.
Ice-shove deformation and glacial R. LeB.; Fenton, M.; and Moss, J. H., and Bonini, W. 1961.
stratigraphy of Port Washington, Andriashek, L. 1980. Glacier-bed Seismic evidence supporting a
Long Island, New York. Geol. landforms of the prairie region of new interpretation of the Cody
Soc. America Bull. 85:357-64. North America. Jour. Glaciol. terrace near Cody, Wyo. Geol.
Mills, H. H. 1977. Textural 25:457-76. Soc. America Bull. 72:547-56.
characteristics of drift from some Morehouse, D. F. 1968. Cave Moss, J. H., and Kochel, R. C. 1978.
representative Cordilleran development via the sulfuric acid Unexpected geomorphic effects of
glaciers. Geol. Soc. America Bull. reactions. Natl. Speleol. Soc. the hurricane Agnes storm and
88:1135-48. Bull. 30:1-10. flood, Conestoga drainage basin,
. 1980. An analysis of Morey, G.; Fournier, R.; and Rowe, southeastern Pennsylvania. Jour.
drumlin forms in the northeastern J. 1962. The solubility of quartz Geology 86:1-11.
and north-central United States. in water in the temperature Mueller, J. E. 1972. Re-evaluation of
Geol. Soc. America Bull. interval from 25 °C to 300° C. the relationship of master streams
91:2214-89. Geochim. et Cosmochim. Acta and drainage basins. Geol. Soc.
-. 1981. Boulder deposits and 26:1029-43. America Bull. 83:3471-74.
the retreat of mountain slopes or . 1964. The solubility of Mugridge, S. J., and Young, H. R.
"Gully Gravure" revisited. Jour, amorphous silica at 25 °C. Jour. 1983. Disintegration of shale by
of Geology 89:649-60. Geophys. Research cyclic wetting and drying and
Mills, H., and Starnes, D. 1983. 69:1995-2002. frost action. Can. J. Earth Sci.
Sinkhole morphometry in a Morgan, J. P. 1970. Deltas— 20:568-76.

fluviokarst region: Eastern resume. Jour. Geol. Educ. Muir Wood, A. M. 1969. Coastal
Highland Rim, Tennessee, U.S.A. 18:107-17. hydraulics. London: Macmillan.
Zeitf. Geomorph. 27:39-54. Morisawa, M. E. 1962. Quantitative Mukerji, A. B. 1976. Terminal fans
Miotke, F. D. 1973. The subsidence geomorphology of some of inland streams in Sutlej-
of the surfaces between mogotes watersheds in the Appalachian Yamuna plain, India. Zeit.f
in Puerto Rico east of Arecibo. Plateau. Geol. Soc. America Bull. Geomorph. 20:190-204.
Caves and Karst 15:1-12. 73:1025-46. Mullenders, W., and Gullentops, F.
Monroe, W. H. 1969. Evidence of 1964. Development of
. 1969. The age of the pingos of
subterranean sheet solution under drainage systems on an upraised Belgium. In The periglacial
weathered detrital cover in Puerto lake floor. Am. Jour. Sci. environment, edited by T. Pewe.
Rico. In Problems of karst 262:340-54. Montreal: McGill-Queens Univ.
denudation. Internat. Speleol. 1968. Streams, their
. Press.
Cong., 5th, Stuttgart. dynamics and morphology. New

York: McGraw-Hill.
Bibliography 554
Mullen E. H. 1974. Origin of Norris, R. M. 1966. Barchan dunes . 1976. Catenas in different
drumlins. In Glacial of Imperial Valley, California. climates. In Geomorphology and
geomorphology, edited by D. R. Jour. Geology 74:292-306. climate, edited by E. Derbyshire,
Coates, pp. 187-204. S.U.N.Y., Norrman, J. O. 1980. Coastal erosion London: John Wiley & Sons.
Binghamton: Pubs, in and slope development in Surtsey Olyphant, G. 1981a. Allometry and
Geomorphology, 5th Ann. Island. Zeit.f. Geomorph., Suppl. cirque evolution. Geol. Soc.
Symposium. 34:20-38. America Bull. 92:697-85.
Miiller, F. 1962. Zonation in the Nunn, K. R., and Rowell, D. M. . 1981b. Interaction among
accumulation areas of the glaciers 1967. Regelation experiments controls of cirque development:
of Axel Heiberg Island, N.W.T., with wires. Philos. Mag. Sangre Cristo Mountains,
Canada. Jour. Glaciol. 4:302-11. 16:1281-83. Colorado, U.S.A. Jour. Glaciol.
1963. Observations on
. Nye, J. F. 1952a. The mechanics of 27:449-58.
pingos (Beobachtungen uber glacier flow. Jour. Glaciol. Orme. A. R. 1973. Barrier and
Pingos). Can. Natl. Res. Council 2:82-93. lagoon systems along the
Tech. Translation 1073. . 1952b. A
comparison Zululand coast, South Africa. In
Muller, S.W. 1947. Permafrost or between the theoretical and the Coastal geomorphology, edited by
permanently frozen ground and measured long profiles of the DR. Coates. pp. 181-217.
related engineering problems. Unteraar glacier. Jour. Glaciol. S.U.N.Y.. Binghamton: 3rd Ann.
Ann Arbor, Mich.: J. W. 2:103-7. Geomorph. Symposium.
Edwards. . 1957. The distribution of . 1974. Quaternary
Munk, W. H. 1949. Surf beats. Am. stress and velocity in glaciers and deformation of marine terraces
Geophys. Union Trans. ice-sheets. Proc. Royal Soc. between Ensenada and El
30:849-54. London, ser.A:239:l 13-33. Rosario, Baja California, Mexico.
Murphey, J. B.; Wallace, D. E.; and -. 1960. The response of In Geology of peninsular
Lane, L. J. 1977. Geomorphic glaciersand ice-sheets to seasonal California, Pacific sections,
parameters predict hydrograph and climatic changes. Proc. Royal pp. 67-79. AAPG, SEPM, and
characteristics in the southwest. Soc. London, ser. A:256:559-84. SEG.
Water Resour. Res. Bull. 1965. The flow of a glacier Osborn, G. D. 1975. Advancing rock
13:25-38. in a channel of rectangular, glaciers in the Lake Louise area,
Myrick, R. M., and Leopold, L. B. elliptic, or parabolic cross-section. Banff National Park, Alberta.
1963. Hydraulic geometry of a Jour. Glaciol. 5:661-90. Can. J. Earth Sci. 12:1060-62.
small tidal estuary. U.S. Geol. Nye, J. F, and Martin, P. C. S. 1967. Osterkamp, W. R. 1978. Gradient,
Survey Prof. Paper 422-B. Glacial erosion, pp. 78-83. Int. discharge and particle-size
Nanson. G. C, and Young, R. W., Assoc. Sci. Hydrol., Comm. Snow relations of alluvial channels in
1981. Overbank deposition and and Ice. Bern. Kansas, with observations on
floodplain formation on small Oberlander, T.M. 1972. braiding. Am. Jour. Sci.
coastal streams of New South Morphogenesis of granitic boulder 278:1253-68.
Wales. Zeit.f. Geomorph. slopes in the Mojave Desert, Ostrem. G. 1964. Ice-cored moraines
25:332-45. California. Jour. Geology in Scandinavia. Geogr.
Nash, D. 1980a. Forms of bluffs 80:1-20. Annlr.46:2S2-331.
degraded for different lengths of 1974. Landscape inheritance
.
Outcalt. S. I., and Benedict, J. B.
time in Emmet County, Michigan, and the pediment problem in the 1965. Photo-interpretation of two
U.S.A. Earth Surf. Proc. and Mojave Desert of southern types of rock glaciers in the
Landforms 5:331-45. California. Am. Jour. Sci. Colorado Front Range, U.S.A.
— 1980b. Morphologic dating
. 274:849-75. Jour. Glaciol. 5:849-56.
of degraded normal fault scarps. and Drake, C. L. 1982.
Officer. C. B., Owens, E. H., and Harper, J. R.
Jour. Geology 88:353-60. Epeirogenic plate movements. 1977. Frost-table and thaw depths
National Academy of Sciences. 1983. Jour. Geology 90:139-54. in the littoral zone near Pearl Bay.
Proc. Permafrost 4th Internat. Oilier, C. D. 1963. Insolation Alaska. Arctic 30:155-68.
Conf. Washington, D.C.: National weathering: Examples from Owens. L. B.. and Watson. J. P.
Academy Press. central Australia. Am. Jour. Sci. 1979. Rates of weathering and
National Research Council of 261:376-81. soil formation on granite in
Canada. 1978. Proc. Permafrost
3rd Internat. Conf. Ottawa, Ont.:
— 1969. Weathering.
. Rhodesia. Soil Sci. Soc. Am.
Edinburgh: Oliver and Boyd. Proc. 43:160-66.
National Research Council of
Canada.
A
555 Bibliography
Paige, S. 1912. Rock-cut surfaces in Parizek, R. 1969. Glacial ice-contact Peltier, L. 1950. The geographical
the desert ranges. Jour. Geology ringsand ridges. Geol. Soc. cycle in periglacial regions as it is
20:442-50. America Spec. Paper 123:49-102. related to climatic geomorphology.

Pakiser, L. C, and Robinson, R. Park, C. 1977. World-wide variations Ann. Assoc. Am. Geog.
1966. Composition of the in hydraulic geometry exponents 40:214-36.
continental crust as estimated of streams channels: An analysis Perutz, M. F. 1940. Mechanism of
from seismic observations. In The and some observations. J. Hydro!. glacier flow. Proc. Royal Soc.
Earth beneath the continents, 33:133-46. London, ser. A:52: 132-35.
edited by J. Steinhart and T. Parker, G. 1976. On the cause and Pesci, M. 1968. Loess. In
Smith, pp. 620-26. Am. Geophys. characteristic scale of meandering Encyclopedia of geomorphology,
Union, Geophys. Monograph 10. and braiding in rivers. Jour. Fluid edited by R. W. Fairbridge. New
Palmer, A. C. 1972. A kinematic Mech. 76:459-80. York: Reinhold Book Corp.
wave model of glacier surges. Paterson, W. S. B. 1964. Variations Petrie, G., and Price, R. J. 1966.
Jour. Glaciol. 11:65-72. in velocity of Athabasca Glacier Photogrammetric measurements
Palmer, A. N. 1975. Origin of maze with time. Jour. Glaciol. of the ice wastage and
caves. Natl. Speleol. Soc. Bull. 5:277-85. morphological changes near the
37:57-76. . The physics of
1969. Casement Glacier, Alaska. Can.
. 1981a. Hydrochemical glaciers. Oxford: Pergamon Press. J. Earth Sci. 3:827-40.
factors in the origin of limestone 1981. The physics of Petterssen, S. 1964. Meteorology. In
caves. Proc. 8th Intl. Cong. glaciers, 2d ed. Oxford: Pergamon Handbook of applied hydrology,
Speleol., Bowling Green, Ky., Press. edited by V. T. Chow, pp. 3-39.
pp. 120-22. Patton, H. B. 1910. Rockstreams of New York: McGraw-Hill.
1981b. A geological guide Veta Park, Colorado. Geol. Soc. Pettijohn, F. 1941. Persistence of
to Mammoth Cave National America Bull. 22:663-76. heavy minerals and geologic age.
Park. Teaneck, N.J.: Zephyrus Patton, P. C, and Baker, V. R. 1976. Jour. Geology 49:610-25.
Press. Morphometry and floods in small Pewe, T. L. 1955. Origin of the
. 1984. Recent trends in karst drainage basins subject to diverse upland silt near Fairbanks,
geomorphology. Jour. Geol. Educ. hydrogeomorphic controls. Water Alaska. Geol. Soc. America Bull.
32:247-53. Resour. Res. 12:941-52. 66:699-724.
Palmer, V. E., and Palmer, A. N. 1977. Geomorphic response
. . 1959. Sand-wedge polygons
1975. Landform development of of central Texas stream channels (Tesselations) in the McMurdo
the Mitchell Plain of southern to catastrophic rainfall and runoff. Sound region, Antarctica —
Indiana: Origin of a partially In Geomorphology of arid and progress report. Am. Jour. Sci.
karstic plain. Zeit.f. Geomorph. semiarid regions, edited by D. O. 257:545-52.
19:1-39. Doehring, pp. 189-217. S.U.N.Y., 1966. Ice-wedges in
Palmquist, R. C. 1975. Preferred Binghamton: Pubs, in —
Alaska Classification,
position model and subsurface Geomorphology. distribution and climatic
symmetry of valleys. Geol. Soc. Patton, P. C; Baker, V. R.; and significance. In Proc. Permafrost
America Bull. 86:1391-98. Kochel, R. C. 1979. Slack-water Internal. Conf. (Lafayette, Ind.,
. 1979. Geological controls on deposits: A geomorphic technique 1963). Natl. Acad. Sci.-Natl. Res.
doline characteristics in mantled for the interpretation of fluvial Council Pub. 1287, pp. 76-81.
karst. Zeit.f. Geomorph., Suppl. paleohydrology. In Adjustments -. 1969. The periglacial
32:90-106. of the fluvial system, edited by environment. Montreal: McGill-
Panos, V., and Stelcl, O. 1968. D. D. Rhodes and G. P. Williams, Queens Univ. Press.
Physiographic and geologic pp. 225-53: Dubuque, Iowa: -. 1973. Ice wedge casts and
control in development of Cuban Kendall Hunt. past permafrost distribution in
mogotes. Zeit. f. Geomorph. Patton, P. C, and Dibble, D. S. 1982. North America. Geoform
12:117-65. Archeologic and geomorphic 15:15-26.
Paredes, J. R. and Buol, S. W. 1981. evidence for the paleohydrologic 1981. Desert dust: Origin,
Soils in an aridic, ustic, udic, record of the Pecos River in west characteristics, and effect on man.
climosequence in the Maracaibo Texas. Am. Jour. Sci. Geol. Soc. America Spec. Paper
Lake Basin, Venezuela. Soil Sci. 282:97-121. 186.
Soc. Am. Proc. 45:385-91. Patton, P.C, and Schumm, S. A. Pewe, T. L., and Journaux, A. 1983.
Parham, W. E., 1969. Formation of 1975. Gulley erosion, Origin and character of loesslike
halloysite from feldspar: Low northwestern Colorado: A silt in unglaciated south-central
temperature, artificial weathering threshold phenomenon. Geology Yakutia, Siberia, U.S.S.R. U.S.
versus natural weathering. Clays 3:88-90. Geol. Surv. Prof. Paper 1262.
and Clay Minerals 17:13-22.
Bibliography 556
Peyronnin, C. A., Jr. 1962. Erosion of Pohl, E. R. 1955. Vertical shafts in Potter, N., Jr., and Moss. J. H. 1968.
Isles Dernieres and Timbalier limestone caves. Natl. Speleol. Origin of the Blue Rocks block
Islands. Am. Soc. Civil Engineers. Soc. Occasional Paper 2. field deposits. Berks County,
Jour. Waterways and Harbors, Pohl, E. R., and White, W. B. 1965. Pennsvlvania. Geol. Soc. America
1:57-69. Sulfate minerals: Their origin in Bull. 79:255-62.
Phillip,H. 1920. Geologische the central Kentucky karst. Am. Potts, A. S. 1970. Frost action in
Untersuchungen iiber den Mineralogist 50:1461-65. rocks: Some experimental data.
Mechanismus der Gletscher Poldervaart, A., ed. 1955. Chemistry Inst. Brit. Geog. Trans.
Bewegung und die Entstehung der of the Earth's crust. In Crust of 49:109-24.
gletschertextur. Neuer Jb. Miner. Earth. Geol. Soc. America Spec. Powers. R. W. 1962. Arabian Upper
Geol. Palaont. 43:439-556. Paper 62:119-44. Jurassic carbonate reservoir rocks.
Pickup, G. 1977. Simulation Pomeroy, J. S. 1980. Storm-induced In Classification of carbonate
modelling of river channel erosion. debris avalanching and related rocks, edited by W. E. Ham. Am.
In K. J. Gregory. River channel phenomena in the Johnstown Assoc. Petroleum Geologists Mem
changes, pp. 47-60. London: John area, Pennsylvania, with 1:122-92.
Wiley & Sons. references to other studies in Prestegaard. K. L. 1983a. Bar
Pierce, J. W. 1970. Tidal and
inlets Appalachians. U.S. Geol. Survev resistance in gravel bed streams at
washover fans. Jour. Geology Prof. Paper 1191. bankfull stage. Water Re sour.
78:230-34. .
1982. Landslides in the Res. 19:472-76.
Pike, R. J., and Wilson, S. E. 1971. greater Pittsburgh region, 1983b. Variables influencing
.
Elevation-relief ratio, hypsometric Pennsylvania. U.S. Geol. Survev water-surface slopes in gravel-bed
integral, and geomorphic area- Prof. Paper 1229. streams at bankfull stage. Geol.
altitude analysis. Geol. Soc. Porslid, A. E. 1938. Earth mounds in Soc. America Bull. 94:673-78.
America Bull. 82:1079-84. unglaciated Arctic northwestern Price. L. W. 1972. The periglacial
Pillans. B. 1983. Upper Quaternary America. Geogr. Rev. 28:46-58. environment, permafrost, and
marine terrace chronology and Porter, S. C. 1977. Present and past man. Assoc. Am. Geog., Comm.
deformation. South Taranaki, glaciation threshold in the on College Geog. Resource Paper
New Zealand. Geology Cascade Range, Washington, 14.
11:292-97. U.S.A. Topographic and climatic Price, R. J. 1966. Eskers near the
Piper, A. M. 1932. Ground water in controls, and paleoclimatic Casement glacier. Alaska. Geogr.
north-central Tennessee. U.S. implications. Jour. Glaciol. Annlr. 48:111-25.
Geol. Survey Water Supply Paper 18:101-16. 1969. Moraines, sandar.
.
640:69-89. Porter. S. C, andOrombelli, G. kames, and eskers near

