Process Geomorphology
Process Geomorphology
Process Geomorphology
Process Geomorphology
Frontispiece, G. K. Gilbert
standing by perched granite
boulder in Yosemite National Park,
1908
Process Geomorphology
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
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
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.
Preface xi
1
Process Geomorphology — An Introduction 1
Introduction 2
The Basics of Process Geomorphology 6
Summary 31
Suggested Readings 31
VII
1
Contents
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
Suggested Readings 333
10
Glacial Erosion, Deposition, and Landforms 363
Introduction 364
Erosional Processes and Features 364
Deposits and Depositionai Features 379
Summary 403
Suggested Readings 404
Contents
11
Periglacial Processes and Landforms 405
Introduction 406
Permafrost 407
415
Periglacial Processes
Landforms 421
Periglacial
Environmental and Engineering Considerations 434
Summary 442
Suggested Readings 443
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
Suggested Readings 480
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
Suggested Readings 529
Bibliography 531
Credits 569
Index 573
Preface
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-
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.
Process Geomorphology
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
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
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
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.
A
good example of human influence on geomorphic processes is occurring
today San Diego County, California (Kuhn and Shepard 1983). Here the
in
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.
•4
*dr.
w>
» * .
Figure 1.2.
Large landslides on Mars.
Compare with figure 1 1
Chapter 1 6
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.
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
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:
Figure 1.4.
Different time intervals and
associated equilibrium in
/\/\>
'"\y\, = Intervals of
A/\
graded time
10 6 10 7
Time (years)
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
Units 3
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 ,
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,
Work = PE
PE = Fs
PE = 100 lb X 30 ft = 3000 ft lb.
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
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.
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
0-20 34 39 030
57 X 10' 5 (20°)
From Handbook of Applied Hydrology by Ven T Chow Copyright <& 1964 by McGraw-Hill Inc Used with permission of
McGraw Hill Book Company
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.
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
13 Process Geomorphology —An Introduction
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-
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)
/ /
~Horse latitudes -
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
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
GM
g
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
evidence for seafloor spreading, that are virtually inexplicable without some
17 Process Geomorphology —An Introduction
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.
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
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
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
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.
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
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
Table 1.8 Rocks exposed at the surface of the North American continent
(expressed as % of area).
Sedimentary 61 5 52
Volcanic 82 11
Plutonic 3.8 6
Metamorphic and total PG 265 31
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.
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
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
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
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
27 Process Geomorphology —An Introduction
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
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.
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
29 Process Geomorphology —An Introduction
-, 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.
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
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.
31 Process Geomorphology —An Introduction
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.
2
I. Introduction B. Climate Change and
II. The Endogenic Effect Geomorphic Response
A. Epeirogeny 1. Sea Level Fluctuation
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
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
.
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
surfaces at elevations between sea level and 1 km account for 20.8 percent of
.
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, ypeiro^enic (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 such a model the value of F can be expressed in terms of energy, and its
dU Gm
-2
dr
37 Climate and Internal Forces
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,
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).
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
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
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
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
41 Climate and Internal Forces
Fennoscandia '
Recent ( + 1100
)
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
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.
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.
.
Chapter 2 42
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^f^ountairi-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
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
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
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
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).
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"
'
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 arej^engrajl 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.
47 Climate and Internal Forces
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
49 Climate and Internal Forces
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
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
I I
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
3
Climates Processes"
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.
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.
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
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
-120
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)
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
57 Climate and Internal Forces
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
(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
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
Pre-Wisconsin glaciation
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
;
CD
.1 150
T3
CD
CO
"co
100
c
CO
CD
D)
C3
50
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
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)
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.
/> 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-
small number of mineral varieties with an equally limited chemistry. With this
advantage, we can explore weathering efficiently even though the systems in-
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
A-H H+
H KOH
l_j-/
Exchange OH
OH O-H
OH-
OH
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.
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.
67 Chemical Weathering and Soils
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
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 '
*
m w *Ktj^
y
Bfl
t
V f K K
/
P. z' f •
%. Mp If
Chapter 3 68
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.
Figure 3.3.
Eh-pH characteristics of soils Eh
values in millivolts
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
(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
71 Chemical Weathering and Soils
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,
+
, , .
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 + .
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.
73 Chemical Weathering and Soils
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)
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^^IHHiBBI
Composition
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
75 Chemical Weathering and Soils
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
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
,
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
+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:
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
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
77 Chemical Weathering and Soils
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.
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.
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.
79 Chemical Weathering and Soils
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.
