Eco Hydrology
Eco Hydrology
Eco Hydrology
The book begins with the small-scale physics of internal control of forest growth
represented by the vertical fluxes of light, CO2 , water vapor, and sensible heat within
the canopy. The scope then widens to consider the large-scale physics of the external
control offered by the balances of thermal energy and water. This leads to the
specification of the two state variables of a maximally productive canopy structure and
completes the hydrologic surface boundary condition for given climate, soil, and plant
species. Finally, by idealizing the stomatal responses to both light and available water,
and assuming a preferred average state of zero plant stress, necessary relations are
obtained among species, climate, and soil for maximum plant productivity. These new
ideas are used to define the climax bounds of plant habitat and to estimate net
primary productivity in the canopy.
Ecohydrology bridges the fields of hydrology and ecology and proposes new
unifying principles derived from the concept of natural selection. It also has potential
application in determining the response of vegetation to slow variations in climate.
This book will therefore provide fascinating reading for graduate-level students and
research scientists working in the fields of ecohydrology, hydroclimatology, forest
ecology, and surface water hydrology.
Peter S. Eagleson
Massachusetts Institute of Technology
The Pitt Building, Trumpington Street, Cambridge, United Kingdom
The Edinburgh Building, Cambridge CB2 2RU, UK
40 West 20th Street, New York, NY 10011-4211, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
Ruiz de Alarcn 13, 28014 Madrid, Spain
Dock House, The Waterfront, Cape Town 8001, South Africa
http://www.cambridge.org
A Introduction 1
B Overview 2
Part I Biophysics
2 Canopy structure 17
A Introduction 17
B Stand structure 17
C Crown shape 18
D Solid density 20
E Leaf orientation 23
F Leaf type 34
3 Radiant uxes 37
A Introduction 37
B Definitions 39
C Partition of radiance 44
D Nominal average insolation over the growing season 49
E Length of growing season 52
F Penetration of direct radiation into the canopy 56
G Penetration of diffuse radiation into the canopy 66
H Penetration of total radiation into the canopy 67
I Optical optimality 71
J Structural response to the radiation regime 76
vii
viii Contents
4 Turbulent uxes 80
A Background 80
B Introduction to momentum flux 82
C General formulation of momentum flux 85
D Features of momentum flux in homogeneous canopies 91
E Momentum flux in homogeneous leafy canopies: M = 1 102
F Momentum flux in homogeneous stemmy canopies 108
G Momentum flux in non-homogeneous leafy canopies: M = 1 113
H Drag partition 124
I Eddy viscosity for open canopies: M 1 127
J Flux of mass and of sensible heat 127
K Evaporation from a saturated surface 130
L The equivalent atmospheric resistance 133
M Transpiration from saturated stomatal surfaces 134
N Transpiration from a dry leaf surface 137
O Transpiration from a dry forest canopy 139
A Introduction 170
B Model framework 172
C Precipitation 174
D Storm infiltration and surface runoff 176
E Potential (unstressed) transpiration 180
F Bare soil evaporation 182
G Moisture-constrained transpiration 186
H Percolation and capillary rise 188
I Evaporation from surface retention 189
J The mean annual (or seasonal) water balance 190
K The vegetation statespace 191
L The critical moisture state 195
M Conclusions 206
Contents ix
A Photosynthesis 235
B Photosynthetic capacity of an isolated leaf 242
C Light control of productivity of an isolated leaf 245
D Water control of productivity of an isolated leaf 253
E Joint control of leaf productivity by light and water 262
F Bioclimatic optimality for the canopy-average leaf 266
A Summary 321
B Speculations 325
C Opportunities 327
Appendices
Glossary 399
References 409
xi
xii Foreword
a unique manner. The approach is analytical proceeding from the small-scale physics
of the internal controls of forest growth to the large-scale physics which crucially
control such growth. Professor Eagleson has been, throughout the last 40 years, a
fundamental contributor to many of the key ideas and developments underlying the
quantitative description and modeling of these physical processes. Nevertheless, this
book is much, much more than an excellent synthesis of the state of the art. It in fact
sets a new level for the science of ecohydrology.
The first part of the book presents a comprehensive and detailed description of the
radiant and turbulent fluxes within the canopy and between the canopy and the atmo-
spheric boundary layer. It continues with the study of the energy and water balances and
the intimate links which these balances establish between soil, vegetation, and climate.
The depth and focus of the presentation as well as the many novel contributions present
in the first six chapters would be enough to make this book an outstanding contribu-
tion to the understanding of ecosystems. Nevertheless, this part just sets the stage for a
tour de force of imagination and analytical prowess where the author develops optimal-
ity principles for the form and function of natural forests. Thus, the reader embarks on
an exciting exploration of the links between soil, climate, and species characteristics,
necessary to maximize plant productivity. Furthermore, Professor Eagleson proceeds
to derive the criteria for optimal canopy structure in which canopy conductance of
water vapor and carbon dioxide is maximum.
Chapters 7 through 11 present for the first time the new full theory. The driving
general assumption is that evolutionary pressure drives forests towards a bioclimatic
optimal state characterized by a maximum probability of reproductive success assumed
to correspond to maximum biomass productivity. This principle of bioclimatic opti-
mality is analytically structured around the following secondary principles: (1) optical
optimality : foliage arrangement and climate insolation leading to maximum car-
bon assimilation; (2) mechanical optimality : leaf angle at which canopy conduct-
ance of carbon dioxide and water vapor is maximized; (3) thermal optimality : leaf
temperature equals the photosynthetically optimum temperature; and (4) hydrologic
optimality : moisture state at which plant is at incipient stress and the average insola-
tion is a maximum, for open stomates. The author carries extensive comparisons of
the consequences of the above principles against historical data collected in forests of
different species and in different climates. These data cover a wide range of environ-
mental conditions and community types and provide significant support for Professor
Eaglesons theory, which I am sure will become a cornerstone in ecohydrology.
Figure 10.3 is remarkably suggestive of bioclimatic optimality where the vast ma-
jority of the observed species lie inside the range of global maximum productivity.
This book contains the signatures of how the best science is created: it is a fascinating
blend of knowledge and imagination which opens new and exciting avenues to fun-
damental problems in ecohydrology. Its profound originality, physical understanding,
and analytical elegance will have a lasting and major impact in the field. Moreover, the
importance of its ideas and the excitement transmitted by a truly creative mind at work
Foreword xiii
Ignacio Rodriguez-Iturbe
Theodora Shelton Pitney Professor of Environmental Sciences
and Professor of Civil and Environmental Engineering
Princeton University
Preface
This is a research monograph in which I explore and evaluate the biophysical relation-
ship between ambient climate and the form and function of the associated vegetation.
I hope the results will be useful in anticipating the changes to be expected in vegetated
surfaces under conditions of slowly changing climate.
At the 70 percent of Earths surface which is water, the atmospheric boundary con-
ditions are, under idealized circumstances, readily expressible in terms of the classical
laws of thermodynamics and fluid mechanics. For the remaining 30 percent of the
surface however, there is a greater or lesser presence of vegetation, and the boundary
conditions are determined in part by biological behaviors which differently constrain
the moisture and energy fluxes. In such cases the boundary conditions are termed
interactive because the structure and type of vegetation help to determine the cli-
mate through their exchange of heat and moisture with the atmosphere, while the
climate through its supply of moisture, carbon dioxide, and light to the surface helps
to determine the type and structure of the vegetation.
This book is devoted to theoretical generation of these atmospheric boundary con-
ditions at vegetated surfaces under conditions which are necessarily highly idealized.
As John Monteith (1981, p. 753) said: in a topic such as this, progress can be made
only if the number of variables is held to a minimum at every stage in the analysis . . . .
Here I consider only monocultures; I neglect the activities of insects and other animals
(including man), and omit the influence of disease and fire; I assume infinite nutrient
and carbon dioxide reservoirs leaving water and light as the limiting resources; and
I analyze the climatesoilvegetation system as though it operated in its long-term
average state with a neutrally stable atmosphere and no lateral advection of energy or
water from adjacent landscapes. Furthermore, I sever the feedback link from surface to
xv
xvi Preface
atmosphere and thus consider the vegetation as a passive responder to climatic forcing
at this level of approximation.
Ecologists, micrometeorologists, and plant scientists will likely find this approach
to be nave and may even be offended by my neglect of much important natural
complexity. Engineers, on the other hand, will be familiar with and may welcome
the reduction of an intricate multidisciplinary problem to a small set of simple, albeit
approximate, rules. In either case, I ask the reader to withhold judgement pending the
presentation of evidence.
Believing, after Darwin, that biology is an expression of physical optimality, my
hypothesis holds natural selection responsible for both the form and the function
of vegetation, and I seek its analytical formulation. However, selecting the modes
of expression of natural selection from such complexity involves (at best) informed
conjecture, and as G.K. Chesterton put it, There are no rules of architecture for a
castle in the clouds!
Denying altruism in these organisms, I assume natural selection to be expressed
through each individual tree rather than through the community of trees constitut-
ing the forest. In their classic paper, Parkhurst and Loucks (1972, p. 505) reasoned
Natural selection leads to organisms having a combination of form and function opti-
mal for growth and reproduction in the environments in which they live. Accordingly,
I imagine selection pressure to maximize reproductive success of the plant through
maximization of its biomass (and hence seed) production, and I formulate this produc-
tivity maximum in terms of the physical variables defining the climate, the soil, and
what are then called the optimal features of the vegetation.
This is not intended to be a micrometeorological textbook, an ecophysiological
textbook, or even a hydrological textbook. It is an effort to draw out, from among the
separate sciences of hydrology, micrometeorology, and ecophysiology, those behavio-
ral characteristics that together may contribute to natural selection, and to connect
them in a way that retains sufficient simplicity and scientific fidelity to be useful to
those charged with forecasting the components of climate change. Accordingly I have
tried to select models of the process physics and biology which are simple enough to
permit their combination without losing sight of the goal (and without unnecessarily
discouraging the reader!). I have tried to stick with the traditional models as best
fitting this bill. In multidisciplinary endeavors, all the rich scientific detail of each
contributing field cant be retained in their joining lest the resulting complexity negate
the utility of the result.
The book is logically divided into two parts:
Part I contains Biophysics. Following a summary in Chapter 1 of what lies ahead, it
presents the small-scale physics of those internal controls on forest growth, the vertical
fluxes within the canopy of light, carbon dioxide, water vapor, and sensible heat.
Chapter 2 reports the observed distribution and orientation of leaf area in the crown and
the canopy, and Chapter 3 examines the effect of solar altitude and foliage orientation on
the penetration of solar radiation into the canopy. Chapter 4 gives considerable detail of
Preface xvii
the vertical turbulent flux of horizontal momentum within canopies of various geometry
and, by analogy, the fluxes of sensible heat and mass as well. The canopy resistance
controlling these vertical fluxes is derived by building a seriesparallel resistive model
which bridges the scales from stomate to canopy. The scope then widens to consider
the large-scale physics of the external controls on forest growth offered by the thermal
energy balance (Chapter 5) and the water balance (Chapter 6). While Part I is mostly
familiar material several new ideas are developed.
Part II is best described as Darwinian ecology. With all details now in hand,
Chapter 7 develops an optimum canopy structure which specifies the two canopy state
variables (i.e., the canopy conductance and canopy cover) completing the hydrologic
surface boundary condition for given climate, soil, and plant species. Next, through
idealization of (1) the species- and temperature-sensitive photosynthetic capacity of
the leaf, and (2) the species-, soil- and climate-sensitive leaf water stress, Chapter 8
demonstrates the necessary relations among species, climate, and soil for maximization
of plant productivity under the additional assumption of a preferred zero-stress average
state. These ideas are then used in Chapter 9 to define the climax bounds of plant
habitat, and in Chapter 10 to estimate net primary productivity at the canopy scale.
The material of these four chapters is entirely new and represents the essence of the
book. Chapter 11 presents a brief summary along with some speculation and identifies
opportunities for useful extensions of this work.
The work is guided and evaluated along the way by comparisons of the optimal
vegetal features with observations from the literature whenever possible. These com-
parisons are favorable for some features and less favorable (if not unfavorable) for
others. The data are too few for useful statistical assessment, but I find the results
encouraging considering the enormous complexity of this ecological system.
The exposition has two threads:
1. Description of the observed physical characteristics of natural (and some
agricultural) vegetation canopies, and computation of the vertical fluxes therein
of momentum, mass, sensible heat, and light. While much of this is a collection
of material already present in the literature there is a revealing new
representation of effective canopy resistance that is helpful in understanding
possible processes of natural selection, and there is a new finding of optical
optimality that allows quantitative comparison of canopy carbon dioxide
supply and demand.
2. Search for the mechanisms by which natural selection is expressed in forest
form and function. This material is new and has not appeared before in print.
This breadth of scope gives the book utility both as a graduate level text for the
developing interdisciplinary field of ecohydrology, and as a reference for researchers
in landscape ecology. In view of the former, the writing style is tutorial with attention
given to completeness of detail.