Pissart, A. 1970. The pingos of 1980. Catastrophic rockfall of Breidamerkurjokull. Iceland. Inst.
Prince Patrick Island September 12, 1717. on the Brit. Geog. Trans. 46:17-43.
(76°N-120°W). Natl. Res. Italian flank of the Mont Blanc . 1970. Moraines at
Council of Canada, Tech. Trans. Massif. Zeit f Geomorph. Fjallsjokull. Iceland. Arc. Alp.
1401. 24:200-18. Res. 2:27-42.
Plummer. L. 1975. Mixing of Post, A. S. 1960. The exceptional . 1973. Glacial and
seawater with calcium carbonate advances of the Muldrow, Black fiuvioglacial landforms. New
groundwater. In Quantitative Rapids and Sustina glaciers. Jour. York: Hafner.
studies in the geological sciences, Geophys. Research 65:3703-12. Pritchard. G. B. 1962. Inuvik,
edited by E. Whitten, pp. 219-36. . The recent surge of
1966. Canada's new Arctic town. Polar
Geol. Soc. America Mem. 142. Walsh glacier, Yukon and Alaska. Record 11:71:145-54.
Plummer, L.; Vacher, H.; Mackenzie, Jour. Glaciol. 6:375-81. Ragan. R. M. 1968. An experimental
F.; and Land. L.
Bricker, O.; . 1967. Effects of the March investigation of partial area
1976. Hydrochemistry of 1964 Alaskan earthquake on contributions. Intl. Assoc. Sci.
Bermuda: A case history of Survey Prof.
glaciers. U.S. Geol. Hvdrol. Pub. 76:241-51.
groundwater diagenesis of Paper 544-D. Rahn. P. H. 1966. Inselbergs and
biocalcarenites. Geol. Soc. 1969. Distribution of surging nickpoints in southwestern
America Bull. 87:301-16. glaciers in western North Arizona. Zeit.f. Geomorph.
Plummer, N.; Wigley, T: and America. Jour. Glaciol. 8:229-40. 10:217-25
Parkhurst. D. 1978. The kinetics Potter, N., Jr. 1972. Ice-cored rock . 1967. Sheetfloods.
of calcite dissolution in CO\ water glacier.Galena Creek, northern streamfloods. and the formation of
systems at 5° to 60°C and 0.0 to Absaroka Mountains. Wyoming. pediments. Ann. Assoc. Am.
1 atm. C0 2 Am. Jour. Sci.
. Geol. Soc. America Bull. Geog. 57:593-604.
278:179 216. 83:3025-57.
557 Bibliography
. 1976. Coulee alignment and Reheis. M. J. 1975. Source Richardson, H. W. 1942. Alcan-
the wind in southern Alberta. transportation and deposition of America's glory road, parts I and
Canada: Discussion. Geol. Soc. debris on Arapaho Glacier. Front II. Engr. News Record
America Bull. 87:157. Range. Colorado, U.S.A., Jour. 129:25:81-96 and 27:35-42.
Rains. R.. and Shaw, J. 1981. Some Glaciol. 14:407-20. . 1943. Alcan-America's glory
mechanisms of controlled moraine Reiche. P. 1943. Graphic road, part III. Engr. News Record
development. Antarctica. Jour. representation of chemical 130:1:131-38.
Glaciol 27:113-28. weathering. Jour. Sed. Petrology . 1944. Controversial Canol.
Rankin. J. K. 1952. Development of 13:58-68. Engr. News Record 132:2:78-84.
the New
Jersey shore. In 3rd Reid. Layman, J.; and Frostick,
I.; L. Rieke. R. D.: Vinson, T. S.; and
Coastal Engr. Conf. Proc, edited 1980. The continuous Mageau. D. W. 1983. The role of
by J. W. Johnson, pp. 306-17. measurement of bedload specific surface area and related
Cambridge. Mass.: Council of discharge. Jour. Hydraul. Res. index properties in the frost heave
Wave Research. 18:243-49. susceptibility of soils. In Proc.
Rapp. A. 1960. Recent development Reid. M.: MacLeod. D. A.: and
J. Permafrost 4th Internat. Conf,
of mountain slopes in Karkevagge Cresser, M. 1981. The assessment pp. 1066-71. Natl. Acad. Sci.
and surroundings, northern of chemical weathering rates Rigsby. G. P. 1960. Crystal
Scandinavia. Geogr. Annlr. within an upland catchment in orientation in glacier and in
42:65-206. North-East Scotland. Earth Surf. experimentally deformed ice.
Raudkivi. A. J. 1967. Loose Proc. and Landforms 6:447-57. Jour. Glaciol' 3:589-606.
boundary hydraulics. Oxford: Rendell, H. 1982. Clay hillslope Ritter. D. F. 1967a. Terrace
Pergamon Press. erosion rates in Basento Valley, development along the front of the
Rawitz. E.: Engman, E. T.; and S. Italy, Geogr. Annlr. Beartooth Mountains, southern
Cline. G. D. 1970. Use of the 64A:141-47. Montana. Geol. Soc. America
mass balance method for Renwick, W.; Brumbaugh. R.: and Bull. 78:467-84.
examining the role of soils in Loeher, L. 1982. Landslide .
1967b. Rates of denudation.
controlling watershed morphology and processes on Jour. Geol. Educ. 15, C.E.G.S.
performance. Water Resour. Res. Santa Cruz Island, California: short rev. 6:154-59.
6:115-23. Geogr. Annlr. 64A: 149-59. -. 1972. The significance of
Ray. L. L. 1951. Permafrost. Arctic Revue de geomorphologie stream capture in the evolution of
4:196-203. dynamique. 1967. Field methods a piedmont region, southern
Rav. R. J.; Krantz. W. B.: Caine, for the study of slope and fluvial Montana. Zeit.f. Geomorph.
T. N.; and Gunn. R. D. 1983a. A processes. Rev geomorph. 16:83-92.
mathematical model for patterned dxnamique 17:145-88. -. 1975. Stratigraphic
ground: Sorted polygons and Rhodes. D. D. 1977. The b-f-m implications of coarse-grained
stripes, and underwater polygons. diagram: Graphical representation gravel deposited as overbank
In Proc. Permafrost 4th Internat. and interpretation of at-a-station sediment, southern Illinois. Jour.
Conf.. pp. 1036-41. Natl. Acad. hydraulic geometry. Am. Jour. Geology 83:645-50.
Sci. Sci. 277:73-96. -. 1982. Complex river terrace
1983b. A model for sorted
. Rice, A. 1976, Insolation warmed development in the Nenana Valley
patterned ground regularity. Jour. over. Geology, pp. 6 1 -62. near Healy, Alaska. Geol. Soc.
Glaciol. 29:317-37. Rich. J. and
L. 1935. Origin America Bull. 93:346-56.
Raymond. C. E 1971. Flow in a evolution of rock fans and W. F; and
Ritter. D. F.; Kinsey,
transverse section of Athabasca pediments. Geol. Soc. America Kauffman, M. E. 1973. Overbank
Glacier. Alberta. Canada. Jour. Bull. 46:999-1024. sedimentation in the Delaware
Glaciol. 10:55-84. Richards, K. S. 1976a. Channel River valley during the last 6,000
Reed, B.: Galvin. C. J.: and Miller. width and the riffle-pool sequence. years. Science 179:374-75.
J. P. 1962. Some aspects of Geol. Soc. America Bull. Rittman, A. 1962. Volcanoes and
drumlin geometry. Am. Jour. Sci. 87:883-90. their activity. New York: Wilev-
260:200-210. 1976b. The morphology of

. Interscience.
Reeve. I. J. 1982. A splash transport sequences. Earth Surf.
riffle-pool Robin. G. deQ. 1976. Is the basal ice
model and its application to Proc. and Landforms 1:71-88. of a temperate glacier at the
geomorphic measurement. Zeit.f. 1979. Prediction of drainage pressure-melting point? Jour.
Geomorph. 26:55-71. density from surrogate measures. Glaciol. 16:183-95.
Water Resour. Res. 15:435-42. Robin, G. deQ., and Weertman, J.
. 1982. Rivers. London: 1973. Cyclic surging of glaciers.
Methuen and Co. Jour. Glaciol. 12:3-18.
.
Bibliography 558
Robinson, G. 1966. Some residual . 1975. Geomorphology. St. Amand, P. 1957. Geological and
hillslopes in the Great Fish River Boston: Houghton Mifflin Co. geophysical synthesis of the
Basin, South Africa. Geogr. Jour. Runnells, D. D. 1969. Diagenesis, tectonics of portions of British
132:386-90. chemical sediments, and the Columbia, the Yukon Territory,
Robinson, L. A. 1977. Marine erosive mixing of natural waters. Jour. and Alaska. Geol. Soc. America
processes at the cliff foot. Marine Sed. Petrology 39: 88- 201 1 1 1 Bull. 68:1343-70.
Geol. 23:257-71. Russell, R. 1943. Freeze-thaw Sakamoto-Arnold, C. M. 1981.
Ronov, A., and Yaroshevsky, A. frequencies in the United States. Eolian features produced by the
1969. Chemical composition of Am. Geophys. Union Trans. December 1977 windstorm in
the Earth's crust. In The Earth's 24:125-33. southern San Joaquin Valley,
crust and upper mantle, edited by Russell, R. J. 1967a. Aspects of California. Jour. Geology
P. Hart, pp. 37-57. Am. Geophys. coastal morphology. Geogr. Annlr. 89:129-37.
Union, Geophys. Monograph 13. 49A:299-309. Sallenger, A. D. 1979. Beach-cusp
Rossby, C. G. 1941. The scientific 1967b. River and delta
. formation. Marine Geol.
modern meteorology.
basis of In morphology. Louisiana State 29:23-37.
Climate and man, U.S. Dept. Univ., Coastal Studies Inst. Tech. Salter, P., and Williams, J. 1965.
Agri. Yearbook, pp. 599-655. Rept. 52. The influence of texture on the
Rothlisberger, H. 1972. Water Russell, R. J., and Mclntire. W. G. moisture characteristics of soils, I.
pressure in intra- and sub-glacial 1965. Beach cusps. Geol. Soc. A comparison of
critical
channels. Jour. Glaciol. America Bull. 76:307-20. techniques for determining the
11:177-203. Russell-Head, D. S., and Budd, W. F. available-water capacity and
Rouse, L. J., Jr.; Roberts, H. H.; and 1979. Ice-sheet flow properties moisture characteristic curve of a
Cunningham, R. 1978. Satellite derived from bore-hole shear soil. Jour. Soil Sci. 16:1-5.
observation of subaerial growth of measurements combined with ice- Saucier, R. T, and Fleetwood, A. R.
the Atchafalaya Delta, Louisiana. core studies. Jour. Glaciol. 1970. Origin and chronologic
Geology 6:405-8. 24:117-30. significance of Late Quaternary
Rubey, W. W. 1938. The force Rust, B. R. 1972. Structure and Ouachita River,
terraces,
required to move particles on a process in a braided river. Arkansas and Louisiana. Geol.
stream bed. U.S. Geol. Survey Sedimentology 18:221-45. Soc. America Bull. 81:869-90.
Prof. Paper 189-E. . 1977. Mass flow deposits in Savage, J. C, and Paterson, W. S. B.
1952. Geology and mineral
.
a Quaternary succession near 1963. Borehole measurements in
resources of the Hardin and Ottawa, Canada: Diagnostic the Athabasca Glacier. Jour.
Brussels quadrangles (in Illinois). criteria for subaqueous outwash. Geophys. Research 68:4521-36.
U.S. Geol. Survey Prof. Paper Can. J. Earth Sci. 14:175-84. Savat. J. 1981. Work done by splash:
218. Ruxton, B. P., and Berry, L. 1961. Laboratory experiments. Earth
Ruhe, R. V. 1952. Topographic Weathering profiles and Surf. Proc. and Landforms
discontinuities of the Des Moines geomorphic position on granite in 6:275-83.
lobe. Am. Jour. Sci. 250:46-56. two tropical regions. Rev. Savigear, R. 1952. Some observations
1964. Landscape
.
ge'omorph. dynamique 12:16-31. on slope development in South
morphology and alluvial deposits Ruxton, B. P., and McDougall, I. Wales. Inst. Brit. Geog. Trans.
in southern New Mexico. Ann. 1967. Denudation rates in 18:31-51.
Assoc. Am. Geog. 54:147-59. northeast Papua from potassium- Sawkins, J. 1869. Report on the
1965. Quaternary
.
argon dating of lavas. Am. Jour. geology of America. Mem. Geol.
paleopedology. In The Quaternary Sci. 265:545-61. Survey
of the United States, edited by Ryckborst, H. 1975. On the origin of Schalscha, E.; Appelt, H.; and
'
H. Wright and D. Frey, pingos. J. Hydro/. 26:303-14. Schatz, A. 1967. Chelation as a
pp. 735-64. Princeton, N.J.: Ryder, J. M. 1971a. The stratigraphy weathering mechanism. I. Effect
Princeton Univ. Press. and morphology of paraglacial of complexing agents on the
1967. Geomorphic surfaces alluvial fans in south-central solubilization of iron from
and surficial deposits in southern British Columbia. Can. J. Earth minerals and granodiorite.
New Mexico. New Mexico State Sci. 8:279-98. Geochim. el Cosmochim. Acta
Bur. Mines and Min. Res. Memo — . 1971b. Some aspects of the 31:587-96.
18. morphometry of paraglacial Schatz. A. 1963. Chelation in
. 1969. Quaternary alluvial fans in south-central nutrition, soil microorganisms and
landscapes in Iowa Ames: Iowa British Columbia. Can. J. Earth soil chelation. The pedogenic
Slate Univ. Press. Sci. 8:1252-64. action of lichens and lichen acids.
Jour. Agri. and Food Chem.
11:112-18.
559 Bibliography
Schatz, A.; Cheronis, N.; Schatz, V.; . 1971. Fluvial Schumm, S. A., and Parker, R. S.
and Trelawney, G. 1954. geomorphology. In River 1973. Implications of complex
Chelation (sequestration) as a mechanics, edited by H. W. Shen. response of drainage systems for
biological weathering factor in chs. 4 and 5. Ft. Collins, Colo.: Quaternary alluvial stratigraphy.
pedogenesis. Pennsylvania Acad. Colorado State Univ. Nat. Phys. Sci. 243:99-100.
Sci. Proc. 28:44-57. -, ed. 1972. River morphology. Scott, A. J., and Fisher, W. L. 1969.
Schowengerdt, R. A., and Glass, Stroudsburg, Pa.: Dowden, Delta systems and deltaic
C. E. 1983. Digitally processed Hutchinson, and Ross. deposition. In Delta systems in
topographic data for regional Geomorphic
1973. the exploration for oil and gas,
tectonic evaluation. Geol. Soc. thresholds and complex response edited by W. Fisher, L. Brown, A.
America Bull. 94:549-56. of drainage systems. In Fluvial Scott, and J. McGowen. Univ.
Schumm, A. 1956. Evolution of
S. geomorphology, edited by Texas, Austin, Bureau Econ.
drainage systems and slopes in M. Morisawa, pp. 299-310. Geology.
badlands at Perth Amboy, New S.U.N.Y., Binghamton: Pubs, in Scott, K. M., and Gravlee, G. C, Jr.,
Jersey. Geol. Soc. America Bull. Geomorphology, 4th Ann. Mtg. 1968. Flood surge on the Rubicon
67:597-646. -. 1977. The fluvial system. River, California — Hydrology,
. 1960. The shape of alluvial New York: John Wiley & Sons. hydraulics, and boulder transport.
channels in relation to sediment 1980. Some applications of U.S. Geol. Survey Prof. Paper
type. U.S. Geol. Survey Prof. the concept of geomorphic 422-M.
Paper 352-B. thresholds. In Thresholds in Seed, H.; Woodward, R.; and
. 1962. Erosion of miniature geomorphology, edited by Lundgren, R. 1964. Clay
pediments in Badlands National D. Coates and J. Vitek, mineralogical aspects of the
Monument, South Dakota. Geol. pp. 473-86. London: Allen and Atterberg limits. Am. Soc. Civil
Soc. America Bull. 73:719-24. Unwin Ltd. Engineers Proc. 90:SM4: 107-31.
. 1963a. A tentative Schumm, S. A.; Bean, D. W.; and Selby, M. J. 1966. Methods of
classification of alluvial river Harvey, M. D. 1982. Bed-form- measuring soil creep. J. Hydrol.
channels. U.S. Geol. Survey Circ. dependent pulsating flow in 5:54-63.
477. Medano Creek, Southern . 1967. Aspects of the
. 1963b. Sinuosity of alluvial Colorado. Earth Surf. Proc. and geomorphology of the Greywacke
channels on the Great Plains. Landforms 7:17-28. ranges bordering the lower and
Geol. Soc. America Bull. Schumm, S. A., and Chorley, R. J. middle Waikato basins. Earth Sci.
74:1089-1100. 1964. The fallof Threatening Jour. 1:1-22.
. 1963c. Disparity between Rock. Am. Jour. Sci. . 1980. A rock mass strength
present rates of denudation and 262:1041-54. classification for geomorphic
orogeny. U.S. Geol. Survey Prof. . 1966. Talus weathering and purposes: With tests from
Paper 454-H. scarp recession in the Colorado Antarctica and New Zealand.
-. 1965. Quaternary Plateaus. Zeit.f Geomorph. Zeit.f. Geomorph. 24:31-51.
paleohydrology. In The 10:11-36. 1982. Hillslope materials
Quaternary of the United States, Schumm, S. A., and Hadley, R. F. and processes. New York: Oxford
edited by H. E. Wright and D. G. 1957. Arroyos and the semiarid Univ. Press.
Frey, pp. 783-94. Princeton, N.J.: cycle of erosion. Am. Jour. Sci. Senstius, M. 1958. Climax forms of
Princeton Univ. Press. 255:161-74. chemical rock-weathering. Am.
1967a. Rates of surficial
. Schumm, S. A., and Khan, H. R. Scientist 46:355-67.
rock creep on hillslopes in western 1972. Experimental study of Sevon, W. D. 1969. Sedimentology of
Colorado. Science 155:560-61. channel patterns. Geol. Soc. some Mississippian and
-. 1967b. Meander wavelength America Bull. 83:1755-70. Pleistocene deposits of
of alluvial rivers. Science Schumm, and Lichty, R. W.
S. A., northeastern Pennsylvania. In
157:1549-50. 1963. Channel widening and Geology of selected areas in New
1968. River adjustment to
. floodplain construction along Jersey and eastern Pennsylvania.
altered hydrologic regimen, Cimarron River in southwestern New Brunswick, N.J.: Rutgers
Murrumbidgee River and Kansas. U.S. Geol. Survey Prof. Univ. Press.
paleochannels, Australia. U.S. Paper 352-D. Sharp, R. P. 1940. Geomorphology of
Geol. Survey Prof. Paper 598. .1965. Time, space and the Ruby-East Humboldt Range,
-. 1969. River metamorphosis. causality in geomorphology. Am. Nevada. Geol. Soc. America Bull.
Am. Soc. Civil Engineers, Jour. Jour. Sci. 263:110-19. 51:337-72.
Hydraulics Div., HYl:255-73.
Bibliography 560
. 1951. Features of the firn on Shepard. F. P.; Emery. K. O.; and .