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
cz 3 11 Kaolinite
Dickite
Silica sheet Nacrite
Crystal unit
o ">
Alumina sheet Halloysite (4H 2 0)
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
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.
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 :
This suggests that other clay minerals in the weathered residuum are unstable
and will ultimately change to kaolinite.
Chapter 3 82
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
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
^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
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.
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
85 Chemical Weathering and 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).
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,
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
by combining this parameter with the bulk density (dry weight of soil/unit
87 Chemical Weathering and Soils
Figure 3.12.
Production and destruction of
organic matter in humid climates.
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
P* = w -w s d
X 100
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
Horizon 3 Characteristics
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
situ, reddish hues, iron and aluminum concentrations as sesquioxides, and stable
89 Chemical Weathering and Soils
Symbol 3 Meaning
g strong gleying
h illuvial humus
ir illuvial iron
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
In his most recent edition, Birkeland (1984) retains the K horizon to in-
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
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.
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.
91 Chemical Weathering and Soils
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
Table 3.7 Orders and suborders of the Soil Conservation Service classification
Udalf
INCEPTISOL Aquept
Ustalf
Andept
Umbrept ULTISOL Aquult
Ochrept Ochrult
Umbrult
ARIDISOL Orthid
Argid OXISOL
MOLLISOL Rendoll
At boll
Aquoll
Alton
Udoll HISTOSOL
Ustoll
Table 3.8 Formative elements in names of soil orders in the Soil Conservation
Service classification.
item of classification
93 Chemical Weathering and Soils
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
Soils highly
weathered 9. Oxisols
Organic
soils
(usually
have more 10. Histosols
than 25%
organic
matter)
It is obvious even from a brief description that the S.C.S. system allows
Forma- Derivation of
tive Formative Mnemon- Connotation of Formative
Element Element icon Element
Orders in S.C.S.
Classification Approximate Equivalents
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
S or s —f (cl, o, r, p, t . . .)
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
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,
mate, duration of soil development, and landscape evolution will make the
meaning of catena profile changes more difficult to interpret. Nonetheless, the
97 Chemical Weathering and Soils
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
.
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 .
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
Rainfall (cm)
Figure 3.17.
50 75
The depth of carbonate zones in
soils as related to mean annual
rainfall
20 30
Rainfall (inches)
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.
101 Chemical Weathering and Soils
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
Bca y 200 cm
or I
Zone of carbonate
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
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.
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-
changes, the dominant factors of soil formation in any given area must also
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-
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-''i^i-pOonaAna^-'\ floor -
—t^T^.^ Mountains
-1— ^?\, 5a" - : ;•
Valley Coalescent_ i^'
^Ssfer" Mountains S^;*-'"—
^,^i^5;Organ
rf er races
_ ^>^^$^C~~ ^-i^^-'t ^
Piedmont
slope
Bedrock
1200 meters. Between the mountains and the valley (the piedmont slope) exist
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
105 Chemical Weathering and Soils
-Torriorthents
Figure 3.21.
Terraces and river alluvium in the
valley border subarea near Las
Cruces, N.M
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-
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
III IV
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)
Holocene
0-100 (Dune) 3 (Dune) 3
100-7500 Fillmore 6 Organ I
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.
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
Suggested Readings The following references provide greater detail concerning the concepts dis-
cussed in this chapter.
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
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 occur^al 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
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
Na-Montmorillonite
Osage, Wyo 1400-1600
lllite
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
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,
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)
115 Physical Weathering, Mass Movement, and Slopes
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
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 .
117 Physical Weathering, Mass Movement, and Slopes
effect usually occurs after the parentrock haTalready been converted into soil.
Pla nt roots commonly grow in fragtnrps^o f the parent rock an d p hysically pry
THe solid material apart Nonetheless, co mpared with_other pjgcesse s,_rQojiet
.
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
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
119 Physical Weathering, Mass Movement, and Slopes
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-
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,
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.
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.
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
water table zero. Therefore, the effective normal stress and the total normal
stress are the same since
a = a 0.
123 Physical Weathering, Mass Movement, and Slopes
<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.
Chapter 4 124
Table 4.2 Average Atterberg limits in clays saturated with common cations
3
Montmorillonite lllite Kaolmite Halloysite"
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
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
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
1 2 3 4 5
r : 7 r : 6 r 5: r : 4 r : 1
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
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:
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.
Earthquakes
Regional tilting
Removal of underlying support
Natural (undercutting, solution, weathering, etc.)