My interest in the role of vegetation in climate arose from my water balance work
in 1975 and has been advanced with the inspiration, advice, and assistance of many.
xviii Preface
Peter S. Eagleson
Cambridge, MA
2002
Acknowledgments
Many have contributed to this work in important ways and deserve special mention at
the outset: Dara Entekhabi translated the original data of Appendix F into accessible
format; Andrea Rinaldo graciously edited an early draft of Chapter 4; Lelani Arris
donated an exhaustive catalogued library of early ecological references that eased the
literature search; David Benney assisted with mathematical complexities and gave
continuing moral support; William D. Rich provided the artful digital images of many
interesting trees from both the Arnold and Holden Arboreta; Beverly G. Eagleson and
Peter S. Eagleson, Jr. contributed additional photographs; Robert H. Webb supplied the
image of the creosote bush community and F. Eugene Hester that of the Okefenokee
swamp; and John MacFarlane prepared the line drawings. The author thanks them
each sincerely.
In addition, the author is indebted to: the MIT Department of Civil and Environmen-
tal Engineering for generous financial assistance with manuscript preparation through
resources of the Edmund K. Turner Professorship; the Rockefeller Foundation for pro-
vision of a Bellagio Center Residency during a crucial early stage of this long project;
and his colleagues at the Parsons Laboratory for keeping him eager for each new day
at the office.
Authors note
The photographs of individual trees presented here were chosen, for the most part, as
illustrations of their infinite variety and their intrinsic beauty rather than to illustrate
particular ideas put forth in the text. Note that species gathered from afar and open-
grown in an arboretum are likely to display different adaptive characteristics than they
would were they part of a homogeneous stand within their natural habitat.
xix
Notation
A = surface area, m2
A = exponential variable (Eq. D.1), arbitrary dimensions
Acm = surface area of conical monolayer (Eq. A.89), m2
A0 = reference value of exponential variable, same dimensions as A
AU = astronomical unit = 1.496 108 km
A D = horizontal component of the effective drag-producing area of the solid matter, m2
Ab = cross-sectional area of right circular cylinder circumscribing the crown, m2
Al = one-sided area of single leaf, cm2
AL = leaf area, m2
A0 = modified gravitational component of infiltration rate (Eq. 6.17), mm h1
As = horizontal shadow area per unit horizontal area of canopy (Eq. 3.18), m2
As = lateral surface area of cone, m2
Acm = surface area of conical monolayer, m2
Acr = horizontal cross-sectional area of the crown, m2
A1 = single-sided surface area of needles lying in and comprising the monolayer crown
surface, m2
A = water balance parameter for quadratic soil (Eq. C.22), dimensionless
a = scattering dimension, m
a( ) at ( )/L t = single-sided foliage area homogeneity function, dimensionless
acr = crown one-sided leaf area density = (one-sided leaf area)/(crown volume), m2 m3
aD = canopy drag area density, m2 (drag area) / m3 (crown)
ad = extinction parameter for horizontal foliage shear stress (Eq. 4.5), dimensionless
aL = canopy leaf area density = single-sided leaf area per unit of canopy volume, m1
am = extinction parameter for horizontal momentum flux (Eq. 4.6), dimensionless
at = canopy foliage area density = single-sided foliage area per unit of canopy volume
(Eqs. 4.14, 4.15, A.3), m1
xx
Notation xxi
o
ISL = relative saturation insolation, dimensionless
ISW = desiccation or water-critical insolation, i.e., the insolation at intersection
of the asymptotes of the leaf transpiration-insolation curve and thus the
maximum insolation for fully open stomates (light non-limiting) (Eq. 8.26),
cal cm2 min1 or MJ m2 h1 or W m2
ISW = canopy average (i.e., effective) leaf desiccation insolation (Eq. 8.32),
cal cm2 min1 or MJ m2 h1 or W m2
I I /I0 , dimensionless
o
I = insolation in wave length interval around (Eq. 3.6), cal cm2 min1
or MJ m2 h1 or W m2
i = storm intensity, mm h1
j = storm counting variable
K = turbulent transfer coefficient, cm2 s1
K (1) = effective saturated hydraulic conductivity of soil, cm s1
K c = carbon dioxide eddy diffusivity, cm2 s1
K h = thermal eddy diffusivity, cm2 s1
K m = kinematic eddy viscosity (eddy momentum diffusivity) (Eq. 4.37), cm2 s1
K v = water vapor eddy diffusivity, cm2 s1
K m (0) = kinematic eddy viscosity at canopy top, cm2 s1
K mo = K m ( )/K m (0), dimensionless
Kmo K m ( )/K m (0) = canopy-average eddy viscosity (Eq. 4.105), dimensionless
)d = canopy-average eddy viscosity for dense monolayer crown, dimensionless
(K o
PD = canopy average above-ground assimilation capacity (Eq. 10.10), gs h1 / unit of
one-sided leaf area, g(CO2 ) m2 h1
PL = actual biological productivity, g(CO2 ) m2 h1
Pr = rate of respiration (Eq. H.2), g(CO2 ) m2 h1
Ps = light-saturated rate of productivity (equal to Psm when Tl = Tm ) (Eq. E.9),
gs m2 y1
PT = annual total above-ground carbon demand (Eq. 10.18), g(CO2 ) m2 y1
Pt = total rate of photosynthesis (Eqs. 8.2, H.1), g(CO2 ) m2 h1
PW = climatic potential productivity as limited by available water, g(CO2 ) m2 h1
Psm = saturation rate of CO2 assimilation per unit of (one-sided) leaf area when Tl = Tm
(Eq. 8.1), g(CO2 ) m2 h1 or gs m2 h1 , as noted
Pwm = desiccation rate of CO2 assimilation as limited only by available water,
g(CO2 ) m2 h1
P = P/Psm = normalized net photosynthesis (Eqs. 8.12, 8.15), dimensionless
o
Soo = solar irradiance on horizontal plane at Earths surface (Eq. 3.2), W m2
s = instantaneous amount of some fluid quantity (e.g., momentum, heat, or solute
mass), quantity g1
s = soil moisture concentration (s = 1 at saturation), dimensionless
sc = critical soil moisture concentration, dimensionless
so = spacetime average soil moisture concentration in the root zone, dimensionless
soc = critical spacetime average root zone soil moisture concentration (Eq. 6.58),
dimensionless
s5 = soil moisture concentration at s = sc = 5 bar, dimensionless
s = soil moisture for quadratic soil (Eq. C.21), dimensionless
s(0) = soil moisture concentration (its maximum value) at end of average storm,
dimensionless
s( ) = interference (or shielding) function (Eq. 4.11), dimensionless
s = temporal mean of fluid quantity, s
s = turbulent temporal fluctuation of fluid quantity, s
T = temperature, C
Tl = leaf temperature, C
Tl = crown-average leaf temperature, C
Tm = optimal leaf temperature = temperature at which Psm is maximum for a given
species, C
T0 = climatic atmospheric temperature at screen height above canopy, C
T0K = climatic atmospheric temperature at screen height above canopy, K
Ts = surface temperature, C
TsK = surface temperature, K
t = time, s
t = thickness of foliage layer, mm
tb = time between storms, h
tb = interstorm time available for transpiration, h
tb = interstorm time available for bare soil evaporation, h
te = time between exhaustion of surface retention and cessastion of soil moisture
evaporation, h
tj = Julian day, dimensionless
to = time after exhaustion of surface retention (t = t ) at which f e = E ps , h
tr = storm duration, h
ts = time to stress (Eq. C.6), h
ts = time during which the vegetation transpires (Eq. 6.43), h
t = time at which evaporation exhausts surface retention, h
u = horizontal wind velocity, m s1
u = integration variable (Eq. 10.11), dimensionless
u 0 = horizontal wind velocity at canopy top, m s1
u = average wind velocity, m s1
u = shear velocity 0 / (Eq. 4.27), m s1
Ve = volume of soil moisture (per unit of surface area) available for exchange with
atmosphere during average interstorm period (Eqs. 6.41, 6.59, 6.67, 6.68,
8.27), mm
xxx Notation
= slope, e/T , of saturation vapor pressure vs. temperature curve (Eq. 4.134),
Pa K1
( ) = finite increment of ( )
S = average carryover (from dormant season to growing season) soil moisture
storage (Eq. 6.69), mm
z = lumped vertical intercrown diffusion distance for canopy of conical
crowns, m
z L = thickness of interleaf atmospheric layer, m
= solar declination, degrees
= 1/m tr , h1
f dt L t = monolayer foliage volume density (Eq. 7.22), dimensionless
= Psm /ISL = dry matterradiation quotient (Eq. 8.9), an empirical constant
with the approximate value 0.81 gs MJ(tot.)1
= infinitesimal radius, mm
= surface absolute roughness, mm
= characteristic of hemispherical segment, dimensionless
o = longwave emissivity of the surface, dimensionless
(z) = residual of exponential integral
( ) = surface runoff function (Eq. 6.27, Fig. 6.5), dimensionless
o stomated (i.e., assimilating) leaf area
illuminated (i.e., projected) leaf area
perimeter
arc chord (Eq. 8.8), dimensionless
= normalized evapotranspiration (Eq. 6.57), dimensionless
= angle between the ray, r, and an upward normal, n, to an arbitrarily oriented
surface
L = leaf angle (i.e., the angle between the leaf surface and the horizontal), degrees
o = particular leaf inclination, degrees
os = supplement of a particular azimuthal angle, degrees
z = Sun zenith angle = complement of h (cf. Fig. 3.2b), degrees
g(solid matter)
= empirical conversion factor g(CO 0.5 (Eqs. 8.7, 10.3)
2 assimilated)
= insolation extinction coefficient, dimensionless
b(1) = extinction coefficient of an isolated leaf with no scattering = one-dimensional
shadow of a solitary leaf (Eq. 3.20), dimensionless
b(2) = extinction coefficient for a single layer of leaves with idealized scattering,
dimensionless
b = decay or extinction coefficient for direct radiation, dimensionless
d = decay or extinction coefficient for diffuse radiation, dimensionless
o = shape parameter or distribution index of storm depth, dimensionless
= scatterless extinction coefficient for total clear sky radiation, dimensionless
= wavelength of radiation, nm
= roughness density or frontal area index (Eq. 4.106), dimensionless
= latent heat of vaporization of water = 597.3 cal g1 (at 0 C)
0 = scale parameter of probability density function of storm depth, cm1
NIR = 700 to 4000 nm
PAR = 400 to 700 nm
UV = less than 400 nm
xxxii Notation
irradiation
(I.e., quantity of energy arriving at a surface during a given time interval)
(From Iqbal, M., An Introduction to Solar Radiation, Table B.1, p. 378, Copyright
c 1983
Academic Press, with kind permission from Academic Press)
irradiance
(I.e., rate of energy arriving at and normal to a surface)
(After Iqbal, 1983, Table B.2)
miscellaneous
1 Angstrom = 1010 meters
1 calorie = 4.187 Joule
xxxiv
Units xxxv
Larcher (1983, p. XVII).
Hicks and Chabot (1985, p. 262).
Gates (1980, p. 84).
Gates (1980, p. 85).
Common and scientic names
xxxvi
Common and scientic names xxxvii
A Introduction
The biophysical system
The physical interaction between a vegetation canopy and its atmospheric and soil
environments is governed by both the plant structure and the biochemistry of the
individual plants. The spacing of the individual plants; their height and diameter; the
depth and shape of their crowns and root systems; the size, shape, number, color,
texture and spatial arrangement of their leaves along with the associated pods, stems,
twigs or shoots and branches, all contribute to the instantaneous vertical exchanges of
momentum, mass and energy between canopy and atmosphere and/or to the extraction
of moisture and nutrients from the soil. Plant biology modulates these fluxes through
such transient mechanisms as stress-induced variability of leaf stomatal resistance to
transpiration and to CO2 assimilation, short-term changes in leaf attitude, and seasonal
changes in the color and density of the foliage.
1
2 Introduction and overview
The question
Can we formulate this complex interaction in a way that is at once simple enough and
yet sufficiently exact to reveal the principal natural selection pressures that determine
the observed configurations and functionings of natural plant communities?
Background
The early work of Darwin (1859) on natural selection led to the concept of ecological
optimality. This connection between natural selection and the principles of physics was
recognized by Lotka (1922) who proposed that natural selection tends to make the en-
ergy flux through the system a maximum, so far as is compatible with the constraints to
which the system is subject . . ., and further that in the struggle for existence, the ad-
vantage must go to those organisms whose energy-capturing devices are most efficient
in directing available energy into channels favorable to the preservation of the species.
Rosen (1967, p. 7) proposed a more general connection between natural selection
and the environment. In his words On the basis of natural selection, then, it may be
expected that biological organisms, placed for a sufficiently long time within a specific
set of environmental circumstances, will tend to assume characteristics which are opti-
mal with respect to these circumstances. Parkhurst and Loucks (1972, p. 505) refined
this to the form commonly used today. That is Natural selection leads to organisms
having a combination of form and function optimal for growth and reproduction in the
environments in which they live.