1974. Variation of
upper Seward Glacier. St. Elias LaFond. E. C. 1941. Rip currents: mainstream length with basin
Mountains. Canada. Jour. A process of geological area in river networks. Water
Geology 59:599-621. importance. Jour. Geology Resour. Res. 10:1167-77.
.1953. Deformation of a 49:337-69. Shumskii. P. A. 1964. Principles of
vertical bore hole in a piedmont Shepard. F. P., and Grant. U. S.. IV. structural glaciology. New York:
glacier. Jour. Glaciol. 2:182-84. 1947. Wave erosion along the Dover Publications.
1960. Glaciers. Eugene. southern California coast. Geol. Shuster, E. T., and White. W. B.
Ore.: Univ. Oregon Press. Soc. America Bull. 58:919-26. 1971. Seasonal fluctuations in the
1963. Wind ripples. Jour. Shepard, F. P., and Inman, D. L. chemistry of limestone springs: A
Geology 71 :617 -36. 1950. Nearshore circulation possible means for characterizing
1964. Wind-driven sand
. in related to bottom topography and carbonate aquifers. J. Hydrol.
Coachella Valley, California. wave refraction. Am. Geophvs. 14:93-128.
Geol. Soc. America Bull. Union Trans. 31:2:196-212. Sidle. R..and Swanston. D. 1982.
75:785-804. Shepard. and Wanless, H. R.
F. P.. Analysis of a small debris slide in
1966. Kelso Dunes. Mojave 1971. Our changing coastlines. coastal Alaska. Can. Geotechnical
Desert. California. Geol. Soc. New York: McGraw-Hill. Jour. 19:167-74.
America Bull. 77:1045-74. Shepherd. R. G.. and Schumm, S. A. Development of tower
Silar. J. 1965.
. 1979. Intradune flats of the 1974. Experimental study of river karst ofChina and North
Algodones chain. Imperial Valley. incision. Geol. Soc. America Bull. Vietnam. Natl. Speleol. Soc. Bull.
California. Geol. Soc. America 85:257-68. 27(2):35-46.
Bull. 90:908-16. Shlemon. R. J. 1975. Subaqueous Silvester. R. 1966. Wave refraction.
.1980. Wind-driven sand in delta formation — Atchafalaya In Encyclopedia of oceanography.
Coachella Valley. California: Bay, Louisiana. In Deltas, edited edited by R. W. Fairbridge.
Further data. Geol. Soc. America by M. Brousard, pp. 209-21. pp. 975-76. New York: Reinhold
Bull. 91:724-30. Houston: Houston Geologic Book Corp.
Sharp. R. P., and Noble. L. H. 1953. Society. Simmons. G.. and Richter. D. 1976.
Mudflow of 1941 at Wrightwood, Yamada, I.; and
Shoji. S.; Microcracks in rocks. In The
southern California. Geol. Soc. Kurashima, K. 1981. Mobilities Physics and chemistry of
America Bull. 64:547-60. and related factors of chemical minerals and rocks, edited by
Sharpe. C. F. S. 1938. Landslides elements in the topsoils of R.G.J. Strens. pp. 105-37.
and related phenomena. New andosols in Tohuku. Japan: 2. London: John Wiley & Sons.
York: Columbia Univ. Press. Chemical and mineralogical Simons, D. B.. and Richardson. E. V.
Shaw. J. 1972. Sedimentation in the compositions of size fractions and 1962. Resistance to flow in
ice-contact environment, with factors influencing the mobilities alluvial channels.Am. Soc. Ci\il
examples from Shropshire of major chemical elements. So/7 Engineers Trans. 127:927-52.
(England). Sedimentology Sci. 132:330-46. 1963. Forms of bed
.
18:23-62. Short. A. D. 1979. Three- roughness in alluvial channels.

. 1980. Drumlins and large- dimensional beach stage model. Am. Soc. Civil Engineers Trans.
scale flutings related to glacier Jour. Geology 87 :553-7 1. 128:284-302.
folds. Arc. Alp. Res. 12:287-98. Short, A. D.. and Wright. L. D. . 1966. Resistance to flow in
Shawe. D. R. 1963. Possible wind- 1981. Beach systems of the alluvial channels. U.S. Geol.
erosion origin of linear scarps on Sydnev Region. Aust. Geog. Survey Prof. Paper 422- J.
the Saga Plain, southwestern 15:8-l'6. Simpson. D. 1964. Exfoliation in the
Colorado. U.S. Geol. Survey Prof. Shreve. R. L. 1966a. Statistical law upper Pocahontas Sandstone.
Paper 475-C: 138-42. of stream numbers. Jour. Geology Mercer County. West Virginia.
Shepard. F. P. 1950. Beach cycles in 74:17-37. Am. Jour. Sci. 262:545-51.
southern California. U.S. Army — 1966b. Sherman landslide.
. Skempton. A. W. 1953. Soils
Corps Fngrs.. Beach Erosion Alaska. Science 154:1639-43. mechanics in relation to geology
Board Tech. Memo 20. -. 1967. Infinite topological!} )'orkshire Geol. Soc. Proc.
1952. Revised nomenclature random channel networks. Jour. 29:33-62.
for depositional coastal features. Geology 75:178-86. . 1964. The long-term
Am. Assoc Petroleum Geologists — 1968. The Black hawk
. stability of clay slopes.
7 5-102
Bull. 36:1902 12. landslide. Geol. Soc. America Geotechnique 2:
. 1963. Submarine geology. Spec. Paper 108.
New York: Harper & Row.
561 Bibliography
Small, R.; Clark, M.; and Cawse, Smith, H. T

U. 1948. Giant glacial Stalker, A. MacS. 1960. Ice-pressed
T. J. 1979. The formation of grooves northwest Canada. Am.
in drift forms and associated deposits
medial moraines on Alpine Jour. Sci. 246:503-14. in Alberta. Canada Geol. Survey
glaciers. Jour. Glaciol. 22:43-52. . The Hickory Run
1953. Bull. 57.
Smalley, I.J. 1966. Drumlin boulder field, Carbon County, Stanley, D. J.; Krinitzsky, E. L.: and
formation. A rheological model. Pennsylvania. Am. Jour. Sci. Compton, J. R. 1966. Mississippi
Science 151:1379. 251:625-42. River bank failure, Fort Jackson,
1970. Cohesion of soil
.
Smith, N. D. 1970. The braided Louisiana. Geol. Soc. America
particles and the intrinsic stream depositional environment: Bull. 77:850-66.
resistance of simple soil systems to Comparison of the Platte River Stanley, J. M., and Cronin, J. E.
wind erosion. Jour. Soil Sci. with some Silurian clastic rocks, 1983. Investigations and
21:154-61. north-central Appalachians. Geol. implications of subsurface
-. 1981. Conjectures, Soc. America Bull. conditions beneath the Trans
hypotheses, and theories of 81:2993-3014. Alaska Pipeline in Atigun Pass. In
drumlin formation. Jour. Glaciol. . 1971. Transverse bars and Proc. Permafrost 4th Internal.
27:503-5. braiding in the Lower Platte Confi. pp. 188-93. Natl. Acad.
1
Smalley, I. J., and Unwin, D. J. 1968. River, Nebraska. Geol. Soc. Sci.
The formation and shape of America Bull. 82:3407-20. Stanley. S. R.. and Ciolkosz, E. J.
drumlins and their distribution 1974. Sedimentology and 1981. Classification and genesis of
and orientation in drumlin fields. bar formationin the upper spodosols in the central
Jour. Glaciol. 7:377-90. Kicking Horse River, a braided Appalachians. So/7 Sci. Soc. Am.
Smart, J. S., and Wallis, J. R. 1971. outwash stream. Jour. Geology Proc. 45:912-17.
Cis and trans links in natural 82:205-23. Stearns, S. R. 1966. Permafrost
channel networks. Water Resour. Snodgrass, D.; Groves, G.; (perenially frozen ground). U.S.
Res. 7:1346-48. Hasselmann, K.; Miller, G.; Army Corps Engrs., Cold Regions
Smith, D. D., and Wischmeier, W. H. Munk, W.; and Powers, W. 1966. Res. and Eng. Lab, Cold Regions
1962. Rainfall erosion. Advances Propagation of ocean swell across Sci. and Eng. (A2). 1
in Agron. 14:109-48. the Pacific. Phil. Trans. Royal Steers, J. A. 1962. The sea coast.
Smith, D. G. 1976. Effect of Soc. London, ser. A:259:431-97. London: Collins.
vegetation on lateral migration of SoilSurvey Staff. 1951. Soil survey- Steinemann, S. 1954. Results of
anastomosed channels of a glacier manual. U.S. Dept. Agri. preliminary experiments on the
meltwater river. Geol. Soc. Handbook 18, Soil Conserv. Serv. plasticity of ice crystals. Jour.
America Bull. 87:857-60. . 1 960. Soil classification, a Glaciol. 2:404-12.
Smith, D. G., and Smith, N. D. comprehensive system 7th — . 1958. Flow and
1976. Sedimentation in approximation. U.S. Dept. Agri., recrystallization of ice. Intl.
anastamosed river systems: Soil Conserv. Serv. Assoc. Sci. Hydrol. Pub.
Examples from alluvial valleys 1975. So/7 taxonomy. U.S. 39:449-62.
near Banff, Alberta. Jour. Sed. Dept. Agri. Handbook 436, Soil Steinen, R. P.; Harrison, R. S.; and
Petrology 50:157-64. Conserv. Serv. Matthews, R. K. 1973. Eustatic

Smith, D. I. 1969. The solution -. 1981. Replacement chapter low stand of sea level between
erosion of limestone in an arctic to Handbook 18, Soil Survey 125,000 and 105,000 B.P.:
morphogenetic region. In Manual, released May 1981. U.S. Evidence from the subsurface of
Problems of the karst Dept. Agri., Soil Conserv. Serv. Barbados, West Indies. Geol. Soc.
denudation, edited by O. Stelcl, Sonu, C. J. 1972. Field observation America Bull. 84:63-70.
pp. 99-110. Brno. of nearshore circulation and Stevens, R., and Carron, M. 1948.
Smith, D. I., and Atkinson, T 1976. meandering currents. Jour. Simple field test for distinguishing
Process landforms and climate in Geophys. Research minerals by abrasion pH. Am.
limestone regions. In 77:18:3232-47. Mineralogist 33:31-49.
Geomorphology and climate, Soucie, G. 1973. Where beaches Stewart, J. H., and LaMarche, V. C,
edited by E. Derbyshire. London: have been going: Into the ocean. Jr. 1967. Erosion and deposition
John Wiley & Sons. Smithsonian 4:3:55-61. produced by the flood of Dec.
Smith, G. D. 1942. Illinois loess- Springer, M. E. 1958. Desert 1964 on Coffee Creek, Trinity
variations in its properties and pavement and vesicular layer of County, California. U.S. Geol.
distribution; a pedologic some desert soils in the desert of Survey Prof. Paper 422-K:l-22.
interpretation. Univ. Illinois Agr. the Lahontan Basin, Nevada. So/7
Exp. Sta. Bull. 490:137-84. Sci. Am. Proc. 22:63-66.
A
Bibliography 562
Stoddart, D. 1969. Climatic Stringfield, V. T., and LeGrand, . 1958. The karstlands of
geomorphology. In Introduction to H. E. 1969. Hydrology of Jamaica. Geogr. Jour.
fluvial processes, edited by carbonate rock terranes — 124:184-99.
R. Chorley, pp. 189-201. London: review. J. Hydrol. 8:349-413. -. 1973. Karst landforms. New
Methuen and Co. Sugden, D. E., and John, B. S. 1976. York: Columbia Univ. Press.
Stokes, W. L. 1964. Incised, wind- Glaciers and landscape. London: ed. 1981. Karst
,
aligned stream patterns of the Edward Arnold Ltd. geomorphology. New York:
Colorado Plateau. Am. Jour. Sci. Sunamura, T. 1975. A laboratory Academic Press.
262:808-16. study of wave-cut platform Swineford, A., and Frye, J. C. 1951.
Stone, R. 1968. Deserts and desert formation. Jour. Geology Petrography of the Peorian loess
landforms. In Encyclopedia of 83:389-97. in Kansas. Jour. Geology
geomorphology, edited by R. W. 1976. Feedback relationship
. 59:306-22.
Fairbridge, pp. 271-79. New inwave erosion of laboratory Swinnerton, A. C. 1932. Origin of
York: Reinhold Book Corp. rocky coast. Jour. Geology limestone caverns. Geol. Soc.
Strahler, A. N. 1950. Equilibrium 84:427-37. America Bull. 43:662-93.
theory of slopes approached by . 1977. A relationship Swinzow, G. K. 1969. Certain
frequency distribution analysis. between wave-induced cliff aspects of engineering geology in
Am. Jour. Sci. 248:800-814. erosion and erosive force of waves. permafrost. Eng. Geol.
— . 1952a. Dynamic basis of Jour. Geology 85:613-18. 3:177-215.
geomorphology. Geol. Soc. . 1978. Mechanisms of shore Taber, S. 1929. Frost heaving. Jour.
America Bull. 63:923-38. platform formation on the Geology 37:428-61.
1952b. Hypsometric (area- southeast coast of the Izu 1930. The mechanics of
.
altitude) analysis of erosional peninsula, Japan. Jour. Geology frost heaving. Jour. Geology
topography. Geol. Soc. America 86:211-22. 38:303-17.
Bull. 63:1117-42. 1982. A predictive model for . 1943. Perennially frozen
1957. Quantitative analysis wave-induced erosion, with ground in Alaska: Its origin and
of watershed geomorphology. Am. application to Pacific coasts of history. Geol. Soc. America Bull.
Geophys. Union Trans. Japan. Jour. Geology 90:167-78. 54:1433-1548.
38:913-20. 1983. Processes of sea cliff
. -. 1953. Origin of Alaska silts.
1958. Dimensional analysis and platform erosion. In CRC Am. Jour. Sci. 251:321-36.
applied to fluvially eroded handbook of coastal processes Tan, K. H. 1980. The release of
landforms. Geol. Soc. America and erosion, edited by P. Komar, silicon, aluminum, and potassium
Bull. 69:279-99. pp. 233-65. Boca Raton, Fla.: during decomposition of soil
-. 1964. Quantitative CRC Press. minerals by humic acid. Soil Sci.
geomorphology of drainage basins Sundborg, A. 1956. The river 129:5-11.
and channel networks. In Klaralven, a study of fluvial Tanner, W. F. 1958. The equilibrium
Handbook of applied hydrology, processes. Geogr. Annlr. beach. Am. Geophys. Union
edited by V. T. Chow, 38:280-91. Trans. 39:889-91."
pp. 4-39-4-76. New York: Suzuki, T, and Takahashi, K. 1981. . 1974. The incomplete flood
McGraw-Hill. An experimental study of wind plain. Geology 2:105-6.
.
1965. Introduction to abrasion. Jour. Geology Tator, B. A. 1952. Pediment
physical geography. New York: 89:509-22. characteristics and terminology.
John Wiley & Sons. Svasek, J. N., and Terwindt, J. H. J. Ann. Assoc. Am. Geog.
-. 1966. Tidal cycle of changes 1974. Measurements of sand 42:295-317.
on an equilibrium beach. Jour. transport by wind on a natural 1953. Pediment
.
Geology 74:247-68. beach. Sedimentology 21:31 1-22. characteristics and terminology