Human activity (mining)
Lateral pressure
Natural (swelling, expansion by freezing, water addition)
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,
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
129 Physical Weathering, Mass Movement, and Slopes
_ 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
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
Bluff Peat
1 1
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
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).
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)
EXTREMELY
SLOW TO MODERATE
Rockslide
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
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
135 Physical Weathering, Mass Movement, and Slopes
Type of Material
Translational
Many Earth slide
Units Rock slide Debris slide
From Varnes 1978. T R B Special Report 176 Landslides Used with permission of the Transportation Research Board National Academy ot Sciences
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
137 Physical Weathering, Mass Movement, and Slopes
cohesion (c), effective normal stress (a ), and friction (0) such that
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
1 39 Physical Weathering, Mass Movement, and Slopes
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:
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
Index Description
Classification D/L — maximum depth of displaced mass prior to its displacement over
maximum length
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
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°.
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
(B) (C)
Table 4.7 Average values of the depth/length ratio in different types of landslips
as calculated for different areas.
Flows 1.58
Planar slides 6.33
Rotational slides 20.84
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
145 Physical Weathering, Mass Movement, and Slopes
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
Figure 4.27.
The relationship between mass
strength and profile angle for all
I I I I I I I I
10 20 30 40 50 60 70 80 90 100
Angle of slope (S°)
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
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
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.
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 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
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.
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
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.
Suggested Readings The following references provide greater detail concerning the concepts dis-
cussed in this chapter.
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
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.
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-
Figure 5.1.
Basic drainage patterns.
Descriptions are given in
table 5.1.
Dry-<?
Basic Significance
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.
Rectangular Joints and/or faults at right angles Lacks orderly repetitive quality of trellis
pattern; streams and divides lack regional continuity
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
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
157 The Drainage Basin —Development, Morphometry, and Hydrology
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
-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
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.
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.
..:
North
(A) (B)
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
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.
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.
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.
-
Linear Morphometry
- I lP ~ 1 1 where
wher
Total stream length L = L,R b s 1
I I
V u—1 Ju = RL
2D
Areal Morphometry
Basin shape RF = —
Lb
Drainage density D= —
1L
F = —
N
Stream frequency s
C = —
1
Constant of channel maintenance
D
Relief Morphometry
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.
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).
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.
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.
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,
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).
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.
171 The Drainage Basin —Development, Morphometry, and Hydrology
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
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-
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
Figure 5.15.
Growth of two parallel drainage
patterns in Ontonagon region,
the
northern Michigan. (After Hack
1965)
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)
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.
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
175 The Drainage Basin —Development, Morphometry, and Hydrology
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
Glacial Deposits
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
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
Capillary fringe
ft'' Water table Capillary water
x
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
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
V= K ^' ~ "2
'
Recharge
Figure 5.21.
Confined aquifer and piezometric
surface
Artesian Well
Piezometric
surface
Q = KIA
where
:re Q is discharge,
dis A is the cross-section area of an aquifer (w X d), and
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
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.
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.
181 The Drainage Basin —Development, Morphometry, and Hydrology
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.
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,
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
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
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
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.
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 •
•
• >•
1.1
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
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).
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.
1 89 The Drainage Basin —Development, Morphometry, and Hydrology
200
100
« 80 :
1 60-
40-
20-
(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
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.
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
191 The Drainage Basin —Development, Morphometry, and Hydrology
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
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
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
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
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.
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
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.
(1) precipitation and vegetation, (2) basin size, (3) elevation and relief. (4) rock
type, and (5) human activity.
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
E
a)
to
10
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
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
197 The Drainage Basin —Development, Morphometry, and Hydrology
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)
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-
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)
Figure 5.35.
Relation between infiltration
^W i i i I <W
1.0
Infiltration capacity (in/hr)
1 99 The Drainage Basin —Development, Morphometry, and Hydrology
2000
Figure 5.36.
"D „ Changes in sediment yield and
<u
03
~o 600
200
1800 1860 1900 1960
Construction
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 -
Gully
Channel Channel
banks bed
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).
201 The Drainage Basin — Development, Morphometry, and Hydrology
Table 5.5 Sediment budget for a small drainage in northern California for the years 1971-1974
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
Average
Annual
Suspended
Load Denudation
Basin Location Area (mi 2 ) (tons x 10 3 ) (in/1000 yr)
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.
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)
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,
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.
Suggested Readings The following references provide greater detail concerning the concepts dis-
cussed in this chapter.
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-
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.
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-
.