Methodology
We seek to define, in terms of the key structural and behavioral parameters of a mono-
cultural plant community, those conditions under which the reproductive potential of
the individual plants is maximum for a given climate and soil. We assume the resulting
community will prevail in the given environment. While our interest is in natural sys-
tems, primarily forests, we will also examine the behavior of crops to the extent that this
may aid in generalization. We formulate our optimization arguments mechanistically
in the belief that quantification is the key to understanding.
B Overview
Objective
We wish to find a set of general biophysical relations that define the optimum natural
habitat for a given vegetation species and that alternatively will define the maximally
productive stable (i.e., climax) vegetation community for a given climate and soil.
Overview 3
the net primary productivity, NPP. It is our thesis that the physical parameter values
leading to maximum productivity in a given environment are those that we should find
in a climax community and that the analytical expression of this optimum equilibrium
may provide a useful means of coupling the vegetated land surface to the atmosphere
in a slowly changing climate.
For pedagogical elegance the work is organized, beginning with Chapter 2, as the
classical physical science textbook would be; understanding is developed analytically
and in a reductionist manner from first principles, building gradually toward NPP
in Chapter 10. However in this overview, in order not to lose sight of the forest for the
trees, we omit mathematical detail and begin with the discussion of productivity.
Factors of productivity
We define the environment of a forest in terms of: the average length, m , of its
growing season for particular species; the climatic time averages over this season of the
insolation, I0 , the precipitation, P , and the daylight-hour atmospheric temperature,
T0 ; under the assumption that the reservoir of nitrogen and other nutrients in the soil,
and that of CO2 in the free atmosphere, remain non-limiting to production. With these
simplifications we recognize the principal productive needs for: light to keep the leaf
stomates open for uptake of CO2 , and to fuel its assimilation; nitrogen to nourish the
formation of plant tissue; water to keep leaf stomates open by maintaining plant turgor,
to transport the nitrogen from soil to plant, and to regulate plant temperature through
evaporative cooling; and finally, the crowns turbulent flux capacity to evacuate water
vapor at a rate meeting the plants demand for water-borne nitrogen, and to supply
atmospheric CO2 at a rate meeting that demand for even the lowest leaf. We find
biophysical bases for optimizing each of these.
We begin by assuming for the moment that the capacity of the turbulent flux to
supply CO2 to the leaves exceeds the light-driven demand of the leaves for CO2 . Such
demand-limited assimilation is illustrated by the classical experimentally-determined
photosynthetic capacity curve for an isolated C3 leaf at constant leaf temperature,
Tl . This curve defines the rate of CO2 assimilation, P, as a saturating function of
the variable insolation, I, maximizing at the value Ps for large I. As sketched in
Fig. 1.1, for the temperature Tl = Tm producing maximum assimilation, the net photo-
synthesis is defined by two species-dependent parameters, the saturation rate of maxi-
mum carbon assimilation per unit of basal leaf area, Psm , and the saturation insolation,
The proportionality constant is likely to vary with species making uncertain those conclusions
about interspecies competition that are based solely on NPP.
This neglects the fact that competitive advantage in a mixed community may shift with age
leading possibly to a dominant stable species that is not climax in the globally optimum sense
used here for the monoculture.
Time-averaged quantities remain undifferentiated in notation because we write all our dynamic
relationships in terms of climatic time-averages.
Insolation is the flux density of solar radiation on a horizontal surface.
This crude assumption is an expedient that allows us to proceed and must be remembered when
evaluating our results.
Overview 5
pe stry
ng
ao i
at hem Photosynthetic capacity curve
ni
om c
St hoto
P
I SL I SW
0
0
Insolation, I
Fig. 1.1. Definition of the leaf light characteristic. Unit basal leaf area of C3 species;
water and nutrients non-limiting; optimal temperature; ambient CO2 .
ISL , at which the asymptotes of the photosynthetic capacity curve intersect. We call
these asymptotes the leaf light characteristic, and show that provided water is not
limiting, ISL marks the insolation at which the stomates are effectively fully open.
For maximum productivity and maximum efficiency of light utilization, the optimal
bioclimatic state for a single leaf of the given species sets the climatic value, I0 , of the
insolation equal to ISL , with Tl = Tm , as is shown in Fig. 1.2. We also show for isolated
leaves that ISL = Psm / with constant over a wide range of species as is shown in
Fig. 1.3. The relationship Psm = ISL is therefore the biochemical assimilation capacity
for C3 leaves.
Consider now an increasing climatic insolation, I0 , which causes decreasing av-
erage soil moisture concentration, so , when the other climatic variables remain fixed.
With I0 exceeding ISL (and hence non-optimal), but not yet causing water-limitation
due to a generous value of the fixed precipitation, the stomates remain fully open and
the individual leaf continues to transpire at its climatically potential rate, E v = E ps .
However, at the particular climatic insolation, I0 = ISW (which equals or exceeds ISL ),
the declining average soil moisture concentration reaches a critical value, soc , at
the end of the average interstorm period, causing the stomates to begin closing and
transpiration to decline. This situation is illustrated in Fig. 1.4 where we define the
evaporation function by its asymptotes which in this case we call the leaf water char-
acteristic and which intersect at the desiccation or water-critical insolation, ISW . We
show how ISW is estimated from the water balance (cf. Chapter 8). We also show that
at constant temperature, the transpiration rate is an approximate surrogate for produc-
tivity of the same species (cf. Appendix E) allowing us to use the productivity, P,
rather than the transpiration rate as the ordinate in Fig. 1.4, and making Pwm = Psm .
The range ISL I0 ISW is thus bioclimatically optimal for the isolated leaf because
it maximizes leaf productivity (cf. Fig. 1.7).
E ps is the potential rate of evaporation from a wet, simple surface (cf. Appendix B).
6 Introduction and overview
I
(a)
Optimum
Species 2
Psm2
1 I0 = ISL2
ISW
I
(b)
1.6
ity C3 crops
pac
ca
Psm (gs m2 h1)
n
tio
1.2 ila
s sim gs
la < >= 0.81
m ica MJ(tot)
he 1
oc
0.8 Bi
Sitka Deciduous trees
spruce
Coniferous trees
Goethalsia
0.4 Beech Loblolly pine
Arctic Red oak
willow
White oak
0
0 0.4 0.8 1.2 1.6 2.0 2.4
ISL (MJ(tot) m2 h1)
Fig. 1.3. Biochemical assimilation capacity of C3 leaves. Data from Table 8.2.
Overview 7
Pwm = Psm
Stomata
Stomata fully open
P closing
Stomata
closed
ISL ISW
0
Fig. 1.4. Definition of the leaf water characteristic. Unit basal leaf area; light and
nutrients non-limiting; ambient CO2 ; fixed species, soil, precipitation and temperature.
To proceed toward canopy production, we must expand Psm from isolated leaf to
full canopy. Expanding in the vertical direction we multiply first by the leaf area
index, L t , to incorporate the total basal leaf area per unit of crown basal area, and
next by the so-called carbon demand function, fD ( L t ) (cf. Chapter 10) to average
the leaf assimilation rate induced by a decaying insolation over the crown depth. This
function is based upon our following proposed conditions for optical optimality in
multilayers:
The reflection coefficient of the photosynthetically active component, PAR, of the
incident radiation is small and nearly invariant over the range of vegetated latitudes,
thus the incident light is used optimally for photosynthesis when in the time average:
(1) the upper leaf surfaces are in full sunlight and the crown basal area is in full
shadow at the average solar altitude, h ; by including scattering we show
(cf. Chapter 3) that this mandates an important geometrical relation between
the solar altitude and the leaf angle, resulting in = , as is sketched for a
pair of opaque leaves in Fig. 1.5 and as is verified from observations of full
crowns of translucent leaves in Fig. 1.6; and
(2) the insolation at the lowest leaf is the minimum for leaf metabolism (i.e., the
so-called compensation intensity, Ik ); we show that its relative value, Ik /I0 , is
a species constant (cf. Appendix H) that fixes the value of the insolation
L t is called the (horizontal) momentum absorption index, or the horizontal leaf area index, since
is the cosine of the leaf angle, L , with the horizontal.
is the light extinction coefficient.
8 Introduction and overview
In this vertical expansion from isolated leaf to whole crown, we average the
insolation over the depth of the crown at optical optimality to get the crown-average
L
h
b
Fu
ll i
1
llu
b
mi
nat
Leaf Leaf
c
ion
a
h L
(1)
b
0.6
Evergreen trees
0.4
Deciduous trees
0.2
0
0 0.2 0.4 0.6 0.8 1.0
Light extinction coefficient,
Overview 9
1 fI (Lt ) I0
L =
ISL
0 1 2 3
Crown
ic
P
W
ist
ate
er
Po =
ct
rc
Species
ra
ha
ha
ra
tc
cte
limited
gh
limited
ris
Li
tic
I0 = ISL I0 = ISW T0 = Tm
1
0
0 1 2 3
I0
Io =
ISL
Fig. 1.7. Optimal bioclimatic state and the supporting evidence. Data from Table 9.1.
Eight communities from desert shrub to rainforest. Plotted circle shows mean, box
shows one standard deviation, and bar shows range (fixed C3 species, soil and
precipitation; ambient CO2 ; nutrients non-limiting).
insolation, Il = f I ( L t )I0 . We assume all leaves in the crown to have the same light
characteristic making the crown-average I SL equal the isolated leaf ISL , whereupon
the optimal bioclimatic state for the canopy becomes ISL = Il = f I ( L t )I0 . This is
supported in Fig. 1.7 by observations over a range of communities from desert shrub
to rainforest (cf. Chapter 9). Together with Fig. 1.3, it defines the climatic assimilation
potential, Psm = f I ( L t )I0 , fixing the maximum CO2 assimilation rate for given
climate and species. The uppermost leaves in the canopy will reach their desiccation
moisture state, ISW , first and at the lower canopy-average radiance, ISW = f I ( L t )ISW .
Comparison of the relative magnitudes of Il , ISL , and
ISW , in all their permutations, de-
fines the range of natural habitats in water and light space and reveals the boundaries of
this space as climax conditions which compare favorably in Fig. 1.8 with observations
from the same wide range of communities (cf. Chapters 8 and 9).
Expanding crown productivity in the horizontal direction we multiply by the frac-
tion of the ground surface covered by crown basal areas (i.e., the canopy cover, M).
Being a canopy property, M depends upon the collective action of the individual plants
in satisfying their individual water needs and is not maximized by Dawkinss selfish
gene. We show that the canopy demands water over the full growing season at the time-
averaged rate MEv , where E v kv E ps is the canopy transpiration rate (cf. Chapter 6),
10 Introduction and overview
1.2
Water-limited
climax 4
7
8 2
1.0
5 1
9
6
0.8
0.6 Subclimax
fI ( L t )ISW
ISL
natural habitat
W =
3
0.4
Light-limited
climax
0.2
0
0 0.2 0.4 0.6 0.8 1.0
M
Fig. 1.8. Climax conditions as bounds to natural habitat. Data from Table 9.2.
1, Beech; 2, oak; 3, Goethalsia (rainforest); 4, Sitka spruce; 5, ponderosa pine;
6, loblolly pine; 7, pinyon-alligator juniper; 8, pinyon-Utah juniper; 9, creosote bush.
and kv is the canopy conductance (cf. Chapter 6). In order to maintain the water
balance, M will be less than 1 when E v exceeds the climatically available water sup-
ply, and such a vegetation community is termed water-limited. Otherwise M = 1
and the community is light-limited with an excess of water. We recognize that the
proportionality between soil-to-plant nitrogen flux and the flux of water, when cou-
pled with the sensitivity of CO2 assimilation to nitrogen availability, provides a degree
of productivity-motivated selection pressure to maximize the plant transpiration rate,
E v . This may be accomplished by maximizing kv , through minimizing the normalized
canopy resistance, rc /ra (at given temperature) as is shown in Fig. 1.9 (cf. Chapter 7).
For given we find this minimization to favor tapered crowns when L t is large. For
We show that the open-stomate condition of the optimal bioclimatic state allows approximation of
the canopy flux resistance, rc /ra , as being independent of stomatal resistance. We then use the
canopy-average eddy viscosity to derive rc /ra in terms of and L t .
Overview 11
0.8 = 0.5
rc
ns
ra
0.6
= 0.5
= 0.3
0.4
0.2
0
0 2 4 6 8 10
Lt
Oak
Beech
Sugar maple
Lt 6
Observed
range
Observed
mean
rc
2
Maximum vapor flux minimum ra ) )
at constant temperature with M = 1
0
0 0.2 0.4 0.6 0.8 1.0
or
40
36
1
American beech
Loblolly pine
1
32
28
Loblolly pine
T0 (C)
24
European beech
Arctic pine
20
16
Sitka spruce
12
12 16 20 24 28 32 36 40
Tm (C)
Bowen ratios on the order of 101 for a wide range of communities (cf. Table 9.2),
and therefore that nominally g(T l ) g(T0 ) = g(Tm ) = 1.