1968. Quantitative Swanson, D. 1972. Magma supply Ann. Assoc. Am. Geog. 43:47-53.
geomorphology. In Encyclopedia rate at Kilauea Volcano, Ten Brink, N. W. 1974. Glacio-
of geomorphology, edited by 1925-1 97 1. Science 175:169-70. isostasy: New data from West
R. W. Fairbridge, pp. 898-912. Sweeting, M. M. 1950. Erosion Greenland and geophysical
New York: Reinhold Book Corp. cycles and limestone caverns in implications. Geol. Soc. America
Strakhov, N. 1967. Principles of the Ingleborough District of Bull. 85:219-28.
lithogenesis. London: Oliver and Yorkshire. Geogr. Jour. Terzaghi, K. 1936. The shearing
Boyd. 115:63-78. resistance of saturated soils. Proc.
1st Internal. Conf. on Soils hiech
and Foundation Eng. 1:54-66.
563 Bibliography
. 1943. Theoretical soils Thorne, C. R., and Lewin, J. 1979. -. 1982. Accelerated erosion:
mechanics. New York: John Bank processes, bed material Process, problems, and prognosis.
Wiley & Sons. movement, and platform Geology 10:524-29.
. 1950. Mechanism of development in a meandering Trainer,F., and Heath, R. 1976.
landslides. In Application of river. In Adjustments of the Bicarbonate content of
geology to engineering practice, fluvial system, edited by D. D. groundwater in carbonate rock
edited by S. Paige. Geol. Soc. Rhodes and G. P. Williams, in eastern North America.
America Berkey Vol., pp. 83-123. pp. 117-37.Dubuque, Iowa: J. Hydrol. 31:37-55.
1962. Stability of steep Kendall-Hunt. Trenhaile, A. S. 1971. Drumlins:
slopes on hard unweathered rock. Thorne, C. R., and Tovey, N. K. Their distribution, orientation and
Geotechnique 12:251-70. 1981. Stability of composite river morphology. Can. Geog.
Thie,J. 1974. Distribution and banks. Earth Surf. Proc. and 15:113-26.
thawing of permafrost in the Landforms 6:469-84. . 1977. Cirque elevation and
southern part of the discontinuous Thornes, J. B. 1970. Hydraulic Pleistocene snowlines. Zeit. f
permafrost zone in Manitoba. geometry of stream channels in Geomorph. 21:445-59.
Arctic 27:189-200. the Xingu-Araguaia headwaters. 1979. The morphology of
.
Thomas, H. P., and Ferrell, J. E. Geogr. Jour. 136:376-82. valley steps in the Canadian
1983. Thermokarst features Thornton, E. B. 1973. Distribution of Cordillera. Zeit.f. Geomorph.
associated with buried sections of sediment transport across the surf 23:27-44.
the Trans Alaska pipeline. In zone. Proc. 13th Conf. on Coast. Tricart, J. 1967. Le modele des
Proc. Permafrost 4th Internal. Eng., pp. 1049-68. regions periglaciaires. In Traite
Conf, pp. 1245-50. Natl. Acad. Thorp, J., and Smith, G. 1949. de geomorphologie 2, edited by
Sci. Higher categories of soil J. Tricart and A. Cailleux. Paris:
Thomas, R. H. 1979. West Antarctic classifications. Order, Suborder, SEDES.

Ice Sheet: Present-day thinning and Great Soil Groups. Soil Sci. 1968. Notes
.
and Holocene retreat of margins. 67:117-26. geomorphologiques sur la

Science 205:1257-58. Thrailkill, J. 1968. Chemical and karstification en Barbade
Thomas, R. H.: MacAyeal, D. R.; hydrologic factors in the (Antilles): Mem. docums. cent.
Bentley, C. R.; and Clapp, J. L. excavation of limestone caves. docum. cartogr. geogr. 4:329-34.
1980. The creep of ice, Geol. Soc. America Bull. -. 1969. Geomorphology of
geothermal heat flow, and 79:19-45. cold environments, trans, by
Roosevelt Island, Antarctica. . 1972. Carbonate chemistry Edward Watson. New York: St.
Jour. Glaciol. 25:47-60. of aquifer and stream water in Martin's Press.
Thorn, C. E. 1976. Quantitative Kentucky. J. Hydro!. 16:93-104. Tricart,J., and Cailleux, A. 1972.
evaluation of nivation in the Thrailkill, and Robl, T. 1981.

J., Introduction to climatic
Colorado Front Range. Geol. Soc. Carbonate geochemistry of vadose geomorphology. London:
America Bull. 87:1169-78. water recharging limestone Longman.
1979. Bedrock freeze-thaw
. aquifers. J. Hydrol. 54:195-208. Trimble, S. W. 1977. The fallacy of
weathering regime in an alpine Tinkler, K. J. 1982. Avoiding error stream equilibrium in
environment, Colorado Front when using the Manning contemporary denudation studies.
Range. Earth Surf. Proc. and equation. Jour. Geology Am. Jour. Sci. 277:876-87.
Landforms 4:21 1-28. 90:326-28. Trimble, S. W., and Lund, S. W.
Thorn, C. E., and Hall, K. 1980. Todd. D. K. 1959. Groundwater 1982. Soil conservation and the
Nivation: An arctic-alpine hydrology. New York: John Wiley reduction of erosion and
comparison and reappraisal. Jour. & Sons. sedimentation in the Coon Creek
Glaciol. 25:109-24. , ed. 1970. The water Basin, Wisconsin. U.S. Geol.
Thorne, C. R. 1982. Processes and encyclopedia. Port Washington, Survey Prof. Paper 1234.
mechanisms of river bank erosion. N.Y: Water Information Center. Troll, C. 1958. Structure soils,
In Gravel-bed rivers, edited by R. Townsend, D. W., and Vickery, R. P. solifiuction, and frost climates of
Hey, J. Bathurst, and C. Thorne, 1967. An experiment in the Earth. U.S. Army Corps
pp. 227-72. New York: John regelation. Philos. Mag. Engrs., Snow, Ice, Permafrost
Wiley & Sons. 16:1275-80. Res. Est. Trans. 43.
Toy, T. J. 1977. Hillslope form and Trudgill, S. 1976.The marine erosion
climate. Geol. Soc. America Bull. of limestones on Aldabra Atoll,
88:16-22. Indian Ocean. Zeit.f. Geomorph.,
Suppl. 26:164-200.
Bibliography 564
Tuan, Ti-Fu. 1959. Pediments in Vagners. V. J. 1966. Lithologic Vitek,J. D. 1983. Stone polygons:
southeastern Arizona. Berkeley, relationship of till to carbonate Observations of surficial activity.

Calif.: Univ. Calif. Pubs, in bedrock in southern Ontario. M.S. In Proc. Permafrost 4th Internal.
Geog. 13. thesis.Geology Dept.. Univ. Conf., pp. 1326-31. Natl. Acad.
. 1962. Structure, climate and Western Ontario. Sci.
basin land forms in Arizona and Van Arsdale, R. 1982. Influence of Vivian, R. 1970. Hydrologie et
New Mexico. Ann. Assoc. Am. calcrete on the geometry of erosion sous-glaciaires. Rev. geog.
Geog. 52:51-68. arroyos near Buckeye, Arizona. alp. 58:241-64.
Tucker, M. J. 1950. Surf beats: Sea Geol. Soc. America Bull. . 1980. The nature of the ice-
waves of minute period.
1 to 5 93:20-26. rock interface: The results of
Proc. Royal Soc. London, ser. A, Van Dorn, W. G. 1965. Tsunamis. In investigation on 20,000 :
of the m
202:565-73. Hydroscience advances 2. New rock bed of temperate glaciers.
Twidale, C. R. 1962. Steepened York: Academic Press. Jour. Glaciol. 25:267-77.
margins of inselbergs from 1966. Tsunamis. In
. Vivian, R., and Bouquet, G. 1973.
northwestern Eyre Peninsula, Encyclopedia of oceanography. Subglacial cavitation phenomena
South Australia. Zeit.f. edited by R. W. Fairbridge, under Glacier d" Argentiere. Mont
Geomorph. 6:51-69. pp. 941-43. New York: Reinhold Blanc, France. Jour. Glaciol.
. 1964. Erosion of an alluvial Book Corp. 12:439-52.
bank at Birdwood, South Van Heukon, T. K. 1977. Distant Wahrhaftig, C. 1965. Stepped
Australia. Zeit.f. Geomorph. source of 1976 dustfall in Illinois topography of the southern Sierra
8:189-211. and Pleistocene weather models. Nevada. Geol. Soc. America Bull.
1967. Origin of the Geology 5:693-95. 76:1165-90.
piedmont angle as evidenced in Vanoni, V. A. 1941. Some Wahrhaftig, C, and Cox, A. 1959.
South Australia. Jour. Geology experiments on the transportation Rock glaciers in the Alaska
75:393-411. of suspended load. Am. Geophys. Range. Geol. Soc. America Bull.
1968. Weathering. In Union Trans., 22nd Ann. Mtg., 70:383-436.
Encyclopedia of geomorphology, pt. 3:608-20. Waitt, R. B.. Jr.1980. About forty
edited by R. W. Fairbridge, . 1946. Transportation of last-glacial Lake Missoula
pp. 1228-32. New York: Reinhold suspended sediment by water. jokulhlaups through southern
Book Corp. Am. Soc. Civil Engineers Trans. Washington. Jour. Geology
. 1972. Evolution of sand 3:67-133. 88:653-79.
dunes in the Simpson Desert, Vanoni, V. A., and others. 1966. Walcott. R. I. 1972. Late Quaternary
central Australia. Inst. Brit. Geog. Sediment transportation movements in eastern
vertical
Trans. 56:77-110. mechanics. Initiation of motion. North America. Quantitative
-. 1978. On the origin of Am. Soc. Civil Engineers, Jour. evidence of glacio-isostatic
pediments in different structural Hydraulics Div. rebound. Rev. Geophys. and
settings. Am. Jour. Sci. 92:HY2:291-313. Space Physics 10:849-84.
278:1138-76. Varnes, D. J. 1958. Landslide types Walker, A. S. 1982. Deserts of
Twidale, C. R., and Bourne, J. A. and processes. In Landslides and China. Am. Scientist 70:366-76.
1975. Episodic exposure of engineering practice, edited by Wallace. R. E. 1977. Profiles and
inselbergs. Geol. Soc. America E. Eckel, pp. 20-47. Washington, ages of young fault scarps, north-
Bull. 86:1473-81. D.C.: Highway Research Board central Nevada. Geol. Soc.
Ueta, H. T, and Garfield, D. E. Spec. Rept. 29. America Bull. 88:1267-81.
1968. Deep core drilling program . 1978. Slope movement types . Geometry and rates of
1978.
at Byrd Station 1967-68. U.S. and processes. In Landslides, change of fault-generated range
Arctic Journal 3:1 1 1-12. edited by R. Schuster and fronts, north-central Nevada. U.S.
U.S. Army Corps of Engineers. 1971. R. Krizak. pp. 11-33. Geol. Survey Jour, of Res.
National shoreline study. Washington, D.C.: Trans. Res. 6:637-49.
Washington, D.C.: U.S. Army Board, Natl. Acad. Sci. Walters. J. C. 1978. Polygonal
Corps Engr. Vernon, P. 1966. Drumlins and patterned ground in central New
Ursic, S. J., and Dendy, F. E. 1965. Pleistocene ice flow over the Ards Jersey. Quat. Res. 10:42-54.
Sediment yields from small Peninsula. Jour. Glaciol. 6:401-9. Ward. W. H. 1945. The stability of
watersheds under various land Verstappen. H. Th. 1964. Karst natural slopes. Geogr. Jour.
uses and forest covers Proc. Fed. morphology of the Star 105:170-97.
Inter-Agency Sedimentation Conf. Mountains (central New Guinea) Warnke. D. A. 1969. Pediment
(1963). U.S. Dept. Agri. Misc. and its relation to lithology and evolution in the Halloran Hills,
Publ. 970:47-52. climate. Zeit.f. Geomorph. central Mojave Desert. California.
8:40-49. Zeit.f. Geomorph. 13:357-89.
565 Bibliography
Warren, A. 1970. Dune trends and Weertman, J. 1957. On the sliding of Westgate, J. A. 1968. Linear sole
their implications in the central glaciers. Jour. Glaciol. 3:33-38. markings in Pleistocene till. Geol.
Sudan. Zeit.f. Geomorph., Suppl. . 1964. The theory of glacier Mag. 105:501-5.
10:154-80. sliding. Jour. Glaciol. 5:287-303. Weyl, P. K. 1958. The solution
.1971. Dunes in the Tenere . 1967. An examination of the kinetics of calcite. Jour. Geology
Desert. Geogr. Jour. 137:458-61. Lliboutry theory of glacier sliding. 66:163-76.
Washburn, A. L. 1956. Classification Jour. Glaciol. 6:489-94. Whalley, W. B. 1983. Rock
of patterned ground and review of -. 1979. The unsolved general glaciers —
Permafrost features or
suggested origins. Geol. Soc. glacier sliding problem. Jour. glacial relics. In Proc. Permafrost
America Bull. 67:823-66. Glaciol. 23:97-111. 4th Internal. Conf,
. 1967. Instrumental Weertman, J., and Weertman, J. B. pp. 1396-1401. Natl. Acad. Sci.
observations of mass-wasting in 1964. Elementary dislocation Whalley, W. B.; Douglas, G. R.; and
the Mesters Vig district. theory.New York: Macmillan. McGreevey, J. P. 1982. Crack
Northeast Greenland. Medd. om Weinert, H. 1961. Climate and propogation and associated
Gr0nland 166:4. weathered Karroo dolerites. weathering in igneous rocks. Zeit.
. 1970. An approach to a Nature 191:325-29. f. Geomorph. 26:33-54.
genetic classification of patterned . 1965. Climatic factors Whillans, I. M. 1978. Inland ice
ground. Acta Geogr. Lffdz. affecting the weathering of sheet thinning due to Holocene
24:437-46. igneous rocks. Agri. Meterol. warmth. Science 201:1014-16.
. 1973. Periglacial processes 2:27-42. Whitaker, R. H.; Buol, S. W.;
and environments. London: Welder, F. A. 1959. Processes of Niering, W. A.; and Havens,
Edward Arnold Ltd. deltaic sedimentation in the lower Y. H. 1968. A soil and vegetation
. 1980. Geocryology. New Mississippi River. Louisiana pattern in the Santa Catalina
York: John Wiley & Sons. State Univ., Coastal Studies Inst. Mountains, Arizona. Soil Sci.
Wasson, R. J. 1977. Catchment Tech. Rept. 12. 105:440-51.
processes and the evolution of Wellman, P. 1982. Surging of Fisher White, E. L., and Reich, B. M. 1970.
alluvial fans in the lower Derwent Glacier, eastern Antarctica: Behaviour of annual floods in
Valley,Tasmania. Zeit. f. Evidence from geomorphology. limestone basins in Pennsylvania.
Geomorph., Suppl. 21:147-68. Jour. Glaciol. 28:23-28. J.Hydrol. 10:193-98.
Wasson, R. J., and Hall, G. 1982. A Wells, J. T; Prior, D. B.; and White, E., and White, W. 1979.
long record of mudslide movement Coleman, J. M. 1980. Flowslides Quantitative morphology of
at Waerenga-O-Kuri, New in muds on extremely low angle landforms in carbonate rock
Zealand. Zeit f. Geomorph. tidal flats, northeastern South basins in the Appalachian
26:73-85. America. Geology 8:272-75. Highlands. Geol. Soc. America
Watts, S. H. 1979. Some Wells, S. 1976. Sinkhole plain Bull. 90:385-96.
observations on rock weathering, evolution in the central Kentucky 1983. Karst landforms and
.
Cumberland Peninsula, Baffin karst. Natl. Speleol. Soc. Bull. drainage basin evolution in the
Island. Can. J. Earth Sci. 38:103-6. Obey River basin, north-central
16:977-83. . 1977. Geomorphic controls Tennessee, U.S.A. J. Hydrol.
Wayne, W. J. 1967. Periglacial of alluvial fan deposition in the 61:69-82.
features and climatic gradient in Sonoran Desert, southwestern White, S. E. 1976a. Is frost action
Illinois, Indiana and western Arizona. In Geomorphology in really only hydration shattering?
Ohio, east-central United States. arid regions, edited by A review. Arc. Alp. Res. 8:1-6.
In Quaternary paleoecology, D. Doehring, pp. 27-50. S.U.N.Y. 1976b. Rock glaciers and
.
edited by E. Cushing and Binghamton: 8th Geomorph. block fields, review and new data.
H. Wright, pp. 393-414. New Symposium. Quat. Res. 6:77-98.
Haven, Conn.: Yale Univ. Press. Wentworth, C. K. 1938. Marine White, W. A. 1966. Drainage
1981. Ice segregation as an
. beach formation: Water level asymmetry and the Carolina
origin for lenses of nonglacial ice weathering. Jour. Geomorphology capes. Geol. Soc. America Bull.
in "ice-cemented" rock glaciers. 1:6-32. 77:223-40.
Jour. Glaciol. 27:506-10. Wescott, W. A., and Ethridge, F. G. White, W. B. 1960. Termination of
Wear, J., and White, J. 1951. 1980. Fan-delta sedimentology passages in Appalachian caves as
Potassium fixation in clay mineral and tectonic-Hallahs Fan delta, evidence for a shallow phreatic
studies as related to crystal southeast Jamaica. Am. Assoc. origin. Natl. Speleol. Soc. Bull.
structure. Soil Sci. 71:1-14. Petroleum Geologists Bull. 22:43-53.