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
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
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
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
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
F = r
?r
where d is depth and g is gravity. The Froude number is important because it
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
f^ow? ^ \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
Table 6.1 Manning roughness coefficients (n) for different boundary types.
Manning
Roughness
Boundary n(ft 1/6 )
From Handbook of Applied Hydrology by Ven T Chow Copyright© 1964 by McGraw-Hill Inc Used with permission of
McGraw-Hill Book Company
(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
= 2 1o g7
f+ 1.0,
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
Accelerating flow
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
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
McGraw-Hill Book Company
211 Fluvial Processes
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
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-
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).
213 Fluvial Processes
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
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
power law provides a sound theoretical basis for competence studies, but it is
less satisfying in practice because accurate measurement of bed velocity is
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
215 Fluvial Processes
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
(/-/)
30 40 50 60 70 80
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.
217 Fluvial Processes
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
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
<" = -^ = 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-
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
219 Fluvial Processes
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*^_ ^^^i^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
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 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
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
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
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,
w = aQ b
d = cQf
v = kQ™,
223 Fluvial Processes
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
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-
many cases the original stream bottom, and its gradient represents the channel
225 Fluvial Processes
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
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-
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)
(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
227 Fluvial Processes
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
(D
E
T5
0)
CO
aa)
a 10
T3 CD
C
CD
Q.
(0
3 O
CO
1
nil ti I I I I
(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.
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
229 Fluvial Processes
North River
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)
- «
m"
Chapter 6 230
1000
i
800 =
40
i i
60
i i 1
100
I
areas
alluvial terrace
1 l 400 600
I i
1000
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.
231 Fluvial Processes
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-
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
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
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
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)
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
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).
235 Fluvial Processes
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.
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
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)
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)
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)
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-
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
(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.
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.
239 Fluvial Processes
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.
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
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.
—
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-
_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).
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).
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-
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
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.
245 Fluvial Processes
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
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
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
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-
>- 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
247 Fluvial Processes
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 > r„
\ ':
<
Layered
L.
•t, < < t„ Uniformly \ ". Tc To
resistant- moderately \_ '^
nonresistant resistant
Initial profile
Time 1
Time 2
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
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
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.
2h Qi =* — — S-d-
-p^
d~p~
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
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.
251 Fluvial Processes
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 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):
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
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,
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
cussed in this chapter.
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.
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
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
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.
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.
259 Fluvial Landforms
Figure 7.2.
Map showing complex distribution
of various types of deposits on a
portion of the Mississippi River
floodplam near Grand Tower, III
ir deposits Channel
Overbank deposits Ponds
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
261 Fluvial Landforms
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
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
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)
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
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.
263 Fluvial Landforms
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
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
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).
Average Thickness of
River Basin Flood Depositon (feet)
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
wm
200 300
Time (years)-
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
267 Fluvial Landforms
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.
(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.
269 Fluvial Landforms
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
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
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-
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)
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
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
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
273 Fluvial Landforms
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.
275 Fluvial Landforms
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.
Fan Morphology
Plan View Broad fanlike Broad fanlike Broad fanlike Broad fanlike to
symmetrical symmetrical symmetrical elongated
Depositional Processes
Return 1-50 yr Discrete events 0-few days Seasonally Seasonally constant to 3000-6000 yr Discrete
Interval constant discrete events
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
277 Fluvial Landforms
'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
Ste^fc
:
Gravel in abandoned washes.
:
Varnish cover.
Gravel in desert pavement areas.
Weathered and varnished.
Gravel in modern washes.
Unweathered and no varnish.
Undifferentiated gravel.
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)
Figure 7.20.
Relation of fan area to drainage-
basin area for anumber of fans in
0.01
1.0 0.1 10 100 1000
Drainage-basin area (A D ), in square miles
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
281 Fluvial Landforms
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
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
283 Fluvial Landforms
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}
ii i i I I I I I wmmi^^^mm
5 10
Scale (miles)
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
285 Fluvial Landforms
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.
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
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."
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
287 Fluvial Landforms
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
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
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
289 Fluvial Landforms
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
291 Fluvial Landforms
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)
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
293 Fluvial Landforms
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
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.
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
295 Fluvial Landforms
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; 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.
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
'channel-Shelf '
meander belts marine coastal basin)
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).
297 Fluvial Landforms
']/ 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.
(A) (B)
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.
Figure 7.30.
Emergence of new land in
(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.
301 Fluvial Landforms
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
Suggested Readings The following references provide greater detail concerning the concepts dis-
cussed in this chapter and also contain more extensive bibliographies on the
various topics.