Finally, in order to get the potential (i.e., carbon demand-limited) annual production
for a given species, we multiply by the average length of the growing season, m ,
and to convert the gross primary productivity calculated in g(CO2 ) into net primary
productivity, NPP, measured in grams of (above-ground) solid matter, gs , we multiply
by the widely accepted nominal empirical conversion factor, = 0.50 gs g1 .
2.0
CO
= 0.30 2 6 8
downward suppl (a)
y 4
p=p
s 3 7
(b)
p =pD 5
9
NPP ,
p
1.0 and 12
dem 2
CO 2
1 10 11
0
0 1.0 2.0 3.0 4.0
Species, Lt
Fig. 1.12. Productive gain of C3 multilayer canopies. Curve (a) = canopy CO2 demand
compared to monolayer demand; curve (b) = canopy downward CO2 supply compared
to monolayer supply. Data points: 1, Creosote bush; 2, ponderosa pine; 3, loblolly pine;
4, oak; 5, rainforest (Ghana); 6, rainforest (Congo); 7, red spruce; 8, beech
(Tennessee); 9, beech (Eastern Europe); 10, pinyon-alligator juniper; 11, pinyon-Utah
juniper; 12, sugar maple. Symbol shape indicates crown shape.
14 Introduction and overview
NPP/ pS = fS (,L t ) for cylindrical crowns shown as Curve (b) in Fig. 1.12 for the
commonly observed value = 0.30. Here pS is the carbon supply analog of pD
which we assume to equal pD for the L t = 1 monolayer. With this critical assumption,
comparison of the supply and demand functions shows a critical absorption index,
L t , separating the region ( L t L t ) in which the atmosphere can supply the
entire CO2 demand from that ( L t > L t ) in which a growing fraction of the CO2
demand must be met from below by decaying plant matter. We show that for tapered
crowns (i.e., cones or hemispheres) the atmospheric CO2 supply is monotonically in-
creasing with L t so that these crown shapes can meet the CO2 demand at all L t
without local recycling. Limited observations seem to support our derivation of this
optimum productivity as shown by the plotted points on Fig. 1.12.
We find = 0.81 gS MJ(tot)1 .
Without the need here to have the denominator on the left-hand side of Eq. 1.1 represent
monolayer productivity we are free to extract from Psm its dependence upon species through
f I ( L t ) and to move this function to the right-hand side in Eq. 1.2.
Overview 15
the species-dependence of the potential (i.e., light-limited) NPP is now all on the
right-hand side. Equation 1.2 is plotted as the solid line in Fig. 1.13 and shows a broad
global maximum over a particular range of L t . The range of observed tree species is
indicated by the spread of the vertical dashed lines at the respective species-constant
L t . Their clustering in the range of the global maximum NPP is taken as confirmation
of our fundamental assumption.
Extreme climates call for extreme L t in order to conserve heat (large L t ) or
water (small L t ) leading to locally optimal NPP that are smaller (due to CO2 or water
limitation respectively) than this global, light-limited maximum. While the above
development of productivity and its underlying optical and bioclimatic optimalities is
at the heart of this work, we also present and evaluate applications of these ideas to
natural habitats and to ecotone location, and we discuss their potential use to assess
some of the apparent consequences of global climate change.
Part I
Biophysics
2
Canopy structure
Basic physical features of the forest canopy are defined, observations are pre-
sented, and simplified models are proposed.
A Introduction
The diffusion of water vapor, CO2 , momentum, heat and light vertically through the
canopy governs forest growth under the simplifying assumptions employed herein (see
Chapter 1). In turn, these diffusions are ultimately controlled by the magnitude, distri-
bution, and orientation of the solid matter comprising the forest canopy. This chapter
is devoted to generalized quantitative description of these geometrical characteristics.
B Stand structure
We are interested in the collective physical interaction between the environment
(atmosphere and soil), and a homogeneous community of plants called a stand. The
stand is composed of individual plants, each consisting of a crown supported above the
substrate (i.e., land surface) by a trunk or main stem. The collection of crowns in the
stand constitutes the plant canopy which carries that name because it is often elevated,
umbrella-like, above the ground. Each crown is composed of crown elements which
include the branches, and the foliage elements such as leaves, pods, and stems.
Assuming the individual plants to be identical and with circular plan form, the plan
view of the stand (see Fig. 2.1) shows circular disks of foliage nested at a greater
or lesser density depending upon the availability of some growth-limiting resource.
Letting that limiting resource be water and defining Rosens (1967) cost functional
as water-induced plant stress, Eagleson (1978f ) assumed the optimal density of this
nesting to be that which, for the given climate and soil, produced the maximum average
soil moisture concentration, because this condition minimizes plant stress. He defined
As we will see in Chapter 6, Salvucci (1992) showed that a somewhat greater water use is
achieved at a slightly greater vegetated fraction in return for somewhat higher risk of stress.
17
18 Canopy structure
(b)
the fraction of surface area covered by the projected crown areas (i.e., the disks) as the
canopy cover, M, and showed its variability in water-limited situations. We use this
definition here. When the disks are as close as possible (i.e., M = Mmax ) we refer to
the canopy as closed as in Fig. 2.1a, and when the disks are less densely packed as in
Fig. 2.1b we refer to the canopy as open. From idealized geometrical considerations
Mmax = 0.84, however in discussion of this extreme state, we will refer to it as M = 1.
In nature, the trees may depart from the circular planform assumed here, allowing a
packing density approaching Mmax = 1. The spaces between disks represent area that
may be occupied by an understory and/or by bare soil. For simplicity here we consider
it to be bare soil.
The individual crowns are described by their shape, solid density, leaf orientation,
and leaf type. Each of these physical features of the crown influences its absorption
of radiation, as well as the fluxes of water vapor, carbon dioxide, and heat to or from
the leaves. Therefore, through selection and/or adaptation, we expect the density,
arrangement, types, and orientation of the leaves to reflect the incident radiation regime
(among other factors) in some way (Russell et al., 1989a, p. 32).
C Crown shape
Trees have long been classified in part according to the shape of their crowns (see
Baker, 1950, pp. 71 et seq.). Sinnot (1960) observed that plant species are often sym-
metrical in their external shape with respect to a central vertical axis which allows
modeling individual plants as simple geometrical figures filled with plant elements.
We follow this precedent limiting our geometrical classes to cylindrical (i.e., flat top),
hemispherical (i.e., rounded), and conical (i.e., pointed). These are illustrated in
Fig. 2.2 along with the defining measures of median size: b = diameter of crown,
h = canopy height (i.e., height of tree), h c = depth of crown, and h s = stem height
(i.e., height of crown bottom). Note that we show the geometry common for a single
tree within an extensive closed canopy (i.e., M = 1) of identical trees and that the
crown does not extend all the way to the ground. As the canopy opens (i.e., M < 1), the
penetration of light increases and foliage may be found closer and closer to the ground.
Eagleson (1978a) called M the canopy density which is ambiguous in three-dimensional studies
such as this, so we adopt here the two-dimensional term canopy cover.
Crown shape 19
b b b
Canopy
b/2
hc
h Crown
z
hs
D Solid density
The diffusion of water vapor, CO2 , momentum, heat, and light through the canopy is
ultimately controlled by the magnitude, distribution, and orientation of the solid matter
comprising the canopy. The primary constituent of this solid matter is the leaves or
needles.
Crown and canopy internal properties are assumed to be spatially variable in only
the vertical direction. To study such distributed properties it is convenient to consider
that portion of the crown or canopy lying above a given elevation, z. To do so we prefer
to measure vertical distance downward from an origin at the top of the crown as is
indicated by the variable in Fig. 2.2. Taking into account the elevated crown bottom,
we define this variable over the range 0 1 in the dimensionless form:
hz
= . (2.1)
h hs
The crowns solid density is commonly expressed through the leaf area index, L,
which is defined, after Ross (1975, p. 24), as the area of leaves (upper side only) within
that vertical circular cylinder which just encloses the crown divided by the horizontal
cross-sectional area of that cylinder. The leaf area index is a dimensionless quantity
which varies with season, age, and species, and is a critical parameter controlling both
the physical and the biological processes of plant canopies (Chen and Black, 1992).
For homogeneous leaf area density, the leaf area index is written:
4acr (h h s ) 1 cr
L= Acr ( ) d = acr (h h s ) , (2.2)
b 2
0
The other crown elements (e.g., branches, twigs, stems, and pods) may be incor-
porated into a more inclusive plant (or foliage) area index, L t , defined in the same
manner as the leaf area index. If L is the area index of the other crown elements, then
L t = L + L = plant or foliage area index, PAI, (e.g., Shaw and Pereira, 1982). These
other crown elements also intercept light and contribute to the drag of the canopy. Their
importance relative to that of the leaves is a function of species (as well as season and
age) as can be seen in Table 2.1. From this table we note that for most deciduous
Needle leaves have thickness on the order of the leaf width and thus one or more of the needle
surfaces may have transverse curvature. In this case the term projected leaf area is
conventionally used to describe the needle shadow area obtained when the light source is directly
above the horizontal needle.
Solid density 21
trees the other crown elements constitute less than 10 percent of the total area, but for
agricultural crops the percentage rises markedly. Measurements for coniferous trees
are few and hard to find, but Jarvis et al. (1976, p. 179) suggest that for radiation at
least, the effective foliage units in coniferous canopies are the twigs rather than the
individual needles. We assume that the energetic and dynamic importance of these
crown elements is proportional to their relative magnitude. Our developments here
will be in terms of the plant area index, L t . However, because reported observations
of this parameter are few, we will use available values of L as a first approximation to
L t wherever necessary.
For simplicity we assume herein that the solid matter is distributed in one of two
fashions (cf. Fig. 2.3): (1) multilayer disbursed either homogeneously through-
out the crown (i.e., acr is constant) forming what we call a homogeneous multi-
layer crown, or non-homogeneously such that the foliage area index is constant
with radius from the trunk; and (2) monolayer concentrated in a thin layer of
uniform density at the crown surface forming what we call a homogeneous monolayer
crown.
The crown cross-section variation, Acr ( ), is species-dependent of course and is
idealized in Fig. 2.2. For acr = constant, dL(z)/dz is proportional to Acr ( ) from
Species L L Lt = L + L L /L t Reference
Leafy plants
Oak 4.60 0.48 5.08 0.09 Rauner (1976)
Maple 5.02 0.16 5.18 0.03 Rauner (1976)
Aspen 4.73 0.17 4.90 0.03 Rauner (1976)
Linden 4.26 0.30 4.56 0.07 Rauner (1976)
Beans 6.25 0.85 7.10 0.12 Thom (1971)
Stemmy plants
Maize 0.18 Den Hartog and Shaw (1975)
Maize 0.30 Ross (1981, p. 78)
Willow 4.50 1.50 6.00 0.25 Cannell et al. (1987)
Wheat 0.74 Ross (1981, p. 80)
The leaf area index may be estimated in any one of several ways: (1) destructive testing
involving cutting and measuring the individual leaves (e.g., Thom, 1971); (2) inclined point
quadrats which involve statistical analysis of the contacts between the leaves and a thin probe
passed at various angles through the crown (e.g., Levy and Madden, 1933; Warren Wilson, 1960,
1963; Philip, 1965); and (3) the quality of light reaching the canopy substrate (e.g., Evans and
Coombe, 1959; Jordan, 1969; Lang et al., 1985, 1991; Lang, 1986; Lang and Xiang, 1986; Lang
and McMurtrie, 1992).
We will see in Chapter 7 that this is precisely obtained only for cylindrical crowns due to the need
for the same light intensity at the lowermost leaf.
This is the situation which must prevail for tapered crowns in order to achieve the same minimum
light intensity at the base of the crown for all radii.
22 Canopy structure
Eq. 2.2. Typical observed forms of dL( )/d for multilayers are generalized by Ross
(1981, Fig. 59b, p. 216) and are reproduced here as Fig. 2.4. According to Ross (1981),
grasses and young conifers have a dL(z)/dz that most closely resembles Type 1. He
reports Type 2 to be quite common, while horse beans have the more unusual Type 4
distribution. There is ample additional observational evidence that canopy leaf area
index has a Type 2 variation in a wide variety of crops and trees. For example see
Uchijima (1976, Fig. 13), for rice and maize; Saugier (1976, Fig. 6b), for sunflower;
Amiro (1990a, Fig. 2), for pine; Norman and Jarvis (1974, Fig. 8), for Sitka spruce;
and Rauner (1976, Fig. 2), for oak, linden, and aspen. The last is reproduced here in
Fig. 2.5.
In the present work we will approximate the commonly observed Type 2 distribu-
tion for multilayers by the uniform distribution designated Type 3 in Fig. 2.4.