64:374-99. 1969. Conceptual models for
.
carbonate aquifers. Ground Water

7:15-21.
—
Bibliography 566
. 1976. Geology and biology Williams, G. P., and Guy, H. P. — .1973. Ergs. Sed. Geol.
of Pennsylvania caves. 1973. Erosional and depositional 10:77-106.
Pennsylvania Geol. Survey Gen. aspects of Hurricane Camille in Wilson, L. 1968. Morphogenetic
Rept. 66:1-71. Virginia, 1969. U.S. Geol. Survey classification. In Encyclopedia of
1977. Conceptual models for Prof. Paper 804. geomorphology, edited by R. W.
carbonate aquifers: revisited. In Williams, G. P., and Wolman, M. G. Fairbridge, pp. 717-28. New
Hydrologic problems in karst 1984. Downstream effects of dams York: Reinhold Book Corp.
terrains, edited by R. Dilamarter on alluvial U.S. Geol.
rivers. 1969. Les relations entre
.
les
and S. Csallany, pp. 176-87. Survey Prof. Paper 1286. processus geomorphologiques et le
Bowling Green: Western Williams, J. R. 1970. Groundwater climat moderne comme methode
Kentucky Univ. Press. in thepermafrost region of de paleoclimatologie. Rev. geog.
White, W. B., and Schmidt, V. A. Alaska. U.S. Geol. Survey Prof. phys. et de geologie dyn.
1966. Hydrology of a karst area Paper 696. 11:303-14.
central West Virginia.
in east Williams, L. 1964. Regionalization of . 1972. Seasonal sediment
Water Resour. Res. 2:549-60. freeze-thaw activity. Ann. Assoc. yield patterns of United States
Whitney, M. I. 1978. The role of Am. Geog. 54:597-611. rivers. Water Resour. Res.
vorticity in developing lineation Williams, P. J. 1966. Downslope soil 8:1470-79.
by wind erosion. Geol. Soc. movement at a sub-Arctic location . 1973. Variations in mean
America Bull. 89:1-18. with regard to variations with annual sediment yield as a
Whitney, M. I., and Dietrich, R. V. depth. Canadian Geotech. Jour. function of mean annual
1973. Ventifact sculpture by 3:191-203. precipitation. Am. Jour. Sci.
windblown dust. Geol. Soc. Williams, P. W. 1963. An initial 273:335-49.
America Bull. 84:2561-82. estimate of the speed of limestone Winkler, E. M. 1965. Weathering
Whittecar, G. R., and Mickelson, D. solution in County Clare. Ir. rates as exemplified by
1979. Composition, internal Geogr. 4:432-41. Cleopatra's Needle in New York
structures, and a hypothesis of . 1966. Morphometric City. Jour. Geol. Educ. 13:50-52.
formation for drumlins, Waukesha analysis of temperate karst . 1975. Stone: Properties,
County, Wisconsin, U.S.A. Jour. landforms. Ir. Speleol. 1:23-31. durability in man's environment,
Glaciol. 22:357-71. -. 1971. Illustrating 2d ed. New York: Springer-
Wiegel, R. L. 1964. Oceanographical morphometric analyses of karst Verlag.
engineering. Englewood Cliffs, with examples from New Guinea. Winkler, E. M., and Wilhelm. E. J.
N.J.: Prentice-Hall. Zeit.f. Geomorph. 15(1):40-61. 1970. Salt bursts by hydration
Wilcock, D. N. 1971. Investigation -. 1972a. The analysis of pressures in architectural stone in
into the relations between bedload spatial characteristics of karst urban atmosphere. Geol. Soc.
transport and channel shape. terrains. In Spatial analysis in America Bull. 81:567-72.
Geol. Soc. America Bull. geomorphology, edited by R. J. Woldenberg, M. J. 1969. Spatial
82:2159-76. Chorley. London: Methuen and order in fluvial systems. Horton's
. 1975. Relations between Co. laws derived from mixed
planimetric and hypsometric 1972b. Morphometric hexagonal hierarchies of drainage
variables in third- and fourth- analysis of polygonal karst in New basin areas. Geol. Soc. America
order drainage basins. Geol. Soc. Guinea. Geol. Soc. America Bull. Bull. 80:97-112.
America Bull. 86:47-50. 83:761-96. Wolfe. TE. 1964. Cavern
Wilford, G. E., and Wall, J. R. D. — . 1983. The role of the development in the Greenbrier
1965. Karst topography in subcutaneous zone in karst Series, West Virginia. Xatl.
Sarawak. /. Trop. Geogr. hydrology. J. Hydrol. 61:45-67. Speleol. Soc. Bull. 26:37-60.
21:44-70. Willman, H. B., and Frye, J. C. Wollast. R. 1967. Kinetics of the
Williams, A. T. 1973. The problem 1970. Pleistocene stratigraphy of alteration of K-feldspar in
of beach cusp development. Jour. Illinois. Illinois Geol. Survey buffered solutions at low
Sed. Petrology 43:857-66. Bull. 94. temperature. Geochim. et
Williams, G. 1964. Some aspects of Wilson, B. W. 1966. Seiche. In Cosmochim. Acta 31:635-48.
the eolian saltation load. Encyclopedia of oceanography. Wolman, M. G. 1955. The natural
Sedimentologv 3:257-87. edited by R. W. Fairbridge, channel of Brandywine Creek.
Williams, G. P. 1978. Hydraulic pp.804-1 1. New York: Reinhold Pennsylvania. U.S. Geol. Survej
geometry of river cross sections Book Corp. Prof. Paper 271.
Theory of minimum variance. Wilson, I.G. 1972. Aeolian . 1959. Factors influencing
U.S. Geol. Survey Prof. Paper —
bedforms their development and erosion of a cohesive river bank.
1029. origins. Sedimentology Am. Jour. Sci. 257:204-16.
19:173-210.
567 Bibliography
1967. A cycle of
. Woodruff, N. P., and Siddoway, F. H. Yang, C. T, 1973. Incipient motion
sedimentation and erosion in 1965. A wind erosion equation. and sediment transport. Am. Soc.
urban river channels. Geogr. Soil Sci. Soc. Am. Proc. Civil Engineers, Jour. Hydraulics
Annlr. 49-A:385-95. 29:602-8. Z)/v.99:HYl 0:1 679- 1704.
Wolman, M. G., and Brush, L. M., Wright, H. E., and Frey, D. G., eds. Yeats, R. S. 1978. Neogene
Jr. 1961. Factors controlling the 1965. The Quaternary of the acceleration of subsidence rates in
size and shape of stream channels United States. Princeton, N.J.: southern California. Geology
in coarse noncohesive sands. U.S. Princeton Univ. Press. 6:456-60.
Geol. Survey Prof. Paper 282-G. Wright, J., and Schnitzer, M. 1963. Yefimov, A. I., and Dukhin, I. E.
Wolman, M. G., and Gerson, R. Metallo-organic interactions 1968. Some permafrost
1978. Relative scales of time and associated with podsolisation. Soil thicknesses in the Arctic. Polar
effectiveness of climate in Sci. Soc. Am. Proc. 27:171-76. Record 14:68.
watershed geomorphology. Earth Wright, L. D. 1977. Sediment Young, A. 1960. Soil movement by
Surf. Proc. and Landforms transport and deposition at river denudational processes on slopes.
3:189-208. mouths: A synthesis. Geol. Soc. Nature 188:120-22.
Wolman, M. G., and Leopold, L. B. America Bull. 88:857-68. . 1961. Characteristic and
1957. River flood plains; some Wright, L. D.; Chappell, J.; Thorn, limiting slope angles. Zeit. f.
observations on their formation. B. G.; Bradshaw, M. P.; and Geomorph. 5:126-31.
U.S. Geo!. Survey Prof. Paper Cowell, P. 1979. Morphodynamics -. 1972a. Slopes. Edinburgh:
282-C. of reflective and dissipative beach Oliver and Boyd.
Wolman, M. G., and Miller, J. P. and inshore systems: Southeastern .
1972b. The soil catena: A
1960. Magnitude and frequency Australia. Marine Geol. systematic approach. In
of forces in geomorphic processes. 32:105-40. International geography 1 972,
Jour. Geology 68:54-74. Wright, L. D.; Guza, R. T; and Congr. Int. Geogr. Commun.,
Womack, W. R., and Schumm, S. A. Short, A. D. 1982a. Dynamics of n. 22, v. 1:287-89.
1977. Terraces of Douglas Creek, a high-energy dissipative surf Young, R. W. 1983. The tempo of
northwestern Colorado: An zone. Marine Geol. 45:41-62. geomorphological change:
example of episodic erosion. Wright, L. D.; Nielson, P. N.; Short, Evidence from southeastern
Geology 5:72-76. A. D.; and Green, M. O. 1982b. Australia. Jour. Geology
Woo, M., and Heron, R. 1981. Morphodynamics of a macrotidal 91:221-30.
Occurrence of ice layers at the beach. Marine Geol. 50:97-128. Zeigler, J. M.; Hayes, C. R.; and
base of High Arctic snowpacks. Wright, L. D., and Short, A. D. Tuttle, S. D. 1959. Beach changes
Arc. Alp. Res. 13:225-30. 1983. Morphodynamics of during storms on outer Cape Cod,
Woo, M., and Marsh, P. 1977. Effect beaches and surf zones in Massachusetts. Jour. Geology
of vegetation as limestone solution Australia. In CRC handbook of 67:318-36.
in a small High Arctic basin. Can. coastal processes and erosion, Zenkovich, V. P. 1967. Processes of
J. Earth Sci. 14:571-81. edited by P. Komar, pp. 35-64. coastal development, translated
Wood, A. 1942. The development of Boca Raton, Fla.: CRC Press. by D. G. Fry, edited by J. A.
hillside slopes. Proc. Geol. Assoc. Wyllie, P. 1971. The dynamic Earth. Steers. Edinburgh: Oliver and
53:128-40. New York: John Wiley & Sons. Boyd.
Wood, F. J. 1978. The strategic role Yaalon, D. H. 1975. Conceptual Zirjacks, W. L., and Hwang, C. T.
of perigean spring tides. models in pedogenesis. Can soil- 1983. Underground utilidors at
Washington, D.C.: U.S. Dept. of forming functions be solved? Barrow, Alaska: A
two-year
Commerce. Geoderma 14:189-205. Permafrost 4th
history. In Proc.
Woodcock, A. H. 1974. Permafrost Yair, A.; Lavee, H.; Bryan, R. B.; Internal. Conf, pp. 1513-17.
and climatology of a Hawaii and Adar, E. 1980. Runoff and Natl. Acad. Sci.
volcano crater. Arc. Alp. Res. erosion processes and rates in the
6:49-62. Zin Valley badlands, northern
Woodcock, A. H.; Furumoto, A. S.; Negev, Israel. Earth Surf. Proc.
and Woollard, G. P. 1970. Fossil and Landforms 5:205-25.
ice in Hawaii. Nature 226:873.
Credits
Illustrations © 1970 The Geological Society of Reprinted by permission. Figure 3.19:

America. Reprinted by permission. Figure From Soils and Geomorphology by
2.14: From Brakenridge, G. R., 1981. Peter W. Birkeland. Copyright © 1984 by
Used with permission of Quaternary Oxford University Press, Inc. Reprinted
Chapter 1 Research. Figures 2.15 and 2.16: H. E. by permission. Figure 3.20: From Gile,
Figure 1.6: "From Handbook of Applied Wright and D. G. Frey, eds., The 1975. Used with the permission of
Hydrology by Ven T. Chow. Copyright Quaternary of the United States. Quaternary Research. Figures 3.21 and
© 1964 by McGraw-Hill, Inc. Used with Copyright© 1965 by Princeton 3.22: From Gile et al., 1981. Used with
permission of McGraw-Hill Book University Press. Reprinted with the permission of New Mexico Bureau of
Company." Figure 1. 11: From Patton and permission of Princeton University Press. Mines and Mineral Resources. Figure
Schumm 1975, with permission of The Figure 2.17: "From Langbein and 3.23: From Gile, 1975. Used with the
Geological Society of America. Reprinted Schumm, Transactions of the American permission of Quaternary Research.
by permission. Geophysical Union, Vol. 39, page 1077,
1958, copyrighted by American
Geophysical Union." Chapter 4
Chapter 2 Figure 4.7: From Peltier, L., Annals of
Figure 2.1: From Wyllie, The
P., (ed.). the Association of American
Dynamic Earth. © 1971 John Wiley & Chapter 3 Geographers, vol. 40. © 1950 Association
Sons, Inc., New York. Reprinted by Figure 3.3: From Bass-Becking et al., of American Geographers. Reprinted by
permission. Figure 2.4: Adapted from Ten 1960. © by University of Chicago Press. permission. Figures 4.1 1 and 4.15: From
Brick in Geological Society of America Reprinted by permission. Figures 3.4 and Carson, M. A., and M. J. Kirkby,
Bulletin, vol. 85, page 224. © 1974 The 3.7: From Loughnan, Chemical Hillslope Form and Process. © 1 972
Geological Society of America. Reprinted Weathering of the Silicate Minerals. Cambridge University Press. Reprinted by
by permission. Figure 2.7: From Burnett, © 1969 American Elsevier Publishing Co. permission. Figure 4.12: From Legget in
A. W. and S. A. Schumm, "Alluvial-River Reprinted by permission. Figure 3.9: Geological Society of America Bulletin,
Response to Neotectonis Deformation in Reprinted with permission of Macmillan vol. 78,page 1444. © 1967 The
Louisiana and Mississippi," in Science, Publishing Company from The Nature Geological Society of America. Reprinted
49-50, figure, October 7,
Vol. 222, pp. and Properties of Soils, 6th ed., by by permission. Figure 4.18: From Kirkby,
1983. Copyright 1983 by the American Harry O. Buckman and Nyle C. Brady. M. J., "Measurement and Theory of Soil
Association for the Advancement of Copyright © 1960 by Macmillan Creep," in Journal of Geology, vol. 75.
Science. Reprinted by permission. Figure Publishing Company. Figure 3.12: From © 1967 The University of Chicago Press.
2.10: Copyright© 1968 by Dowden, Senstius,M. W., "Climax Forms of Reprinted by permission. Figures 4.20
Hutchinson & Ross, Inc. Figure 2.1 1: Chemical Rock- Weathering," in and 4.21a: After Varnes 1958 in
H. E. Wright and D. G. Frey, eds., The American Scientist, Vol. 46. © 1958 Landslides and Engineering Practice,
Quaternary of the United States. American Scientist. Reprinted by Special Report 29. Used with permission
Copyright © 1965 by Princeton permission. Figure 3.16: From of the Transportation Research Board.
University Press. Reprinted by permission Hausenbuiller, R. L., Soil Science: Figure 4.24: From Ritter, Dale F.,
of Princeton University Press. Figure Principles & © 1972 Wm. C.
Practices. Zeitschrift Fur Geomorphologie, Vol. 16,
2.12: From Bloom et al., 1974. Used with Brown Dubuque, Iowa. All
Publishers, page 91, figure 4. © 1972 Gebriider
permission of Quaternary Research. Rights Reserved. Reprinted by Borntraeger. Reprinted by permission.
Figure 2.13: From Saucier and Fleetwood permission. Figure 3.17: From Jenny and Figure 4.26: From Dalrymple, et al.,
in Geological Society of America Leonard in Soil Science. Vol. 36:367. Zeitschrift Fur Geomorphologie, Vol. 12,
Bulletin, Volume 81, figure 7, page 879. © 1939 The Williams & Wilkins Co.
569
Credits 570
figure Used with permission of

1 . Gardner, Thomas in Geological Society Methuen and Co., Ltd. Figure 8.9: From
ZeitschriftFur Geomorphologie of America Bulletin, vol. 94, page 671. Williams, G., Sedimentology, vol. 3, page
published by Gebruder Borntraeger © 1982 The Geological Society of 282, figure 19. © 1964 Blackwell
Verlagsbuchhandlung Berlin. Stuttgart, America. Reprinted by permission. Scientific Publications. Reprinted by
Johannesstra/3e 3A, 7000 Stuttgart 1. permission. Figure 8.10: From Wilson,
Figure 4.27: From Selby, M. J., I. G., Sedimentology, vol. 19, page 193,
Zeitschrift Fur Geomorphologie, Vol. 24, Chapter 7 figure 2. © 1972 Blackwell Scientific
page 48. figure 4. Used with permission of
Figure 7.3: From Hickin, in American
Publications. Reprinted by permission.
Journal of Science. Vol. 274, figure 2a,
Zeitschrift Fur Geomorphologie Figures 8.1 and 8.15: From R. A.
1
page 416. © 1974 Kline Geology Bagnold, 1941, "The Physics of Blown
published by Gebruder Borntraeger
Laboratory- Yale University. Reprinted by
Verlagsbuchhandlung Berlin. Stuttgart, Sand and Desert Dunes." Used with
Johannesstra/3e 3A, 7000 Stuttgart 1.
permission. Figure 7.4: From Hickin, in
permission of Methuen and Co., Ltd.
Figure 4.31: From "Forms of bluffs
GSA Bulletin, vol. 86, figure 6, page 491.
Figure 8.18: The Quaternary of Illinois:
degraded for different lengths of time in © 1975 The Geological Society of
Univ. 111. Coll. Agr. Spec. Pub. 14.
America. Reprinted by permission. Figure
Emmet County, Michigan, U.S.A." in
7.8: From Everitt, in American Journal of
Earth Surface Processes and Landforms
by D. Nash. Copyright © 1980 John
Science, vol. 266, figure 5, page 429. Chapter 9
Wiley & Sons, Ltd. Reprinted by © 1968 Kline Geological Laboratory- Yale Figures 9.4, 9.12, 9.15, 9.16, and 9.17:
University. Reprinted by permission. "Reproduced from the Journal of
permission of John Wiley & Sons, Ltd.
Figure 7.20: From Bull, in Journal of Glaciology by permission of the
Geological Education, vol. 16, figure 2, International Glaciological Society."
Chapter 5 page 105. © 1968 National Association of Figure 9.20:From Sharp, R. P., Glaciers.
Figure 5.1: From Bulletin AAPG: 1967. Geology Teachers. Reprinted by © 1960 University of Oregon Press.
Used with permission of the American permission. Figure 7.23: After Cooke, in Reprinted by permission.
Association of Petroleum Geologists. American Journal of Science, vol. 269,
Figure 5.28: "From Kochel, et al.. Water figure 1, page 27. © 1970 Kline Geology
Resources Research, vol. 18, page 1168, Laboratory- Yale University. Reprinted by Chapter 10
1982, copyright by the American permission. Figure 7.25: After Ruxton Figures 10.1 and 10.11: After Boulton
Geophysical Union." Figure 5.32: From and Berry, in Revue de Geomorphologie 1974. Used with permission of Donald R.
Milliman and Meade 1983, Journal of Dynamique, vol. 12, 1961. Reprinted by Coates, Glacier Geomorphology. Figure
Geology, © The University of Chicago permission. Figure 7.27: From Scott and
10.4: From W. V., RGS Research
Lewis,
Press.Used with permission of the Fisher, in Delta Systems in the
Series: 4. ©
1960 Royal Geographical
University of Chicago Press. Figure Society. Reprinted by permission. Figure
5.36: Exploration of Oil and Gas, figure 37,
Wolman 10.8: After McCall, RGS Research
After Geografiska Annaler,
in 1969. Reprinted by permission of the
vol. 49- A, figure 1, page 386. © 1967 Series: 4. © 1960 Royal Geographical
Texas Bureau of Economic Geology.
Almqvist & Wiksel! International. Society. Reprinted by permission. Figure
Figure 7.28: From Russell, in River and
Reprinted by permission. 10.16: After Goldthwait 1951. © The
Delta Morphology, Coastal Studies Series
University of Chicago. Used with
#20, figure 8, page 44. © 1967 Louisiana
permission. Figure 10.22: "Reproduced
State University Press. Reprinted by
Chapter 6 permission. Figure 7.29: Reprinted with
from the Journal of Glaciology by-
Figure 6.2: From Simons and Richardson, permission, from Encyclopedia of

permission of the International
Transactions of the American Society of Glaciological Society."
Geomorphology, Rhodes W. Fairbridge,
Civil Engineers, vol. 128, figure 2,
page
ed., Dowden, Hutchinson & Ross,
289, 1963. Used with permission by the
Stroudsburg, Pa. Figure 7.30: From Chapter 11
American Society of Civil Engineers.
Rouse Geology, vol. 6, page 407.
et al., in
Figure 6.3: From Vanoni, Transactions of Figure 1 1.5: From Gold and Lachenbruch
© 1978 The Geological Society of in"Permafrost: 2d International
the American Society of Civil Engineers,
America. Reprinted by permission. Figure Conference Proceedings," 1973, with
vol. 3, figure 24, page 127, 1946. Used
7.31: From Morgan, in Journal of
permission of the National Academy of
with permission by the American Society
Geological Education, vol. 18, page 1 13.
of Civil Engineers. Figure 6.7: From Sciences, Washington, D.C.
© 1970 National Association of Geology
Hjulstrom 1939. Used with permission of
Teachers. Reprinted by permission.
American Association of Petroleum
Chapter 12
Geologists. Figure 6.19: From Hey & Figures 12. 1 and 12.2: From A. N.
Thome Area, Vol. 7, figure 4, page
in Chapter 8 Palmer, 1984, Journal of Geological
193. ©1975 The Institute of British From Mabbutt, in Aboriginal
Figure 8.1:
Education, v. 32, p. 248. Figure 12.8:
Geographers. Reprinted by permission. Man and Environment in Australia, D. J.
After Williams 1972b in Geological
Figure 6.25: From Schumm and Khan in Mulvaney and J. Golson, editors. © 1971 The ©
Society of America Bulletin.
Geological Society of America Bulletin, Australian National University Press.
Geological Society of America. Reprinted
vol. 83, page 1768.© 1972 The Reprinted by permission. Figures 8.4, 8.5,
by permission. Figure 12.9: "From Baker.
Geological Society of America. Reprinted and 8.8: From R. A. Bagnold, 1941, The
American Geophysical Union Water
by permission. Figure 6.27: From Physics of Blown Sand and Desert
Resources Research. 1973, copyrighted
Dunes." Used with permission of
by American Geophysical Union."
571 Credits
Reprinted by permission. Figure 12.22: Chapter 4 Chapter 10

From the Geological Society of America Figure 4.1: R. R. Dutcher. Figure 4.3: Figure 10.2: D. F. Ritter. Figure 10.5:
Bulletin, vol. 79, figure 14, page 39. R. R. Dutcher. Figure 4.4: F. C. Calkins. A. Werner. Figure 10.6: Austin Post,
© 1968 The Geological Society of Figure 4.5: W. T Schaller, USGS. Figure USGS. Figure 10.9 a: F. E. Mathes,
America. Reprinted by permission. 4.6: D. F. Ritter. Figure 4.16: G. W. USGS. Figure 10.12: R. R. Dutcher.
Stose, USGS. Figure 4.19 a, b: USGS. Figure 10.13 a: J. H. Moss, USGS.
Figure 4.21 b: USGS. Figure 4.23 a: Figure 10.15 a: R. R. Dutcher. Figure
Chapter 13
USGS. Figure 4.23 b: J. B. Hadley, 10.15 b: J. H. Moss, USGS. Figure 10.15
Figure 13.7: "Adapted from Gavin,
USGS. Figure 4.23 c: C. F. Erskine, c: J. Travis. Figure 10.17: D. F. Ritter.
Journal of Geophysical Research, 1968,
USGS. Figure 10.18: R. R. Dutcher. Figure
copyrighted by American Geophysical
10.19: J. Travis. Figure 10.20: C. William
Union." Reprinted by permission. Figure
Horrell. Figure 10.21 b: W. C. Alden,
13. 1 "Adapted from Shepard and
1 :
Chapter 5 USGS. Figure 10.23: D. F. Ritter.
Inman, AGU Transactions, 1950, Figure 5.13: USGS. Figure 5.24: D. F.
copyrighted by American Geophysical Ritter.
Union." Reprinted by permission. Figure Chapter 11
13. 19: From W. A. White, in Geological Figure 11.1: O. J. USGS. Figure
Ferrians,
Society of America Bulletin, vol. 77. Chapter 6 1 1 . 1 1 : O. J. USGS. Figure
Ferrians,
© 1966 The Geological Society of Figure 6.4 a, b: W. W. Emmett, USGS. 11.12: G. W. Holmes, USGS. Figure
America. Reprinted by permission. Figure Figure 6.9: D. F. Ritter. Figure 6.22: 11.13: O. J. Ferrians, USGS. Figure
13.20: From May et al., issue 35, 1983, D. F. Ritter. Figure 6.24: USGS. 11.14: R. Craig Kochel. Figure 11.15:
page 522. Copyright by the American Noel Potter, Jr. Figure 1 .1 6: O. J. 1
Geophysical Union. Reprinted by

Chapter 7 Ferrians, USGS. Figure 1 1.18: D. F.
permission. Figure 13.22: From Ritter.
Figure 7.1: D. F. Ritter. Figure 7.4 b:
Sunamura 1982. © by The University of
USDA. Figure 7.9 b: R. Guillemette.
Chicago Press. Used by permission.
Figure 7.17: USGS. Figure 7.18 a: Chapter 12
USGS. Figure 7.22: R. R. Dutcher. Figure 12.4: D. F. Ritter. Figure 12.10:
Figure 7.24: H. E. Malde, USGS. C. William Horrell. Figure 12.12: D. F.
Photos Ritter. Figure 12.16: USGS. Figure
12.17: W. H. Monroe, USGS. Figure
Chapter 8
Frontispiece 12.18: W. H. Monroe, USGS.
Figure 8.3: D. F. Ritter. Figure 8.7: M. R.
G. K. Gilbert, USGS Campbell,USGS. Figure 8.14 a: G. K.
USGS. Figure 8.14 b: C. E.
Gilbert, Chapter 13
Erdmann, USGS. Figure 8.14 c: USDA. Figures I3.l and 13.2 a, b: R. Craig
Chapter 1
Figure 8.17: D. F. Ritter.
Figure 1.1: © John Shelton. Figure 1.2:
Kochel; Figure 13. 10: UPI; Figure 13.12:
NASA. Figure 1.7: T. S. Lovering, P. Michaels; Figure 13.17: US Coast

USGS. Figure G. W. Stose, USGS. Guard and Geodetic Survey; Figure
1.8: Chapter 9
13.18: USGS. Figure 13.24: USGS.
Figure 1.9: W. T. Lee, USGS. Figure Figure 9.3: US Coast Guard and Geodetic
Figure 13.25: W. H. Monroe, USGS.
1.10: R. R. Dutcher. Survey. Figure 9.18: M. F. Meier, USGS.
Figure 13.26 (both): USGS. Figures
Figure 9.19: M. F. Meier, USGS. Figure
13.29-13.32: R. Craig Kochel.
9.21: D. F. Ritter.
Chapter 2
Figure 2.5: NASA. Figure 2.6:
R. Guillemette. Figure 2.8: USGS.
Figure 2.9: H. Miller Cowling.
Chapter 3
Figure 3.2: D. F. Ritter.
1
Index
Ablation till, 381 Barrier islands, 524-28 Calving, 341