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
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
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
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,
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,
Figure 8.2.
Percentage of soil composed of
particles less than 2 mm in
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
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-
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
V. = 5.75 V. log z/ fo
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-
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
= 363 cm/sec.
r
; = A 7" gD
311 Wind Processes and Landforms
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
where V., is the threshold value of drag velocity, p„ is the density of the air,
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
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
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
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-
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
.003
.002
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
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
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.
317 Wind Processes and Land forms
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
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
Table 8.2 Eolian bed forms and their geometry and possible origin
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
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.
319 Wind Processes and Land forms
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-
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
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.
- —
E ^
; r
«-
o
__ o
: -
° c
- r
_t o —
n
w o n
J.
«—
2§ O TJ
*~
.is
CO u. o £ 5 a
323 Wind Processes and Landforms
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.
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
Figure 8.15.
Stages of seif development from
barchan under influence of
secondary wind direction: P =
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
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
327 Wind Processes and Landforms
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
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
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.
200-300" of loess
120-200" of loess
60-120" of loess
I 36-60" of loess
| <36 of loess
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
331 Wind Processes and Landforms
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
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
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
The following references provide greater detail concerning the concepts dis- Suggested Readings
cussed in this chapter.
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
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
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,
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
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
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
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
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
-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 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
-^Accumulation —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
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.
345 Glaciers and Glacial Mechanics
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",
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
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-
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
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).
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-
Direction of
Figure 9.9.
glacier movement
The process of regelation as it
Figure 9.10.
^^ationzone- Hypothetical diagram showing
why the velocity of glacier flow
reaches a maximum at the
equilibrium line If no change in
Q- A + B + C
Q=A+B+C+D
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
351 Glaciers and Glacial Mechanics
Figure 9.12.
Velocity distribution in glacier
flowing a channel with a
in
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
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.
353 Glaciers and Glacial Mechanics
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)
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
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-
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,
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
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,
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
359 Glaciers and Glacial Mechanics
Secondary Features
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
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.
361 Glaciers and Glacial Mechanics
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
Suggested Readings The following references provide greater detail concerning the concepts dis-
cussed in this chapter.
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.
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
20 30 40 50
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
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-
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.
369 Glacial Erosion, Deposition, and Landforms
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)
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
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
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.
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
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,
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
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
^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
377 Glacial Erosion, Deposition, and Land forms
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.
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.
379 Glacial Erosion, Deposition, and Landforms
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)
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-
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
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
at. *•
*
381 Glacial Erosion, Deposition, and Landforms
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
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
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
383 Glacial Erosion, Deposition, and Landforms
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
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
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
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
385 Glacial Erosion, Deposition, and Landforms
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
of the classification shown in the map in figure 10.14. In this region the
is
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
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
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
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.
389 Glacial Erosion, Deposition, and Landforms
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
391 Glacial Erosion, Deposition, and Landforms
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.
393 Glacial Erosion, Deposition, and Landforms
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
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-
'W
N
.
-*.
395 Glacial Erosion, Deposition, and Landforms
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.
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.
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.
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.
397 Glacial Erosion, Deposition, and Landforms
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
(A)
(B)
399 Glacial Erosion, Deposition, and Landforms
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
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
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
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-
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;
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
Suggested Readings The following references provide greater detail concerning the concepts dis-
cussed in this chapter.
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-
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
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
Chapter 11 410
Base of permafrost
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
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
1973).
Table 11.1 Selected Northern Hemisphere permafrost thicknesses
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)
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)
(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)
(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)
Figure 11.6.
Extent of continuous and
discontinuous permafrost in the
Northern Hemisphere. (After
Fernans et al. 1969)
Continuous permafrost
/
^Discontinuous permafrost
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
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
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.
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
(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,
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
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
421 Periglacial Processes and Landforms
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.
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-
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
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
meters
500th Winter 500th Fall
Fro2Efe=
423 Periglacial Processes and Land forms
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).
425 Periglacial Processes and Landforms
&
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
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.
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
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.
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
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).
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.
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).
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).
leys or disseminated across a wide area. Although we may employ the term
"blockfield" in reference to these accumulations (as have many others), some
431 Periglacial Processes and Landforms
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
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
433 Periglacial Processes and Landforms
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.
\*
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-
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
435 Periglacial Processes and Landforms
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.
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
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
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
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
439 Periglacial Processes and Landforms
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.
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.
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
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
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.