(a) (b)
1
Acr () and dL()/d
Leaf orientation 23
3 14
12
10 2
1 2 3
8 3 1
z (m) 6
1
2
4
0
0.10 0.06 0.02 0.2 0.6 1.0 1.4
as (m2 m3) aL (m2 m3)
Fig. 2.5. Observed vertical distributions of solid matter in typical deciduous forests.
aL = Canopy leaf area density; as = canopy non-leaf area density. 1, Oak; 2, linden; 3,
aspen. (From Rauner, J.L., 1976, Fig. 2, p. 244, Copyright c 1976 Academic Press
Ltd, with kind permission from Academic Press.)
E Leaf orientation
The orientation of the leaves in space is also of great significance to plant behavior.
Leaf projection on the horizontal plane defines the proportion of light flux that the
leaves intercept (cf. Chapter 3), while the projection on the horizontal and vertical
planes is a primary determinant of surface and form drag respectively (cf. Chapter 4).
Leaf orientation is a species characteristic which may, through growth or heliotropism,
adapt to ambient radiation conditions.
In a given species, the leaf distribution and orientation varies in the vertical di-
rection (e.g., Grulois, 1967) but evidence supports our assumption of uniformity of
orientation with respect to angular position (i.e., azimuth) about the stem at a given
elevation (Nichiporovich, 1961a; de Wit, 1965). Therefore, the most common simpli-
fying assumption is that at any elevation the leaves are distributed symmetrically about
the vertical axis of the tree. Here we will extend the assumptions to include assumed
homogeneity of both distribution and orientation throughout the crown in order to
derive a simple analytical framework for leaf orientation which will be useful later
when we take up the absorption of light by the canopy. These developments follow
closely those of Ross (1981).
Distribution functions
We define the orientation of leaves in a volume by the probability density function
(pdf) of normal directions to the top surfaces of the leaves. Consider this volume to
24 Canopy structure
rL
i
R=
r c
b
a
Ai
R=
P y
x R=
or
= sin L L L . (2.4)
out the crown, and noting that i i = 2 , we see that the fraction of the total leaf
area lying within the solid angle is 2
. Let the fraction of this last area having
normals in the rL direction be the function gL (P, rL ). If we sum over all i (that is, over
all rLi ) we obtain, in the limit
gL (P, rL )
d 1, (2.5)
2
from which we see that
gL (P, rL )
= pdf of foliage area orientation and of leaf normals. (2.6)
2
Leaf orientation 25
Since very few leaves have their top side facing downward, few normals lie
in the lower hemisphere of our infinitesimal sphere, making 2 gL
0 therein. Noting
that at any point, P, rL = rL (L , L ), and assuming there to be no preferred azimuthal
orientation, we use Eq. 2.4 to rewrite Eq. 2.5
2
gL (P, rL ) 2 1
d = dL g (P, L ) sin L dL = 1. (2.7)
2 0 0 2 L
or simply
2
gL (P, L ) sin L dL = 1. (2.8)
0
in which
and the cumulative distribution function (i.e., CDF) of leaf orientation angle is by
Plate 2.2. Typical hemispherical crown. Variety of fig, Agra, India. (Photograph by
Peter S. Eagleson.)
26 Canopy structure
definition
L
GL
gL (L ) dL . (2.11)
0
Leaf angles may be estimated in the field either through the methods of inclined point quadrats
(e.g., Warren Wilson, 1963, Philip, 1965, Miller, 1967), of gap frequencies (Lang et al., 1985),
from transmission of direct sunlight (Lang, 1986), or by painstaking use of protractors to measure
the angle of a statistically significant sample of leaves (e.g., Nichiporovich, 1961b).
Leaf orientation 27
rm
ifo
(a) Cumulative distribution
Un
0.8
classes defined by de Wit
ile
(1965); (b) Simple
ph
hil
gio
op
0.6 probability density
an
Pla
GL(L)
Pl
functions corresponding to
e
hil
top
the de Wit (1965) classes.
~
ec
0.4
Er
1 4
0.2 3 2
0
0 20 40 60 80
(a) L (degrees)
1.0
3
sin 2L
0.8
2
0.6 4
g L(L)
~
0.4
2 1
sin L cos L
0.2
0
0 20 40 60 80
(b) L (degrees)
then,
gL (rL )
|cos r
L r| = distribution function for the projection of leaf area
2
on to a plane perpendicular to the vector r(, ). (2.17)
Rape Rye
GL(L) 0.8 0.8
GL(L)
1 3 2 3
0.4 0.4
~
~
0 0
L L
Clover Potato
0.8 0.8
1 3 2
GL(L)
GL(L)
0.4 1 3 2 0.4
~
~
0 0
L L
Sugar Corn
0.8 beet 0.8
GL(L)
GL(L)
1 2 1 2
0.4 0.4
~
0 0
0 30 60 90 0 30 60 90
L (degrees) L (degrees)
giving the spatial average, throughout the crown, of the cosine of the angle between
the leaf normal and the directional vector r.
The radius vectors z, rL , and r pierce the surface of the sphere of Fig. 2.6 at three
points defining the vertices of a spherical triangle whose opposite sides we have labeled
respectively a, b, and c. From the spherical law of cosines (see mathematical tables,
e.g., Abramowitz and Stegun, 1964, p. 79)
cos r
L r = cos b cos c + sin b sin c cos , (2.19)
or
cos r
L r = cos cos L + sin sin L cos ( L ). (2.20)
Leaf orientation 29
1.2
PM
0.8
gL ( L)
AM
~
0.4 PM AM
sin L
sin 2L
0
0 30 60 90
L (degrees)
Plate 2.3. Typical spherical sector crown. Umbrella tree, Ayutthaya, Thailand.
(Photograph by Peter S. Eagleson.)
30 Canopy structure
For the special case in which r is oriented vertically upward (i.e., r z), Eqs. 2.18
and 2.20 give
= cos L , (2.21)
which is the spatial average, throughout the crown, of the cosine of the leaf inclination
angle, L . In the general case, this expectation is evaluated analytically by
2
=
gL (L ) cos L dL . (2.22)
0
Monteith (1969) gives values of as low as 0.24 for the more-or-less vertical leaves
of ryegrass, and as high as 1 for the horizontal leaves of clover. Broadleaved and
conifer trees normally have values in the range 0.30 < < 0.50 (see Jarvis et al.,
1976, and Rauner, 1976). A collection of field estimates is included herein (see
Tables 2.2, 2.3, and 2.4).
Estimates of derived from the distributions of Eqs. 2.122.15 corresponding to
the de Wit (1965) classifications are given in Table 2.5. We see that these distributions
give values of which are generally larger than found for trees but agree well with the
distributions observed for crops, as proposed by de Wit (1965) and as seen in Figs. 2.8
and 2.9. Also in Table 2.5 we give the range of observed for crops and grasses, and
the average for deciduous and coniferous trees as assembled from the literature and
listed in Tables 2.22.4. Note the higher values of observed for crops and grasses
than for trees, and note the general correspondence of the range of observed for
crops and grasses with the range of for the de Wit (1965) distribution types. We
conclude from this that the de Wit distributions are not applicable to trees, but are well
suited for crops and grasses.
Henceforth we will drop the crown average notation, ,
and use in which the averaging is
understood.
Leaf orientation 31
Type aw = L t Lt Reference
Shrub
Chaparral 2.5 Mooney et al. (1977)
Desert shrub 1.0 Larcher (1983)
Average 0.45 2.77 5.4 0.44
a
Average of 5 values.
b
Includes entire crown.
c
Dynamically active crown only.
32 Canopy structure
If we take the leaf area, AL , as statistically distributed with respect to angle in the
same way as is the surface area of a hemisphere (Nichiporovich, 1961b), the foliage
area is oriented equally in all hemispheric directions. That is, the distribution of foliage
orientation is hemispheric. Further assuming axial symmetry, we can use the sketch
of Fig. 2.10 to express the -dependence of leaf area in this case as
D
dAL = ( D sin ) d. (2.23)
2
Scaling by the total surface area of the hemisphere, we obtain the probability density
function of leaf area, f(AL ), which is identical to that of leaf angle,
gL (L ), as
dAL
gL (L ) f(AL ) = AL1
= sin sin L , (2.24)
d
which we recognize as the pdf of leaf inclination for erectophile plants. Equations 2.22
and 2.24 then give, for the average leaf angle
2
= sin L cos L dL = 0.5. (2.25)
0
Type aw = L t Lt Reference
Type aw = L t Lt Reference
This value can be compared with observations in Table 2.5. It appears that coniferous
trees, having = 0.45, and being common to moist, cloudy climates, may exhibit this
dependence of on the direction of light.
For direct (as opposed to diffuse) light, the proposition that the leaf is perpendicular
to the incident light leads directly to the relation
= cos h = sin (h ), (2.26)
2
in which h is the effective solar altitude. In Chapter 3 (Fig. 3.9) we will see that over the
latitude range 0 60 , the daylight-hour average solar altitude at summer solstice
is 27 h 40 whence, from Eq. 2.26, 0.46 0.65. The observed average for
deciduous trees in Table 2.2 is 0.31. We conclude from this that the leaf surface
of deciduous trees is approximately 26 degrees more vertical than is the normal to the
direction of light. In Chapter 3 we will consider these important questions in more
detail.
F Leaf type
There are basically two types of leaves, broad leaves and needle leaves.
Broad leaves
So-called broad leaves are the thin flat leaves, anywhere from 2 cm to 25 cm in
characteristic dimension depending upon species. In plan view, again depending upon
species, they may be star-shaped with a varied number of points, or they may be heart-
shaped with serrated edge, or have multiple lobes. Their stomates are predominantly
on the underside of the leaf giving them the name hypostomatous. However, the leaves
a
See Tables 2.22.4.
Leaf type 35
d L dAL(L )
of grasses and herbaceous broadleaved plants have stomates on both the upper and
lower surfaces and are called amphistomatous.
Needle leaves
There is a wide range of geometries within the group of needle-leaved plants, but we
will identify and treat only the two extremes:
Spruce needles have a + cross-section such as a square with four deeply fluted quadrants
(cf. Zim and Martin, 1987, p. 29).
3
Radiant uxes
Characteristics of incident solar radiation and its partition are described, and
the average insolation during the nominal growing season is determined as
a function of latitude. Penetration of radiation into the canopy is considered
theoretically as a function of solar altitude and leaf angle, and its extinction
coefficient is compared with observation.
Maximum canopy absorption of solar radiation is reasoned to occur when
the light at the lowest leaves is at the biological limit called compensation
intensity, and when by virtue of scattering the coefficient of light absorption
equals the cosine of the leaf angle as is shown by the observations. These
conditions define the state of optical optimality that appears to govern those
systems that are productively limited by available photosynthetically active
light. Systems limited by water and apparently by leaf temperature are observed
to regulate leaf angle in ways that accept or reject heat-producing long-wave
radiation.
A Introduction
The average rate at which radiant solar energy passes through a plane unit area normal
to the Suns rays is referred to here as radiant flux density. At the outer limit of the
Earths atmosphere this quantity is termed the solar constant; it is given the symbol
S oo and has the value (Frohlich and Brusa, 1981):
37
38 Radiant uxes
in which o is the distance between the Sun and Earth in astronomical units, AU
(1 AU = mean SunEarth distance = 1.496 108 km), and h is the solar elevation.
Spherical trigonometry gives
where is the geographical latitude, is the solar declination (see List, 1951), and is
the solar hour angle (with zero at apparent noon and positive values in the afternoon).
For more details of this celestial geometry see the illustration and caption of Fig. 3.1
as taken from Eagleson (1970, p. 404).
Observers
meridian
Z Observers zenith
Sun
North Celestial
celestial P equator
N S
Pole
Solar
altitude
h
North O South
Fig. 3.1. Great circles on the celestial sphere. Consider an observer at point O on the
Earths surface. A coordinate system is established having its vertical perpendicular to
the tangent plane at this point. This coordinate system is defined as topocentric, and
the measures of solar position seen from point O are called apparent measures.
Transferring our tangent plane down the radius until point O coincides with the Earths
center we have the geocentric coordinate system shown in Fig. 3.1. The celestial
sphere (in this case the solar sphere) is defined by the imaginary spherical shell
passing through the Sun and having its center at point O. The observers horizon will
then be the great circle defined by the intersection of a plane, tangent to the Earths
surface at point O, with the celestial sphere. The celestial equator is another great
circle given by the intersection of the celestial sphere with a plane through the Earths
equator. A perpendicular to the equator plane, erected at point O, pierces the celestial
sphere at the celestial pole P; and a perpendicular to the observers horizon plane,
erected at point O, pierces the celestial sphere at the observers zenith Z. A great circle
on the celestial sphere and including both P and Z is called the observers meridian.
The angle between the equator plane and an eastwest plane containing Z and O is
the astronomical latitude and is positive for points O in the northern (terrestrial)
hemisphere. A great circle on the celestial sphere containing the Sun S and both
celestial poles PN and PS is known as the solar hour circle, and the angular distance
between the planes of the meridian and the Suns hour circle is known as the Suns
hour angle . It is positive for Sun positions to the west of the meridian. A plane
containing S, O, and Z defines another great circle along the celestial sphere. This
plane is perpendicular to the observers horizon plane. The solar altitude is defined by
the angle h measured along this great circle, between the observers horizon plane
and the plane containing the solar hour circle. The angular distance between the
celestial equator plane and the Sun, measured from the former (and positive when the
Sun lies north of the Earths equator) and along the solar hour circle, is known as the
solar declination .