Abrasion pH, 66, 67 characteristics, 524, 525 Canada, 260, 266, 367, 414, 423, 424,
Active layer, 407 distribution, 524 437, 440, 492
Active subsole drift, 355 features, 524-28 Canol pipeline, 440
Adsorption, 71, 72 ebb tidal delta, 524, 527, 528 Cantilevers, 218
Advection, 14 flood tidal delta, 524, 527, 528 Capillary fringe, 177
Africa, 317, 327, 328, 329 lagoons, 524, 527 Cation exchange capacity, 71, 75
Alaska, 41, 72, 239, 338, 339, 370, 371, washover fans, 524, 525 Caves, 474-79
381, 391, 394,401,408,411,423, processes, 525-28 origin, 477-79
425, 428, 433, 437, 438, 440-42, inlet formation, 525, 527, 528 models, 477-79
520 overwash, 524, 525, 526 patterns, 476, 477
Allogenic valleys, 469 wind action, 525, 526 branchwork, 476
Alluvial fans, 104, 105, 275-85 Base flow, 157, 158 maze, 476, 477
characteristics, 275, 276 Base saturation, 72, 88 physiography, 474-76
morphology, 278-81 Basin and Range, 42 entrances and terminations, 474, 475
origins, 281-85 Basin morphometry, 164-76 rooms and passages, 475, 476
282-85
depositional processes, Baymouth bars, 522, 523 Chattermarks, 367
entrenchment, 282, 283-85 Beach cusps, 502, 508, 509-10 Chelation, 76, 77
types, 275 Beaches, 498-507 Chemical weathering. See Decomposition
zones,276-78 beach profile, 499-501 Chezy equation, 208
abandoned washes, 276, 278 beach face, 499, 500 Chimney, 474
desert pavement, 276-78 berm, 499, 500 China, 413, 414
modern washes, 276-78 longshore bar, 500, 501 Chlorite, 80, 81
Amazon River, 257, 492 morphodynamics, 501-4 Chute, 258-60
Antarctica, 35, 146, 338, 340, 356, 413 modal state, 504 Cimarron River, 231
Appalachian Mountains, 20, 21, 25, 42, morphodynamic states, 502-4 Cirque glaciers, 372, 373
191,432,447,459 summer profile, 501, 502 mechanics of, 372, 373
Aquifer, 179-81 wave climate, 504 Cirques, 369-75
Arete, 370, 371 winter profile, 501, 502 characteristics, 369, 370
Arizona, 290, 460 Bed material load, 212 headwall erosion, 374, 375
Artesian well, 179 Belt of no erosion, 162, 163 processes of, 374, 375
Atchafalaya River, 298, 299 Bergschrund, 374 origin, 370, 373
Atterburg limits, 124 Bifurcation, 164 Clay minerals, 79-83

in clays, 124 Bifurcation ratio, 166 mixed layer, 79, 8
liquid limit, 124 Biogenic C0 2, 451,452 1 clay, 79, 80
: 1
plastic limit, 124 Blind valleys, 469 kaolinite, 79, 80, 81, 82
Australia, 203, 250, 251, 252, 305, 306, Breaking strength, 125 2:1 clays, 79, 80,81
325, 328, 329,451, 501-3 British Columbia, 238 illite, 80, 81
Buried soils, 103 smectites, 80, 81
Backwash, 490, 500, 504, 510 vermiculite, 80, 81
Bajada, 276 California, 3, 4, 26, 27, 41, 43, 44, 45, Climate, 12-15, 50, 53-62
Bank erosion, 218-20 115, 116, 180, 181, 201, 278, 279, classification, 50
Barbados, 56 280,281, 283,289,312,375,495, climate change, 53-62
501,502, 506, 519
573
5 7 7
Index 574
Climate change, 53-62 Crevasses, 359, 360 Drainage composition. 164-76

cause of, 54 Crevasse splays, 299-301 Drainage density, 167-70, 172. 174, 190,
geomorphic response, 54-62 Cuba, 472 191
precipitation and runoff. 59-61 Drainage networks, 153-55, 161-76
sea level fluctuation, 54-58 Darcy'sLaw, 178,458 basin evolution, 171-76
sediment yield, 59-62 Death Valley, 276, 281 basin morphometry, 164-76
vegetation, 60-62 Debris flow, 282-85 areal relationships. 166-70
Climate-process systems, 51-53 Decomposition, 64-84 linear relationships, 166, 167
Climatic geomorphology, 2, 50-53 end products. 78-83 relief relationships. 170, 171
climate-process systems, 51-53 clay minerals, 79-83 stream orders, 164-67
morphoclimatic zones, 51 hydrous oxides, 83 initiation of channels, 161-64
morphogenetic regions. 51-53 processes 66-71
of, belt of no erosion. 162. 163
morphogenetic systems, 51-53 hydrolysis. 70-71 bifurcation, 164
Coast. 482, 483 ion exchange, 71 critical length, 162
classification, 483 oxidation-reduction, 68 cross-grading. 163
definition, 482, 483 solution. 69 micropiracy. 163
Coastal deposition, 522-28 rates of. 77-84 Drift potential, 326. 327
features. 522-28 chemical analyses, 83-84 Driving force, 11-18
barrier islands, 524-28 mineral stability, 78-83 climate, 1 1-14
baymouth bars, 522, 523 Deltas. 294-302 gravity, 15, 16
spits, 522, 523 dynamics and evolution, 297-302 internal heat, 16-18
Coastal erosion, 513-21 Mississippi delta, 298-302 Drumlins, 395-400
landforms, 516-21 parts of, 295 character and distribution, 397, 398
sea arch, 520, 521 deltaic plain, 295 origin. 397-400
sea caves, 520 delta slope, 295 Dry 470
valleys, 469,
sea cliff, 515-19 prodelta environment, 295 Dunes. 319-27
stacks,520 types, 295-97 components of, 319, 320
wave-cut platform, 519, 520 high-constructive, 295-96 forms. 321-27
processes. 5 1 high-destructive, 296, 297 barchan, 322-24
rates of. 513. 514. 515-18 Denudation, 191-203 linear, 322-24
Coastal processes, 483-85 rates of, 201-3 parabolic, 322-23
data needed. 483, 484 sediment budgets, 200, 201 seif, 323
research center. 483-85 sediment yield, 194-203 origins. 320, 321, 323-27
Coastal storms, 494, 496-98 Desert pavement, 306, 324 relation to wind, 325-27
extratropical, 494, 496, 498 Desert Soil —Geomorphology Project, Dynamic equilibrium concept, 7
storm surge, 498 103-7
tropical. 496, 497, 498 Desert varnish, 306 Earth hypsometry, 35. 36
Col, 370 Differential weathering, 22, 23, 24 Edge waves, 495
Colluvium, 257 Dilatant material, 399 EDTA, 77
Colorado, 113, 133, 243, 288, 3 1 Discharge, 180-82.221, 222 Effective normal stress. 121-23
Colorado Plateau, 38 dominant, 221, 222 Eh. 69. 75. 76
Colorado River, 248 Discontinuous permafrost, 413, 414 Eluviation. 85, 87
Competence, 214-17 Disintegration, 64, 110-18 End moraine, 386. 388
Complex response, 25-27 expansion, 111-16 Endogenic process. 11. 34-50
examples, 26, 27 hydration and swelling. 1 12-16 Energy, 10, 36,37
Complex soils, 103 thermal, 1 1 1 kinetic, 10
Compressive flow, 353, 354 unloading, 1 12 potential, 10
Conduction, 12, 16, 17 growth in voids, 116-18 potential energy field, 36, 37
Cone karst, 470 frost action, 1 1 geoid, 37
Confined aquifer, 179 significance of water, 1 18 spheroid, 37
piezometric surface, 179 Dissipative beach. 502. 503 England. 369
Continuous permafrost, 413, 414 Divergent flow. 234-36 Enhanced creep. 347, 348
Controlling obstacle size, 348 Doline, 463-68 Epeirogeny. 35-41
Convection. 12, 17 collapse doline, 466, 467 cause of. 36-41
Convergent flow. 234-36 morphometry, 467, 468 isostasy. 38-41
Cordilleran, 20, 21 465
solutional doline, 464, potential energy field, 36, 37
Creep, 129-32 Draa. 317, 318, 324 rates of. 39-41
continuous. 1 3 1 Drag velocity, 309 Equilibrium. 7, 8
in glaciers, 343-46 Drainage basin, 153-204 dynamic, 8
rates, 130. 132 basin hydrology, 176-91 static, 8
seasonal, 129, 130 denudation, 191-204 steady-state, 8
drainage patterns, 153-55 Equipotential surfaces, 177, 178
575 Index
Erg, 304, 317 Glacial abrasion, 364-67 mass balance, 341, 342
Ergodic hypothesis, 151 features of, 365-67 origin, 335-38
Eustatic sea level change, 55, 56 chattermarks, 367 firn line, 336
rates and magnitude, 55, 56 grooves, 366, 367 processes, 337, 338
Evapotranspiration, 156 striations, 365, 366 types, 338-41
Exfoliation, 1 15 rates of, 364, 365 polar, 340, 341
Exhumed soils, 103 Glacial budgets, 341-43 temperate, 340, 341
Extending flow, 353, 354 ablation,341-43 Glen power flow law, 345, 346
accumulation, 341-43 Graded river, 248, 249
Fan-delta, 294, 295 equilibrium line, 342 Gravity, 15-16, 36-41
Fan-head trench, 282-85 gross budget, 343 anomalies, 37, 38
Fetch, 486 net specific budget, 343 Bouger, 38
Field capacity, 87 Glacial deposition, 379-403 isostatic, 38
Fiord, 378 depositional framework, 383-403 potential energy field, 36
Firn, 337, 372 zones and features, 384-403 Greenland, 39, 40, 418, 420, 421
Firn line, 336, 342 drift,379-83 Ground moraine, 348, 395, 396
Fixation, 76 Glacial drift, 379-83 Ground water, 176-81
Flood-damage stage, 257 fluvioglacial, 382, 383 aquifers, 179
Flood hydrograph, 157 till, 380, 381 environmental problems, 180, 181
Floodplains, 256-67 types, 381 overdraft, 180, 181
deposits and topography, 257-62 Glacial erosion, 364-78 saltwater intrusion, 181
chutes, 258-60 features of, 365-78 subsidence, 180
clay plugs, 261 chattermarks, 367 movement, 177-79
lateral accretion, 257 cirques, 369-75 Darcy's law, 1 78
meander scrolls, 257-60 glacial troughs,375-79 equipotential surfaces, 177, 178
natural levee, 260, 261 grooves and furrows, 366-67 potential, 177
vertical accretion, 257, 265 striations, 365, 366 rules, 178
origin, 263-67 processes, 364-69 profile, 177
lateral migration, 262-67 abrasion, 364-69 water table, 177
overbank deposition, 264-67 quarrying, 364-69 zones, 177
types, 256, 257 Glacial movement, 343-57 wells, 180, 181
Florida, 181,447,523 complications of, 351-57 artesian, 179
Flow duration curve, 183, 184 extending and compressive flow, cone of depression, 1 80
Flows, 138-44 352-54 drawdown, 180, 181
Flow till, 381 kinematic wave, 355, 356 Grus, 113, 114, 295
Fluid threshold, 310, 311 surging glaciers, 356, 357
Force, 9, 10 internal, 343-51 Hanging valley, 375, 376
definition, 9 controls, 343, 344 Hawaii, 47, 48, 77, 414
ingeomorphology, 11-18 creep, 343-46 Head. See Potential
measurement, 10 Glen power flow law, 345, 346 Heave, 126, 129-31
relation to energy, 10 sliding, 343, 346-49 Hong Kong, 510
relation to mass, 9, 10 controlling obstacle size, 348 Horn, 370, 371
France, 364, 469 enhanced creep, 347, 348 Humus, 85-87
Frost action, 117, 372, 374, 375, 415-19 regelation, 347, 348 Hurricane, 496, 497, 498
frost cracking, 419 slippage, 346, 347 Hydration, 112-16
frost heaving and thrusting, 417-19 velocity and flow, 349-51 of clays, 112, 113
frost pull, 418, 419 Glacial quarrying, 364, 367-69 Hydration shattering, 374, 375, 416
frost push, 418, 419 features of, 367-69 Hydraulic conductivity, 178, 179
frost wedging, 416, 417 process mechanics, 367-69 Hydraulic geometry, 222-28
Frost cracking, 419 Glacial surges, 356, 357 Hydraulic gradient, 179
Frost heaving, 417-19 trigger zone, 357 Hydrologic budget, 176
Frost shattering, 374, 375, 416 Glacial valleys, 375-78 Hydrologic floodplain, 256
Frost thrusting, 417-19 characteristics of, 375-77 Hydrolysis, 70
Froude number, 208 fiords, 378 Hypsometric analysis, 170, 171
Fulvic acid, 77 longitudinal profile, 375, 376-78 Hypsometric integral, 170, 171
origins, 376-78
Gelifluction. See Solifluction Glaciers, 335-62 Ice contact zone, 384-400
Gelifluction lobe, 430 classifications, 338-41 features of, 386-400
Geochemical weathering, 65 definition, 336 eskers, 393-95
Geocryology, 406 ice structures, 358-61 flutes and drumlins, 395-400
Geoid, 37 crevasses, 359, 360 ground moraine, 395, 396

foliation, 359 kames and kettles, 383, 393, 394
stratification, 358 kame terrace, 393
moraines, 386-92
Index 576
Ice fabric, 338, 345 Karst, 69, 445-79 flows, 133, 135, 138-40
Iceland, 47, 48, 389, 394 characteristics, 446, 447 debris avalanche. 139
Ice streaming, 369 definitions, 446 debris flows, 139
Ice structures, 358-61 rocks, 449-50 earthflow and mudflow, 139, 140
secondary, 359-61 terminology, 448 sturzstroms, 138, 139, 142
crevasses, 359, 360 Karst denudation, 454 frost creep, 419, 420
foliation, 359 rates of, 454, 455 morphology, 140-44
stratification, 358 Karst hydrology. 455-62 morphometric indices, 140-44
Ice wedge, 408, 421-24 aquifer types, 458, 459 processes, 126-44
Ice-wedge cast, 424 confined flow, 458, 459 slides, 126, 132-38
Ice wedge polygons, 423, 424 diffuse flow, 458,459 debris slides, 132, 134, 135
Idaho. 268 free flow, 458,459 rockslides, 132, 134, 135
Ignimbrite. 47, 50 springs, 456, 459, 460 rotational slip, 132, 134, 135
Illinois, 52, 194, 258, 259, 260. 328, 329, subcutaneous zone, 457 translational slides, 132, 134, 135
330,331,425,463,465 surface flow, 455, 456 solifluction, 420, 421
lllite. 72, 80, 81 drainage morphometry, 460-62 stability factors, 126-29
306
llluviation, 85, 87, 88, Karst landforms, 462-73 Medial moraine, 386, 395
Impact threshold. 311, 312 closed depressions, 462-69 Melt-out till. 381
Indiana. 329, 447, 459, 464, 477 dolines, 463-68 Mexico, 180
Infiltration, 156-61 uvalas and poljes, 469 Michigan, 151, 172, 173,379
Infiltration capacity, 157, 158, 161 valleys, 469, 470 Mississippi, 45
Inselbergs, 288 Karst processes, 451-55 Mississippi delta. 295-302
Insolation, 12 basic controls, 451 Mississippi River, 58, 219, 257, 259, 260,
Interception, 156 solution, 452-54 261,328,329,383,499
Interlocking friction, 121, 122 mixing corrosion, 453 Missouri, 58, 219, 260, 447. 457
Internal forces, 34-50 vadose flow, 453, 454 Missouri River, 330
diastrophism, 35-46 vadose seepage, 453, 454 Mixing corrosion, 453
epeirogeny, 35-41 Karst rocks, 449, 450 Mobility. See Ion mobility
orogeny, 42, 43 lithology, 449 Mogotes, 472, 473, 474
volcanism, 46-50 porosity and permeability, 449, 450 Mojave Desert, 286, 293-94
Internal friction, 121, 122 Karst topography, 446 Montana, 111, 117, 141, 150, 183, 184,
Internal heat, 16-18 Kentucky, 133, 151,447,468 185, 186,285,324,380,392,398
Ion exchange, 71 Kinematic wave, 355, 356 Montmorillonite, 80, 8 1 , 112, 113
Ionic potential, 72, 73 Kinetic energy, 10 Moraines, 342, 343, 386-92
Ion mobility. 72-77 Knickpoint, 246-48 definition, 386
controls of, 72-77 types, 247-48 formation of, 388-92
chelation, 76-77 types, 386-88
Eh, 75, 76 Lateral accretion, 257-67 Morphogenetic systems. 51-53
fixation, 76 Lateral migration, 262-67, 269-72 Mount St. Helens, 28, 29
leaching, 73-75 rates, 262 Murrumbidgee River. 250-52
pH. 74, 75 Lateral moraine, 386, 387
ionic potential, 72 Leaching, 73-75 Natural levee. 260, 261
Iowa, 330 Length ratio, 166 Neap tide, 493
Isostasy.38-41 Litter, 85 Nebraska. 242. 317, 329
depth of compensation, 38 Littoral drift, 504-7 Needle ice, 419
glacio-isostasy, 38-40 budget. 505 Nevada, 150, 180, 181, 272, 280. 281.
Greenland, 39, 40 rates of. 505-7 289, 307
isostatic adjustment, 39-41 swash transport, 504. 505 New Guinea. 56. 461, 467
rates of, 39-41 Littoral zone, 499 New Hampshire. 414
isostatic anomaly, 38 Lodgement till. 381 New Jersev. 171. 181.494
Italy. 1 12,446 Loess, 327-33 New Mexico. 103-7. 150, 324. 451
Louisiana, 45, 298 New York, 114. 181.366.383.398.456.
Jamaica. 470 460.461,462
.lapan, 515-18 Manning equation, 208-10 New Zealand, 140, 146
.lokulaup, 402 Maryland, 84, 130. 229, 230, 263. 527, Nivation, 372
528 North Carolina, 264, 483. 484. 485. 4^S.
Kansas. 231, 244, 329 Massachusetts, 506, 508 526. 527
Kaolinite, 70. 80. 81. 82,83 Mass movement, 126-44,419-21 Norway, 373
origin, 81 -83 classification, 133, 135
creep, 129-32 Ocean currents. 493. 495. 496
falls and topples. 129, 132, 133, 135 longshore currents. 495. 4 >o l
rip currents. 493. 495