443 Periglacial Processes and Landforms
The following references provide greater detail concerning the concepts dis- Suggested Readings
cussed in this chapter.
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.
Figure 12.1.
Major karst areas of the United
States (Boundaries are
generalized, and some regions
shown black contain
in solid some
nonkarst areas.)
Term Definition
Aggressive water Water having the ability to dissolve rocks. Especially water
containing dissolved C0 2 .
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.
P = j* X 100,
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.
451 Karst —Processes and Landforms
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.
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
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)
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)
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
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
through the system rapidly and with some turbulence will ultimately dissolve
more limestone, whereas stagnant water may become supersaturated with re-
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
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).
Figure 12.3.
The distribution of swallets, sinks,
and springs in a karst region of
southeast New York. (From Baker
1976)
457 Karst — Processes and Landforms
M-. _
Chapter 12 458
Figure 12.5.
Classification of flow types in
T
— Generally low relief surface
Random arrangement of
small cavities
Sinking
stream
Water table
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
459 Karst —Processes and Landforms
Springs After entry into the underground system, water moves as one or
more of the types introduced above. The manner in which groundwater leaves
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).
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'
^^_
Datum is mean
sea level vertical
Siltstone I Shaly limestone exaggeration x 10
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.
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.
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-
Closed Depressions
If someone were to ask what kind of landform best typifies a karst terrain, the
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"
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:
465 Karst —Processes and Landforms
Figure 12.12.
Doline near Dongola, Union
1. Slope. Since ponding or retardation of flow accelerates infiltration, County, III
Figure 12.13.
Sequential development of
collapse dollne
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
geological processes such as river entrenchment can cause this, the association
467 Karst —Processes and Landforms
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-
permeability
v
1
t III
Rapid flow down enlarged fractures
469 Karst — Processes and Landforms
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.
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
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
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
% 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-
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
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
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.
477 Karst — Processes and Landforms
Figure 12.21.
Subtypes ot the maze pattern of
cave development.
Anastomotic Spongework
maze maze
Figure 12.22.
Models of groundwater flow in the
development of caves
Swinnerton (1932)
(B)
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.
479 Karst —Processes and Landforms
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-
Suggested Readings The following references provide greater detail concerning the concepts dis-
cussed in this chapter.
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
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.
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
(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)
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.
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
487 Coastal Zones — Processes and Landforms
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.
v= Vg/T
and
L=T yfgh,
(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
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.
491 Coastal Zones —Processes and Landforms
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
Spring tide lations gradually decrease until the equilibrium level is once again attained.
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
493 Coastal Zones —Processes and Landforms
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
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.
495 Coastal Zones —Processes and Landforms
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
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,
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
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.
499 Coastal Zones —Processes and Landforms
MHW
MLW ^^ |
\
\
\
Berm
k Ridge J
Beach face
Longshore bar I
Runnel
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.
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
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.
501 Coastal Zones —Processes and Landforms
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,
(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
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.
505 Coastal Zones —Processes and Landforms
Figure 13.16.
Drift of sediment along a beach
face caused by waves striking at
an angle to the beach.
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
Table 13.3 Representative rates of littoral drift along coasts of the United States
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.
509 Coastal Zones — Processes and Landforms
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.
Topographic Step Berm, beach face Beach berm-offshore Coastal plains; shores Coastal plain
Association bar system with sufficient deltas
sediment
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
(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
CAPE HATTERAS
5
"~ \
Pamlico
Sound
'
rnlic
pa
River
CAPE LOOKOUT
ATLANTIC OCEAN
V
CAPE FEAR
\
15 15 3 45
CAPE ROMAIN
513 Coastal Zones — Processes and Landforms
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
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
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
Process Description
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
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
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
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)
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
H = —s- e~
c ,
c
. In areas of relatively homogeneous parent material and distinct
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 =
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
519 Coastal Zones — Processes and Landforms
Figure 13.23.
Schematic illustration of cliff
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).
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
.
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
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.
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.
523 Coastal Zones —Processes and Landforms
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
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
-
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.
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
527 Coastal Zones — Processes and Landforms
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)
529 Coastal Zones — Processes and Landforms
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.
The following references provide greater detail concerning the concepts dis- Suggested Readings
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.
.
531
—
Bibliography 532
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
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.
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.
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
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
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
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
.
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
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
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
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.
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
.
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
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
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.
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
.
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.
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
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.
.
. 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-
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
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
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
.
. 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:
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:
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.
. 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
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-
Chapter 3
Figure 3.2: D. F. Ritter.
1
Index
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
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
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
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
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
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