Denitions 39
B Denitions
We define the incident radiation field in terms of its energy, in which case the primary
characteristic is its radiant flux density, R, measured in cal cm2 min1 . The radiant
flux density or radiance for short, is a function of the wavelength, , of the radiation;
the time, t; the location, P, in space; and the direction, r, of the ray. Accordingly, we
write, formally
radiance = R (t, P, r) = rate at which energy in the wavelength interval around ,
and in the increment of solid angle, d, around the ray, r, passes through a unit of
surface area, d , perpendicular to the ray at the particular time, t, and location, P.
This is illustrated in the definition sketch of Fig. 3.2a where is the angle between
the ray, r, and an (upward) normal, n, to an arbitrarily oriented surface. If R is
independent of direction, r, the radiation field is isotropic. Only blackbody radiation
is strictly isotropic, but we will see that diffuse sky radiation under overcast skies
approximates this condition.
While the flux of radiation through a unit area perpendicular to the ray, r, has been
termed radiance, and is independent of the surface on which it is incident, the flux
of radiation through an arbitrarily oriented surface (such as a leaf) is known as the
irradiance of the surface and depends upon the surface orientation. That is, radiance
refers to the ray, and irradiance refers to the surface. The flux of radiation, F , through
40 Radiant uxes
d (solid angle)
r
R
P
z
r
z R
h
y
(b) x
Fig. 3.2. Definition sketches. (a) Elementary cone for radiance. (After Kondratyev,
K.Ya., Radiation in the Atmosphere, Fig. 1.2, p. 10, Copyright c 1969 Academic
Press, with kind permission from Academic Press, Orlando, FL.) (b) Spherical
coordinates for radiant flux.
,
F = R cos nr (3.4)
in which nr is the angle between the vectors n and r, and the operator . . signifies
local spatial averaging over the many surface orientations (e.g., leaves) present at
location, P.
In this work we are interested only in climatic issues. Therefore, we will describe
climatic radiance solely in terms of its first-order dependent variables:
(1) latitude, , and
(2) growing season timing, t, and length, .
Thus, we will deal with radiances that are heavily averaged, both in space at a given
latitude, and in time over an ensemble of growing season months. The characteristic
radiances at a given latitude which result from this process are therefore nominal
and may be expected at best to explain only first-order spatial variations in vegetation
form and function. To indicate that the radiances have been so averaged we give them
an overbar. The time-averaged form of Eq. 3.4 is then
.
F = R cos nr (3.5)
Denitions 41
Plate 3.1. Giant Ficus walk toward the light. At Angkor, Cambodia, spong trees
(Ficus tetrameles nudiflora) step over the wall surrounding the ruins of Ta Prom
seeking access to light available where the temple created a gap in the canopy.
(Photograph by Peter S. Eagleson.)
A special case of particular interest here is when the surface of Fig. 3.2a
has a horizontal orientation. The irradiance of this unit horizontal area is known as the
insolation, I . In terms of averaged quantities we use Eq. 3.5 and Fig 3.2b to write
Direct radiation
Radiation coming directly from the sun without modification by the absorption and
scattering effects of intervening clouds is called direct or beam radiation, Rb , and is in
the spectral range 280 nm < < 4000 nm (Ross, 1981, p. 159). About half of this (van
Wijk, 1963) is in the region of photosynthetically active radiation, PAR, where 400 nm
< PAR P < 700 nm, and about half is in the near-infrared region, NIR, where 700
nm < NIR N < 4000 nm. A small amount occurs in the ultra-violet region, UV,
where UV < 400 nm; however we neglect that here and assume PAR = NIR = Rb /2.
We will concern ourselves primarily with PAR in this work, thus all radiation symbols
refer to PAR unless stated otherwise. The PAR absorbed by the atmosphere and by the
surface heats these receptors to earthly rather than solar temperatures at which they
42 Radiant uxes
Ib (h ) = Rb (h ) sin(h ), (3.7)
in which R b is the (seasonal and zonal) average solar (i.e., direct) radiance in the PAR
band at the top of the canopy for solar altitude, h , and h is the average (daylight-
hour) solar altitude during the growing season.
Diffuse radiation
That part of solar radiation which reaches the vegetation from above after multiple
scattering from atmospheric particles and clouds as well as after reflection from the
vegetation (and ground) and additional atmospheric scattering is called the diffuse sky
radiation, Rd (Ross, 1981, p. 161). Its spectral distribution is different from that of direct
radiation. In general, the angular distribution of diffuse radiation is non-isotropic. In
this case, radiation is incident upon our unit surface area from all sectors of the sky,
and the seasonal and zonal average diffuse insolation, Id , is
in which z is the complement of h . Using Fig. 3.2b, Eq. 3.8 can be written
2
2
Id (h ) = d Rd (r: h ) sin z cos z dz
0 0
2
= Rd (r: h ) sin 2z dz . (3.9)
0
The integrand of Eq. 3.9 defines the diffuse radiance on a horizontal surface. This
radiance is also called the zonal brightness of the sky, D dz , and its angular distribution
is symmetrical with respect to the solar vertical plane.
Magnitudes
On average, the direct and diffuse components of the total radiation are distributed
among the three primary spectral bands (i.e., UV, PAR, and NIR) as is shown in
Table 3.1 taken from Ross (1975, Table II, p. 19). The UV component is negligible for
our purposes. Only a small amount of the total incident energy, less than 10 percent,
Denitions 43
is used directly by the plants in photosynthesis (cf. Gates, 1980, p. 26; Larcher, 1983,
p. 20) where it is stored chemically in the high-energy organic compounds of biomass.
This is from the PAR component. The remaining majority is converted into heat where
it powers transpiration, catalyzes photosynthesis, and exchanges convectively with
the atmosphere. The mixture of direct and diffuse radiation varies with place and time.
Without distinguishing among the spectral components, meteorologists define this situ-
ation by the beam ratio, , which is the direct radiance fraction of total radiance. Values
of under conditions of clear atmospere and industrial cloudiness are presented in
Fig. 3.3 as a function of solar altitude, h , using data from Gates (1980, Table 6.1,
p. 114).
0.6 cloudiness
0.5
0.4
0.3
0.2
0.1
0
0 20 40 60 80
Solar altitude, h (degrees)
Waveband (m)
UV PAR NIR
(0.280.38) (0.380.71) (0.714.0)
Limiting cases
As stated earlier, we will recognize two limiting radiation cases:
(1) Very clear sky in which there is no diffuse radiation ( i.e., beam ratio = 1), the
Suns rays are everywhere at the same inclination, and the averaged insolation
is given by Eq. 3.7.
(2) Overcast sky in which there is no direct radiation (i.e., beam ratio = 0), and
the averaged insolation is given by Eq. 3.9.
Observations show (Poldmaa, 1963, reproduced in Ross, 1981, Fig. 41, p. 164) that
under overcast skies, the diffuse radiation is closely isotropic. Under isotropic condi-
tions, R d is independent of r and Eq. 3.9 reduces to
I d = Rd, (3.10)
which is Lamberts law (Kondratyev, 1969, p. 13). The zonal brightness (cf. Eq. 3.9)
may be normalized to give the relative brightness of the sky, D dz , under isotropic
conditions, as
Ddz
D dz () = = sin 2z . (3.11)
Rd
Equation 3.11 is compared with Poldmaas (1963) observations in Fig. 3.4 as
adapted from Ross (1981, Fig. 41, p. 164). The excellent agreement confirms the
isotropic assumption for the radiance of an overcast sky.
C Partition of radiance
Total beam radiance, Ri , incident upon the canopy is partitioned by the foliage ele-
ments, into three components: absorption (Ra ), reflection (Rr ), and transmission (Rt )
Partition of radiance 45
a
Note that, by definition, irradiance is the flux density
experienced by a surface perpendicular to the beam and
when the surface under consideration is horizontal the
flux density is often referred to as insolation.
Source: Ross, J., Radiative transfer in plant communities,
Table V, p. 22, in Vegetation and the Atmosphere, Vol. 1,
Principles, edited by J.L. Monteith, Copyright c 1975
Academic Press, with kind permission from Academic
Press Ltd, London.
Remember that by definition irradiance is perpendicular to the surface in question.
46 Radiant uxes
graphically in Fig. 3.6 (Larcher, 1983, Fig. 2.4, p. 10, after Gates, 1965) where the full
spectral distribution of the coefficients is shown for a poplar (Populus deltoides) leaf.
Consider next the sensitivity of radiance partition to the angle of radiation incidence.
Ross (1981, Table II.3.2, p. 185) has examined this for isolated leaves of lettuce (thin
and pale) and Chinese hibiscus (thick and glossy) under incident white light, and his
results are presented here in Table 3.3. Note:
0 30 40 50 60 70 75
Lettuce
0.175 0.17 0.16 0.145 0.12 0.11 0.10
0.14 0.145 0.155 0.165 0.19 0.23 0.25
0.685 0.685 0.685 0.69 0.69 0.66 0.66
Chinese hibiscus
0.06 0.055 0.05 0.045 0.04 0.04 0.04
0.065 0.075 0.08 0.085 0.095 0.11 0.13
0.875 0.87 0.87 0.87 0.865 0.85 0.83
Note: Leaves of lettuce (Lactuca sativa) were thin and pale green in color, while
those of Chinese hibiscus (Hibiscus rosa sinensis) were thick and glossy.
Source: Ross, J., The Radiation Regime and Architecture of Plant Stands, Table
II.3.2, p. 185, Copyright
c 1981 Dr W. Junk Publishers, The Hague, with kind
permission from Kluwer Academic Publishers.
UV PAR NIR
100 0
Reflection
Transmission (%)
80 20
Reflection (%)
60 40
Absorption
40 60
20 80
Transmission
0 100
0.4 0.6 0.8 1 2
Wavelength (m)
Fig. 3.6. Spectral partition of radiance by leaf of poplar (Populus deltoides). (From
Larcher, W., Physiological Plant Ecology, corrected printing of the 2nd edn, Fig. 2.4,
p. 10, Copyright c 1983 Springer-Verlag Publishers; after Gates, D. M., Energy,
plants and ecology, Fig. 3, p. 5, Ecology, 46/1, 2, 113, Copyright c 1965 Ecological
Society of America, with kind permission from Springer-Verlag GmbH & Co. KG,
W. Larcher, and the Ecological Society of America.)
Partition of radiance 47
Table 3.4. Partition of total beam irradiance of upper surface of tree leavesa
a
Values are measured hemispherical reflectances and transmittances for normal
incidence with absorptances calculated from Eq. 3.12.
b
See listing of Common and scientific names used, p. xxxvi.
Source: Birkebak, R., and R. Birkebak, Solar radiation characteristics of tree
leaves, Table I, p. 648, in Ecology, 45/3, 646649, Copyright c 1964 Ecological
Society of America, with kind permission from the Ecological Society of America,
Washington, DC.
48 Radiant uxes
Moving now from individual leaves to the optical properties of a dense stand of
trees, Ross (1975, Fig. 11, p. 42) demonstrates theoretically the important sensitivity
of the reflection coefficient to the angle of incidence of the beam radiation as is shown
here in Fig. 3.7. Using the series approximation suggested by Goel (1989, p. 209) we
have added the following empirical approximations of the NIR and PAR curves for
use in estimation:
NIR = 0.60 0.31 h 0.03 (h )2 , (3.13)
and
PAR = 0.08 + 0.01 h 0.03 (h )2 , (3.14)
in which the solar altitude, h , is in radians. Note that NIR is also sensitive to ,
and therefore Eq. 3.13 is a good approximation only for a uniform distribution of leaf
normals.
Neglecting the small UV component we see, as above, that the reflection of the
PAR is small and is relatively insensitive to incidence angle, while the reflection of
NIR is large and is highly sensitive to incidence angle. In principal, this NIR behavior
provides a primary means of plant temperature regulation, climatically through natural
selection and adaptation for those average leaf angles that allow for spilling or retention
of NIR, and transiently through either heliotropism to follow the sun or turgor loss
(i.e., wilting) to spill the radiation. However, Fig. 3.7 shows that over the practical
range 27 h 40 , a 90 change in L will change NIR by at most 10 percent.
0.6
Nominal
range
NIR
Reflection coefficient,
1
0.4
NIR
3,4
2
UV
1
0.2
3
2
PAR
1
PAR
0 2,4
0 30 60 90
Solar altitude, h (degrees)
Fig. 3.7. Theoretical reflection coefficients of dense stands for direct radiation.