1 1 11 1
577 Index
Ocean swell, 487 Periglacial processes, 415-21 Puerto Rico, 472, 473, 474, 521
Ogives, 339, 360 frost action, 415-19 Pyramid Lake, 272
Oregon. 47, 50, 324 segregation potential, 415, 416
Orogeny, 35, 36, 42 mass movements, 419-21 Quaternary geomorphology, 34, 102-7
Oscillatory waves, 487, 488 Permafrost, 406, 407-14
Outwash, 383, 384, 400-403 components of, 407-9 Rainfall-erosion facility, 172
Overbank deposition, 263-67 active layer, 407, 409 Rainfall intensity, 158, 162
rates, 265 permafrost table, 407, 409 Rating curve, 182, 183
Overwash, 494, 498, 524, 525, 526 definition, 407 Recovery time, 222
Oxidation-reduction, 68, 75, 76 distribution, 413, 414 Recurrence interval, 184, 185
continuous permafrost, 413, 414 Reflective beach, 502, 503
Paleosols, 103 discontinuous permafrost, 413, 414 Regelation, 347, 348
Pans. 314 sporadic permafrost, 413, 414 Regolith arch, 466
Partial area concept. 160, 161 talik, 409, 410,414 Relict soils, 103
Patterned ground, 426-29 origin, 410, 41 Relief ratio, 170
classification, 426, 427 geothermal gradient, 410, 41 Resistance, 17-25
circles, 426, 427, 428 thickness, 411, 412 Response time, 356
polygons, 426, 427 Permafrost table, 407 Resultant drift potential, 326, 327
stripes, 426-28 Permanent wilting point, 87 Retrograde movement, 419, 420
origin, 428, 429 Peru, 138,491 Return flow, 160
Ped. 85, 86 pH, 74, 75 Reynolds number, 207
Pediment, 286-94 controls of, 74-75 Riffles and pools, 233-35, 236, 243
definition, 286 Phreatic zone, 177 Rill, 158, 162-64, 193
formative models, 291-94 Physical weathering. See Disintegration Rio Grande River, 103-7
morphology, 287-89 Piedmont, 104, 105, 275-94 River adjustment, 231, 245-53
piedmont angle, 288, 289 Piedmont angle, 288-94 Cimarron River, 231
size and shape, 287 Piezometric surface, 179 gradient, 246-48
slope, 289 Pingo, 424-26 knickpoints,246-48
surface topography, 288, 289 closed-system, 424 Murrumbidgee River, 250-52
processes, 290, 291 open-system, 425 river metamorphosis, 249-52
Pediment association, 290 Piracy. See Stream capture shape and pattern, 249-52
Pedochemical weathering, 65 Plane friction, 121, 122 River bars, 233, 235, 242
Pedogenic regimes, 85-102 Plasticity index, 123 River channels, 232-44
calcification, 99-101 Pocket valleys, 470 anastomosed, 242
laterization, 98, 99 Polje, 469 braided,239-42
podzolization, 98 Polygenetic soils, 103 239-41
origin,
soil forming factors, 85-102 Polygonal karst, 467 meandering, 235-38
Pennsylvania, 42, 266, 432, 455 Pore pressure, 122, 123 characteristics, 235
Perigean spring tide, 493, 494 Potential, 177-79 geometry, 236-38
Periglacial environment, 406-42 Potential energy, 10 pattern continuity, 242-44
definition, 406 Potential surface of sliding, 135, 137, 138 straight rivers, 232-34
engineering in, 434-42 Power flow law, 345, 346 features, 233, 234
approaches, 434, 435 Primary porosity, 450 sinuosity, 232
building foundations, 435, 437 Process, 3-6, 1 River hydrology, 181-91
pipelines. 440-42 definition, 3, 1 basin characteristics, 188-91
relation to landforms, 436 human influence, 3, 4, 27 lag, 188-91
roads and airfields, 437-39 in planetary studies, 4, 5 discharge, 181-83
utilities, 439 types, 1 frequency and magnitude, 183-88
processes, 415-21 Process geomorphology, 6-31 mean annual, 182, 183
Periglacial landforms, 421-34 basic principles, 6-31 mean daily, 182
associated with mass movement, complex response, 26-27 floods,183-91
429-34 delicate balance, 7, 8 mean annual flood, 185, 186
blockfields, 430-32 force, 9-16 gaging stations, 182-84
gelifluction features, 430 process linkage, 27-29 rating curve, 182, 183
rock glacier, 432-34 resistance, 17-25 stage, 184
associated with permafrost, 421-26 thresholds, 25-26 morphometric relationships, 190-91
ice wedge polygons, 423, 424 time framework, 29-31 paleofloods, 187, 188
ice wedges, 421-23 rationale, 1-6 recurrence interval, 184, 185
pingo, 424-26 Proglacial environment, 384, 400-403 Susquehanna River, 188-90
thermokarst, 426 features of, 400-403 River landforms, 255-302
patterned ground, 426-29 sandar, 400-403
1 1 1
Index 578
River mechanics, 206-53 Saltwater intrusion, 181 Slopes, 144-52

bank erosion, 218-20 Sandar, 400-403 angles, 144-48
processes, 218-20 types, 401 evolution models, 150-52
equilibrium, 222-32, 245-53 zones, 402,403 strength equilibrium, 146
channel shape, 231, 232 Sand wedge, 422 talus, 147, 148
graded river, 248, 249 Saprolites, 65 transport-limited, 146
hydraulic geometry, 222-28 Sea cliff, 499, 515-18 weathering-limited, 146
slope factor, 228-31 erosion rates,515-18 Slope wash, 149
time factor, 245-53 Secondary minerals, 79-83 Snowline, 336, 370, 372
flow equations, 208 clay minerals, 79-83 Soil, 64, 84-107
Chezy equation, 208 hydrous oxides, 83 available water capacity, 86
Manning equation, 208 Secondary permeability, 450 classification, 90-94
flow types, 207, 234-36 Sediment budget, 200, 201 definition, 64, 84
convergent, 234, 235, 236 Sediment yield, 192-201 geomorphic significance, 65, 102-7
divergent, 234, 235, 236 delivery ratio, 199, 200 pedogenic regimes, 95, 97-102
helical, 235 factors of, 194-200 physical systems, 102
laminar, 207 universal soil loss equation, 194 Quaternary geomorphology, 102-7
turbulent, 207 Segregation potential, 415, 416, 418 soil catena, 96, 97
frequency and magnitude of work, 221, Seiche, 492 Soil classification, 90-94
222 Sensitive soils, 124, 125 Soils Conservation Service
recovery time, 222 Shaft, 459, 474 components of, 91-94
resistance factors, 207-1 Shear strength, 121-26 development of, 90, 91
River sediment, 21 1-22 rocks, 125, 126 Soil profile, 64, 85-90
deposition, 220, 221 classification, 126 horizons, 85, 87-90
entrainment, 214-17 unconsolidated material, 121-25 descriptive terms, 88-90
competence, 214-17 Sheeting, 1 12 properties, 85-87
critical bed velocity, 214-17 Sheet wash, 290, 291 Soil wedges, 424
critical shear stress, 214-17 Shield volcanoes, 47, 48, 49 Solar constant, 1
stream power, 217 Shoaling, 488, 489 Solifluction, 131,420,421

load types, 211, 212 Shoreline configuration, 508-13 Solution, 69, 452-54
transportation, 211-13 beach cusps, 508-10 South Africa, 118,467, 525
bedload discharge, 212, 213 capes, 511, 512 South Carolina, 264
River terraces, 45, 57-59, 104, 267-74 rhythmic topography, 51 South Dakota, 141
components of, 267, 268 Shoshone River, 273, 274 Spheroid, 37
field problems, 272-74 Sieve deposits, 283 Spheroidal weathering, 115, 116
origin, 269-72 Sinkhole. See Doline Spits, 522,523
types, 267-72 Sinuosity, 232 Splay deposits, 257
depositional, 267, 269, 270 Slides, 126, 134, 135-38 Sporadic permafrost, 414
erosional, 267, 271, 272 Slope erosion, 192-94 Springs, 459, 460
paired, 269 rainsplash, 192, 193 exsurgences, 459
unpaired, 269 wash, 193 resurgences, 460
Roche moutenee, 368, 369 Slope hydrology, 156-61 Spring tide, 492, 493
Rock-cut terrace, 267, 272-74 infiltration, 156-58 Stage, 183
Rock glacier, 432, 433-34 base flow, 1 57 Static equilibrium, 8
origins, 433, 434 infiltration capacity, 157, 158 Steephead, 470
types, 433 slope hydrologic cycle, 56 1 Stoss and lee topography, 368
lobate, 433 evapotranspiration, 156 Strandlines, 39, 40
tongue-shaped, 433 interception, 156 Stream capture, 175, 176, 270
Rogen moraine, 386, 390 Slope material, 18-25 1 Stream power. 217
Ruggedness number, 170 forces on, 120, 121 Strength equilibrium envelope, 147
Runoff, 156-61 effective normal stress, 122, 123 Striations, 365, 366
base flow, 1 57, 1 58 shear stress, 121 Sturzstroms. 138, 139. 142
direct runoff, 156, 157, 158 properties, 1 18-25 Suballuvial bench, 286
Hortonian overland flow, 158 cohesion, 123 Subcutaneous zone, 457, 458, 468
saturated overland flow, 160 effective normal stress, 122, 123 Suction potential. 415
return flow, 160 internal friction, 121, 122 Superimposed ice, 342
subsurface stormflow, 159, 160 shear strength, 121-25 Surf beat. 490
ihroughflow, 159 Slope profiles, 144, 145 Surf-scaling parameter, 502, 503
unit hydrograph, 157 controls of, 148-50 Surf zone. 490
variable source concept, 160. 161 geometry, 144, 145 Susquehanna River. 188-90
limiting angles, 144 Swallei. 455
1
579 Index
Swash, 490, 500, 504,510 Unconfined aquifer, 179 Wave set-up, 493, 495
Swash zone, 490 Unit hydrograph, 157 Wave 487-89
steepness,
Systems, 7, 8 Universal soil loss equation, 194 Weathering, 64-84
Uplift, 39-41 decomposition, 64-84
Tafoni, 114, 115 rates of, 40, 41 processes, 69-7
Talik, 409, 410, 414 U-shaped valley, 375, 376 disintegration, 64
Talus, 147, 148 Utah, 227, 286 Whalcback form, 368
Tarn, 369 Utilidor. 439 Wind, 303-33
Tasmania, 431 Uvala, 469 desert environment, 304-8
Tectonic geomorphology, 41-46 climate, 304
criteria of, 43-46 Vadose flow, 453, 454 surface armor, 306, 307
faulting, 44 Vadose seepage, 453, 454 weathering and soils, 306-8
marine terraces, 44 Vadose water, 177 Wind action, 308-10, 315, 316
river profiles, 45 Ventifact, 312, 313, 314 controls of, 308-10, 315, 316
scarp retreat, 43 Vermiculite, 1 16 drag velocity, 309
uses of, 43-46 Vermont, 160, 161 effect of saltation, 315,316
Tennessee, 466, 468 Vertical accretion, 257-67 turbulence, 310
Tephra, 46 Vietnam, 472 deflation, 307, 314
Terminal moraine, 386 Virginia, 229, 230, 467 Wind 314-32
deposition,
Texas, 114, 180, 181, 187,447 Volcanic landforms, 47-50 features,317-32
Texture ratio, 168, 169 calderas, 50 bedforms, 317-26
Thalweg, 233, 235 cones, 47-50 sand seas, 317
Thermokarst, 426 composite cone, 49, 50 sand sheets, 317
Thresholds, 6, 25-26 shield volcanoes, 47-49 loess,327-32
extrinsic, 25 lava plains and plateaus, 47 Wind erosion,310-14
geomorphic, 26 Columbia River plain, 47 features of, 312-14
intrinsic, 26 Deccan Plain, 47 grooves, 314
Throughflow, 159 Volcanism, 46-50 pans, 314
Tidal bore, 492 ventifacts, 312,313
Tides, 492-94 Wash. See Slope wash yardang, 313, 314
classification, 493 Washboard moraine, 386, 390 processes, 310-12
neap tide, 493 Washington, 28, 29, 47, 48, 49, 239, 353, fluid threshold, 310, 311
perigean spring tide, 493, 494 354, 403 impact threshold, 311, 312
spring tide, 492, 493 Wash load, 212 Wind ripples, 317, 318-19
Till, 380-81 Watershed. See Drainage basin formation, 319
characteristics, 380 Water table, 157-60 Wind transportation, 314-32
types, 381 Wave cut platforms, 44, 519, 520 saltation, 314-16
Time in geomorphology, 8, 29-31 Wave dispersion, 486-88 surface creep, 314, 316
cyclic time, 8, 30 Wave period, 486-88 Wisconsin, 385, 387
graded time, 8, 30 Waves, 486-92 Wyoming, 113, 184, 212, 226, 273, 274,
steady time, 8, 29, 30 breaker types, 489 313
Tombolo, 523 refraction of, 490, 491
Topographic floodplain, 256 types, 487-92 Yardang, 313,314
Tractive force. See River sediment seiche, 492 Yellowstone Plateau, 50
Trans-Alaska pipeline, 440-42 tsunami, 491 Yield stress, 344
Translation waves, 488 waves of oscillation, 487, 488 Young pits, 129
Tropical karst, 470-73 waves of translation, 488 Yugoslavia, 446
features of, 470-72 wave generation, 486-89
cockpits, 470, 471 dispersion, 486, 487 Zero annual amplitude, 409, 410
towers, 470. 472, 473 fetch, 486 Zone of aeration, 1 77
473
origin of, 472, period, 486-88 vadose water, 177
Truckee River, 272 Zone of soil moisture, 177
Tsunami, 491
inch
Wm. C. Brown Publishers
Dubuque, Iowa
SBN D-bq7-D5D4?-5

Process Geomorphology

Uploaded by

Copyright:

Available Formats

Process Geomorphology

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Process Geomorphology

Uploaded by

Copyright:

Available Formats

I

Chairman of the Board Lawrence E. Cremer President

Cover photo: © Kathleen Norris Cooke

Copyright © 1978. 1986 by Wm. C. Brown Publishers. All rights reserved

Library of Congress Catalog Card Number: 85-70934

Printed in the United States of America

Climate and Internal Forces 33

Chemical Weathering and Soils 63

Physical Weathering, Mass Movement, and Slopes 109

The Drainage Basin — Development, Morphometry,

Fluvial Processes 205

Fluvial Landforms 255

Glaciers and Glacial Mechanics 335

Geomorphology has undergone a drastic change in scope and philosophy during

matic and geologic controls. More recently, geomorphologists have recognized

braries of North American colleges and universities. There is less mathematical

I wish to acknowledge the help and guidance I received from numerous

surface forms. probably safe to assume that as humans became aware of

or to control the physical environment in the most advantageous manner. While

Landform data come from widely divergent disciplines. Synthesizing the

support interpretations. But the real test of geomorphic validity is outdoors,

Understanding what process is also serves as a basic component of other

edge about geomorphic processes is basic in the field of environmental science.

effects of cyclic climate change are beginning to produce accelerated erosion

1.2). However, simple landform identification is not enough. Understanding

the genesis of those features requires knowledge of how processes function in

Finally, an understanding of process is critical in itself. Cause

-/ 1. A delicate balance or equilibrium exists between landforms and

disequilibrium and a major response may

5. Geomorphic analyses can be made over a variety of time intervals.

The Delicate Balance

after World War II that the equilibrium approach was number

sophical framework for geomorphic analyses. Dynamic equilibrium suggests

or equilibrium condition within our system is revealed by statistical relation-

The systems approach has become highly sophisticated (Chorley 1962;

1. It emphasizes the intimate relationship between process and form.*

As stated above, equilibrium implies that landforms (and presumably

With the foregoing perspective of time, it is apparent why the concept of

geomorphic analyses. (A) Steady

(B) Graded time— steady state equilibnur

(C) Cyclic time — dynamic equilibrium

Table 1.1 Common systems of units used in mechanical analyses.

Systems Length Mass Force Time

cgs centimeter gram dyne second

where W is weight, m is mass, and g is the acceleration due to gravity. This

where F is force and 5 is distance of displacement. In the fps system of units

where PE is the gain in potential energy.

in mechanics can be thought of as the application of energy, a phenomenon

In process geomorphology landforms represent interaction between driving

Driving Forces Having suggested that energy is exerted on Earth materials

Short-Wave Long-Wave Poleward Transport

20-40 030 034 030

60-90 0.14 0.13 030

sunlight passes through the atmosphere, another segment of the radiation is

Some is transferred by conduction into

vection cell. the descending limb of the

30° Pr evailinQ westerlies

Trade winds \2=^

is effective, however, in stabilizing temperature conditions on a regional basis,

Thus, gravity attraction between two objects is an action-reaction phenom-