1, Horizontal leaves; 2, vertical leaves; 3, uniform distribution of leaf normals;
4, empirical approximations, Eqs. 3.13, 3.14. (From Ross, J., Radiative transfer in
plant communities, Fig. 11, p. 42, in Vegetation and the Atmosphere, Vol. 1, Principles,
edited by J. L. Monteith, Copyright c 1975 Academic Press Ltd, with kind permission
from Academic Press Ltd, London.)
To be shown later in Fig. 3.9.
Nominal average insolation 49
0
0 20 40 60 80
Solar altitude, h (degrees)
Henceforth, we will work solely with climatic variables averaged over the growing season and for
simplicity will discontinue the overbar notation for such time averages.
50 Radiant uxes
Fig 3.8. For overcast skies de Wit (1965) estimated the insolation at h = 90 to
be 20 percent of that for very clear sky and distributed it with solar altitude according
to the isotropic assumption to be discussed in a later section of this chapter. This is
shown as Curve no. 4 of Fig. 3.8. We can replace the variable h of Fig. 3.8 with
latitude, , through Eq. 3.3 provided we assume representative values for the solar
declination, , and the solar hour angle, .
Solar declination
We fix the solar declination at a value representative of the vegetation growing season.
As we will see, the growing season changes in length with latitude and varies some-
what in timing according to the climate. However, the first-order accuracy of these
radiation estimates suggests using a single average value of the declination as being
representative of the growing season irrespective of latitude and climate; we choose
the summer solstice, = 23 . Northern hemisphere daylengths for this declination are
given by Trewartha (1954, p. 12) and are listed here in Table 3.5.
a
From Trewartha (1954, p. 12).
b
o = 2 1
2
2 =sunset sin o
c
cos = 2
1
1 1 =sunrise cos d o
Nominal average insolation 51
60
50
Summer
40 solstice
30
Average for
20 daylight
hours
10 Autumnal
equinox
0
90 70 50 30 10
Latitude, (oN)
of the vegetation. Given the effective solar altitude, the tree appears to achieve this
temperature control by setting the leaf angle at a value which compromises between
reflecting heat-producing NIR while absorbing biomass-producing PAR.
We begin to examine this relationship by setting the effective solar hour angle at
the daylight-hour average.
which gives the desired h vs. relation for the summer solstice plotted in Fig. 3.9.
For comparison purposes, the relation for the autumnal equinox, = 0 , cos = 2 ,
is also given in Figure 3.9. For further comparison, the h vs. relation for noon-
time at the summer solstice is obtained from Eq. 3.3 with = 0 and is also plotted
in Fig. 3.9.
Finally, Fig. 3.9 allows transformation of Fig. 3.8 into the nominal growing season
daylight hour average PAR (on a horizontal surface), I0p , as a function of latitude
Note that this choice minimizes the daily extremes of the angle of incidence of solar radiation on
the leaves and hence minimizes scattering thereby maximizing the radiation available for
photosynthesis.
Remember, we are now omitting the overbar indicator of temporal averaging.
52 Radiant uxes
0.5
Very clear sky line, autumnal equinox; a,
horizontal surface, Iop (cal cm2 min1)
0.3
a
0.2
Overcast sky
0.1
Total and
diffuse light
0
90 70 50 30 10
Latitude, ( N)
(cf. Fig. 3.10). Curves for the autumnal equinox ( = 0 ) are added to the figure for
comparison. At any given location the actual average insolation will lie somewhere
in between the bounding very clear and overcast conditions. Jensen and Haise
(1963) compiled observations of average annual total insolation in the latitudinal range
25 N < < 48 N, and the regression of these values (PAR component only) is added
to Fig. 3.10 for reference.
Summer deciduous trees (e.g., Mediterranean and tropical climates) shed all or part of their
leaves in the dry season, whenever it occurs during the calendar year.
Ahlgren (1957) finds that red pine, white spruce, and black spruce begin activity when the
atmospheric temperature rises above freezing. Fraser (1966) concludes that photoperiod controls
the growth cessation of black spruce. Baldwin (1931) finds the initiation of evergreen growth to
be temperature dependent and its cessation of growth to be undetermined. However, Romell
(1925) finds evergreen growth initiation insensitive to both latitude and altitude.
Length of growing season 53
20 pinyonjuniper
200 Spruce, fir Table 3.6. (a) multilayer,
23
18 temperate climate;
150
21 5 3 (b) monolayer, temperate
m (days)
19 14
b
0
30 35 40 45 50 55 60
Latitude, ( N)
For background on bud break and leaf color change see Valentine (1983), Lechowicz (1984),
Taylor (1974), Gee and Federer (1972), and Mitchell (1936) among others. For models of season
length see Britton (1878a, 1878b), Nemeth (1973), and Federer and Lash (1978).
Various investigators report low-level conifer productivity to super-cooled temperatures (cf. Arris
and Eagleson, 1989b).
Table 3.6. Observations of the growing season of trees
Growing season
1 Russian plain 56 birch, aspen 130 258 128 Foliation to color change Rauner (1976)
2 Canada (Boreal) 52 black spruce 142 236 94 Bud break to terminal Morganstern (1978)
bud set
3 Russian steppe 50.5 oak 120 274 154 Foliation to color change Walter and Breckle
(1985)
4 Canada (Great 47 black spruce 144 250 106 Bud break to terminal Morganstern (1978)
Lakes) bud set
5 Southern Quebec 47 white pine 165 Unknown Marie-Victorin
(1927) a
6 Canada (Acadia) 46 black spruce 144 258 114 Bud break to terminal Morganstern (1978)
bud set
7 New England 45 white pine 133 235 102 199% growth of Baldwin (1931) b
sapling leader
8 New England 45 red spruce 162 234 73 199% growth of Baldwin (1931) b
sapling leader
9 New England 45 balsam fir 157 237 81 199% growth of Baldwin (1931) b
sapling leader
10 New England 45 white spruce 157 231 75 199% growth of Baldwin (1931) b
sapling leader
11 Berlin, NH 45 white cedar 138 247 110 199% growth of Baldwin (1931) b
sapling leader
12 Cupsuptic Lake, ME 45 black spruce 163 242 79 199% growth of Baldwin (1931) b
sapling leader
13 Mt. Desert Isle, ME 44 red spruce, 156 232 76 Growth of leader Moore (1917) a
white pine
14 Mt. Desert Isle, ME 44 balsam fir 152 191 39 Growth of leader Moore (1917) a
15 Austria 43 Austrian pine 141 200 59 Growth of leader Nakashima (1929) a
16 Adirondack Mts., NY 43 red spruce 166 212 46 Growth of leader Rees (1929) a
17 Ithaca, NY 42 white pine 120 227 107 Growth of leader Brown (1912, 1915) a
18 Wauseon, OH 41.5 mixed deciduous 105 288 183 Bud break to leaf fall Smith (1915) b
19 Pennsylvania 40.7 white pine, 152 182 30 Growth of leader Illick (1923) a
Norway spruce
20 North Carolina 35 tulip poplar, 95 297 202 Leaf emergence to Reader et al. (1974) b
red maple leaf fall
21 Flagstaff, AZ c 34 ponderosa pine 169 Avg. date of latest spring Williams and Anderson
frost to avg. date of (1967)
earliest fall frost
22 Flagstaff, AZ c 34 pinyonUtah 215 Avg. date of latest spring Williams and Anderson
juniper frost to avg. date of (1967)
earliest fall frost
23 Flagstaff, AZ c 34 pinyonalligator 182 Avg. date of latest spring Williams and Anderson
juniper frost to avg. date of (1967)
earliest fall frost
24 Arizona 32 ponderosa pine 135 213 78 Growth of leader Pearson (1924) a
a
As quoted by Baldwin (1931).
b
Extensive observations at many sites.
c
2100 m elevation.
56 Radiant uxes
temperature
Larcher, 1983, Fig. 3.35,
Optimum
50 p. 114.)
0
1.2 0.8 0.4 0 0.4 +0.8
Relative leaf temperature
the data of Table 3.6 as observations of the complete growing season, and we will use
their empirical correlations with latitude (cf. Fig. 3.11) in later chapters where season
length plays a role.
Analyzing the data of Table 3.6, we find that the average Julian day at which the
sprucefir season begins is 157 (i.e., early in June) with a standard deviation (sample
size of 24) of only 11 days. This supports the earlier observation of Romell (1925)
that evergreen growth initiation is insensitive to latitude and altitude. In contrast, the
season length is highly sensitive to latitude having an average value of 75 days with
standard deviation of 21 days. This puts the mean cessation of sprucefir growth at
Julian day 232 which is mid-August, the peak of summer heat, suggesting that high
temperature may control termination of growth in these species.
A possible mechanism for this control is the asymmetric sensitivity of evergreen
photosynthetic capacity with variations of the deviation in leaf temperature from its
optimum value. This is sketched in Fig. 3.12 where we see the abrupt shutdown of
photosynthesis when the leaf temperature reaches a critically high value. Note in this
figure that the range of productive temperatures less than the optimum is about twice
that of temperatures greater than the optimum. With this in mind we see that increasing
latitude brings an increasing photosynthetic season as measured from initiation in early
June until cessation when the temperature reaches 0.7 the optimum leaf temperature
for the given sprucefir species.
We discuss this biochemical mechanism further in Chapter 8.
For an exhaustive analytical analysis of this issue see Ross (1981, Chapter II.4, pp. 188238).
Penetration of direct radiation 57
and Nilson (1965), Warren Wilson (1965, 1967), and Anderson (1966), Cowan (1968)
derived this attenuation as follows:
Let Rbo be the incident beam radiance at the top of the vegetation canopy for
solar altitude h . This radiation is intercepted and partitioned progressively by the
leaves as the beam penetrates the stand, leaving Rb unintercepted at vertical depth
. In a cascading process, there is at each leaf surface reflection, transmission, and
absorption, and some of the reflected and transmitted energy is partitioned anew at
the bottom side of other leaves. The sum of the multiple reflections and transmissions
is called scattering, and it results in additional absorption of radiation. Forward (i.e.,
downward) scattering results in a reduced decay of the penetrating beam. Bookkeeping
these many internal canopy reflections and transmissions is a tedious and approximate
task that is at a much higher level of detail than is called for in this work. Cowan
(1968) ignores these details and assumes the change in Rb per unit length of optical
path to be separably proportional to Rb and to the foliage area density in the manner
dRb
= G L t Rb , (3.16)
d/sin(h )
in which
G G(h ) = dimensionless proportionality function representing the spatial
arrangement and orientation of the leaves in relation to the solar altitude,
Lt
1
= L t = homogeneous canopy foliage area density index (i.e., one-sided
leaf area per unit horizontal area per unit of dimensionless crown depth), and
sin(h )
= distance in the direction of the penetrating beam.
The fraction of incoming energy escaping back to the sky constitutes the albedo of the canopy.
For more details of the formulation of this process see Sellers (1989).
58 Radiant uxes
Sun z
rL
R= i
c
b
rs
a
Ai
L
z
R=
h P y
x R=
(Monteith, 1973, p. 52). Then, using Eqs. 3.16 and 3.17, the shadow area, As , per unit
horizontal area of canopy, is
dRb G(h )L t
As = = b L t , (3.18)
Rb d sin(h )
in which G(h ) is called the G-function (Ross and Nilson, 1965). In the above:
G(h )
b = shadow area per unit of foliage area, and
sin(h )
G(h ) is the average, over the canopy, of the projection of a unit foliage area on a plane
normal to the sun direction, rs , at arbitrary point P (see Fig. 3.13 for definitions), and
as such it is the average cosine of the angle between the direction rs , and the effective
leaf normal, rL . The nature of b is further clarified by exploring its relation to solar
altitude and leaf angle using an idealized simple geometry.
z Specular
reflection
r
L
L
+
+ rs
G
h
x
1
h L
rs
L
r
L
G
h
h
1 x
Specular
reflection
Leaf
cot(h )sin L
(1)
b
(b)
Fig. 3.14. Geometry of specular reflection from opaque leaf in two dimensions.
(a) Upward specular reflection; (b) downward specular reflection.
reflection. From either case we use the above definition of the G-function to write
G cos = cos (h + L ) = sin(h + L ), (3.19)
2
so that from Eq. 3.18, the extinction coefficient, b(1) , of an isolated leaf is
Fig. 3.15. Graphical summary of beamleaf geometry. Data keyed to Table 3.7.
, Deciduous; , evergreen; Curve (a), empirical correlation; Curve (b), optical
optimality for multilayers (Eq. 3.51).
From Fig. 3.14 we see that b is the horizontal shadow area per unit of (opaque)
leaf area (Monteith, 1973, p. 52). It is easy to visualize how in reality diffraction at the
leaf edges, coupled with re-reflections from the underside of other leaves and transmit-
tance through the leaf and the interleaf gaps all act to reduce this shadow and hence to
bring b .
In Fig. 3.15 the beamleaf geometry is summarized in h space. The function
= sin(h ) divides the space into regions of positive and negative , while the
function = sin(h /2) separates regions of upward and downward specular reflec-
tion. Assuming vegetation growth is limited to latitudes between the Equator and
60 , the effective solar altitude, h , at summer solstice is 27 h 40 as given
by Fig. 3.9. These limits on h are plotted as the vertical dashed lines on Fig. 3.15.
A few observations of from the literature are listed in Table 3.7 and are plotted
(with identifying numbers) on Fig. 3.15 where we note:
(1) specular reflection is upward in all cases,
(2) leaf angles in the temperate deciduous forest (i.e., points nos. 1, 3, 4, 5, 6) are
larger than that for normality with the beam,
(3) leaf angles at the coldest latitude (i.e., no. 4) produce nearly horizontal specular
reflection,
Penetration of direct radiation 61
(4) leaf angles at the hottest latitude (i.e., tropical rainforest, no. 7) produce nearly
vertical (h + 2 L = 90 ) specular reflection, and
(5) the single evergreen observation (i.e., no. 11) has leaves almost normal to the
beam.
The primary variability in b resides in the three-dimensional angular relationship
between the solar beam and the leaf surface since this determines the distribution of
gaps between the leaves, and it is through these gaps that the beam propagates. The
frequency of gaps along the path of a given solar ray has been studied extensively
in the field by the method of point quadrats which involves passing very thin, long
needles through a vegetation stand in a fixed direction and recording all contacts with
the foliage. These observations have guided theoretical study of the significance of
gap distributions under the leadership of Warren Wilson and Reeve (1959, 1960),
Warren Wilson (1960, 1963), Anderson (1966), and Nilson (1971) and they lead to the
following improved prediction of b using more realistic geometry albeit still without
consideration of scattering.
Latitude, h
Observation Species (degrees) (degrees)a Reference
a
Effective solar altitude as given by daylight-hour average at summer solstice.
62 Radiant uxes
g (L ) = (L o ), (3.23)
where ( ) is the Dirac delta-function. Using Eq. 3.23, Eq. 3.21 reduces to
G(h ) , L h
b = , (3.24)
sin(h ) {1 + 2 [tan(os ) os ]}, L > h
in which L = o . Eq. 3.24 is plotted in Fig. 3.16 as taken from Anderson (1966,
Fig. 1, p. 44). Several specific cases of constant leaf inclination are of particular
interest:
Horizontal leaves (o = 0)
G(h )
b = = 1, (3.25)
sin(h )
Vertical leaves (o = 2
)
With o = 2 , Eq. 3.24 is indeterminate for h < L , so we return to Eqs. G.05 and
G.07 to find (as did Ross, 1981, p. 22) that = 0 and
G(h ) 2
b = cot(h ), 0 h . (3.26)
sin(h ) 2
Leaves perpendicular to solar beam (o = 2
h)
cos(os ) = cot2 (L ) = tan2 (h ). (3.27)
Mathematically, os is the supplement of the azimuthal angle (cf. Fig. 3.13) at which sign change
occurs.
Penetration of direct radiation 63
0
10 30 50 70 90
Foliage inclination, L (degrees)
Plate 3.2. Typical equatorial crown shape. Trees of this equatorial Kenyan savanna
elevate their monolayer leaf surface above the shrubs for cooling and display a
spherical-sector crown of large radius in keeping with the near-zenith position of the
Sun during the growing season. (Photograph by Peter S. Eagleson.)
64 Radiant uxes
The term hemispherical distribution is used interchangeably here with uniform distribution to
describe the leaves having a spatial distribution identical to the surface of a hemisphere. That is,
the distribution of area with hemispherical radius vector is constant, i.e., the fractional area is the
same in all vector directions on the hemisphere.
Penetration of direct radiation 65
Opaque hemispheres
Once again, the geometry dictates
R2
shadow area [1 + sin(h )] 1 1
b = 2
= 1+ . (3.33)
surface area 2 R 2 4 sin(h )
0.2
0.4
0.6
Eq. 3.34
with Lt = 5.0
0.8
Leaf orientation = vertical uniform horizontal
d d d
0.55 0.75
0.84
1.0
1 101 102
Id ( = 1)/Id 0
Fig. 3.17. Extinction of diffuse radiation from a uniform sky. Observations: , maize;
+, cotton. Solid lines = theory, Eq. 3.34 with L t = 5.0; dashed lines = exponential
approximation, Eq. 3.35. (From computations of Ross, 1975, Table VII, p. 34.)
This is strictly true only for horizontal leaves (Ross, 1981, p. 247) as is seen for a special case in
Fig. 3.17.
Penetration of total radiation 67
horizontal, vertical, and uniform leaf orientations in comparison with his observations
in maize and cotton canopies. In that figure we see:
(1) the agreement between the observation and the uniform leaf orientation is
fairly good for maize but is poor for cotton,
(2) the observations agree well with Eq. 3.35 (dashed lines) in which the fitted
values of the extinction coefficient, d , for diffuse radiation are about 0.75 and
0.84 for maize and cotton respectively, and
(3) Equation 3.35 fits the vertical leaf distribution approximately with d = 0.55,
and fits the uniform distribution fairly well with d = 0.75.
I ( )
= exp (b L t ) + (1 ) exp (d L t ). (3.36)
I0
At large effective solar altitudes, and for a clear atmosphere, Fig. 3.3 shows that
1
= O(101 ), (3.37)
whereupon, if b d , Eq. 3.36 is approximately
I ( )
= exp ( Lt ), (3.38)
I0
in which is the scatterless extinction coefficient for total clear sky radiation, and
where for clear sky
b . (3.39)
Overcast sky
For overcast skies, we neglect the beam radiation whereupon the extinction is again
governed by Eq. 3.35 in which
= d . (3.40)
68 Radiant uxes
Estimated scattering
Absent an appropriately simple formulation of the scattering phenomenon we will
estimate the sense and the approximate magnitude of its modification to the scatterless
extinction coefficients derived above. We begin by recalling (cf. Fig. 3.14) that the
effect of scattering is to reduce the size of the shadow. It may be thought of crudely as
scatterless radiation in a crown with reduced foliage area index, L t , where we let
L t = (1 s )L t , (3.41)
with s being a scattering coefficient assumed here to be the sum of the reflectance
and transmittance since these are the partition components that contribute to reduction
of canopy shadow. Using Eq. 3.12 that is
s T + T = 1 T . (3.42)
= (1 s ) b = T b ,
Refer to Sellers (1989) for a description of the complex two stream approximation.
Penetration of total radiation 69
1.0
a
f
0.8 h g
0.6 b
b and
9
11
0.4 10 e
8
5
3 2
c
1 6
4
0.2
d
0
0 20 40 60 80
(a) h (degrees)
1.0
a
0.8 b
d
0.6 c
d and
9
11
0.4 10
e 5
8
3 2
1 6
4
0.2
0
0 20 40 60 80
(b) h (degrees)
Fig. 3.18. Extinction coefficients: theory vs. observation. Solid lines are theoretical
neglecting scattering. Dashed lines are adjusted for estimated scattering. Data are
keyed to Table 3.7: solid points, ; open points, ; , evergreen; , deciduous. Vertical
dashed lines enclose daylight-hour average solar altitude at summer solstice for
vegetation zone: 0 60 . (a) Beam radiation. a, horizontal leaves, Eq. 3.25;
b, uniform leaf distribution, Eq. 3.28; c, vertical leaves, Eq. 3.26; d, opaque vertical
right circular cylinders, Eq. 3.29; e, opaque vertical right circular cones with beam
normal to the slant surface, Eq. 3.32; f, L = 90 h , Eq. 3.24; g, Curve (f)
incorporating estimated scattering as described in text; h, opaque hemisphere,
Eq. 3.33. (b) Diffuse radiation. a, Horizontal leaves, L t = 5, Eq. 3.34; b, uniform leaf
distribution, L t = 5, Eq. 3.34; c, vertical leaves, L t = 5, Eq. 3.34 (a c: Numerical
solution at L t = 5 (Ross, 1975, Table VII, p. 34), fitted with exponential decay in
Fig. 3.17); d, Linearly interpolated d for observed = 0.30 (cf. Fig. 3.15); e, Curve
(d) incorporating estimated scattering as described in text.
70 Radiant uxes
and
= (1 s ) d = T d . (3.44)
For mat (i.e., dull surface) leaves, scattering is independent of light direction (Ross,
1981, p. 137) and we find from Table 3.4 that for trees T = 0.51. Hence, with
scattering
0.51 . (3.45)
We now use the estimate of Eq. 3.45 to transform Curve (d) of Fig. 3.18b into Curve
(e). The resulting scattering (h ) only approximates the observations. Note also that
applying this transformation to Curve (f) of Fig 3.18a (L = 90 h ) reduces it to
an excellent fit (dashed Curve (g)) with the evergreen observations.
Note that the observations of from Table 3.7 are also plotted in Fig. 3.18 as open
symbols, one for the evergreens and five for the deciduous. Considering their values
in comparison with the s corrected for scattering (dashed curves), we get the very
interesting empirical result
. (3.46)
Equation 3.46 is based on a very limited sample and deserves further investigation;
we now assemble additional support for this finding
Additional observations
The literature contains many other estimates of ranging in value from 0.3 to 1.5
(Ross, 1975). The higher values of are associated with large, horizontal leaves and
the lower values with more vertical leaves (Ross, 1975). Without reference to their
geographical location (and hence their h ) we have collected observations of and
of the momentum decay parameter, , for various plant types in Tables 2.2, 2.3, and
2.4. As we will see in Chapter 4
mn , (3.47)
a a mnb c a
a
Observed: Tables 2.2, 2.3, and 2.4.
b
Fig. 4.14.
c
Calculated: = /(mn) = aw /(mn L t ).
Optical optimality 71
0.6
Evergreen trees
0.4
Deciduous trees
0.2
0
0 0.2 0.4 0.6 0.8 1.0
Light extinction coefficient,
where
m = exponent governing the relationship between velocity and shear stress and
having the nominal value 0.5 for the foliage elements of trees (Table 4.1), and
n = the number of sides of each foliage element experiencing drag and having
the nominal value 2 for the foliage elements of trees.
The product, mn, has the nominal value of unity for trees but for crops and grasses
1 mn 1.35 (Fig. 4.14). A typical tabulated field experiment (Tables 2.2, 2.3,
and 2.4) measured either or but not both, and did not distinguish between clear
and cloudy skies. Thus the only possible test of Eq. 3.46 is to compare the averages
of the two parameters across broad classes of vegetation and cloudiness, and to use
the above nominal values of mn to estimate from . This is done in Table 3.8 and
Fig. 3.19 and confirms Eq. 3.46 across the unknown range of cloudiness encompassed
by the data. This finding is central to the arguments of this monograph. Equation 3.46
also increases our confidence in the conclusions arising from Fig. 3.18 by effectively
doubling the sample size for either or .
The equality of Eq. 3.46 was noted by Ross (1981, p. 313) at h = /2, but its more
general applicability seems to have escaped prior notice. As we shall see in Chapters 7
and 8, this result has far-reaching physical significance. Is there selection pressure that
leads naturally toward ?
I Optical optimality
Solar radiation is the energy source for production, and as we assume here, selection
pressure is toward maximizing plant productivity. Therefore, whenever energy is the
resource limiting production it will be used optimally. As we will see in Chapter 6,
canopy water demand will adjust to limited supply by reducing the canopy cover, M,
while leaving the plant water use and productivity relatively unchanged. It is the avail-
ability of light, carbon, or nutrients that will limit plant production. In Chapter 10 we
will consider the carbon and nutrients while in this chapter we deal only with the light.
72 Radiant uxes
Plate 3.4. Excess energy loads must be shed. In tropical climates such as Jodhpur,
India (latitude = 26.5 N), leaves of this asoka tree have a mean angle of about 65
assuring that at the growing season solar altitude of over 30 (cf. Fig. 3.9) excess NIR
will be reflectively dumped. (Photograph by Peter S. Eagleson.)
To ensure optimal productive use of the available solar energy there exist at least
three simple mechanisms through which selection may operate:
(1) Maximize the productive solar energy entering the canopy. This is
accomplished by minimizing the canopy reflectance, PAR . Figure 3.7 shows
PAR to be insensitive to over the practical range of h , so the primary
mechanism for control of the reflection of PAR is the nature of the leaf surface
rather than leaf angle optics.
Larcher (1983, p. 10) suggests that a dense covering of hairs can reduce PAR reflection by a
factor of two or three.
Optical optimality 73
We will return to the important condition of incipient illumination, but first imagine
the two-dimensional simplification of specular reflection from a unit of opaque leaf
surface shown in Fig. 3.14 to be extended to a single-layer array of such leaves as
sketched in Fig. 3.20. Considering only downward scattering, the above two conditions
are met in this simplification by the geometrical arrangement sketched which, on the
daylight-hour average, puts the basal area of the crown-average leaf in full shadow
and transmits light downward to the next leaf layer through the unshadowed gap.
Thus in tandem, reflection from the adjacent leaf reduces the one-dimensional leaf
shadow to the optimal group condition
b(2) = , (3.48)
as long as h > 90 L . From the sketch of Fig. 3.20, at optimality the solar altitude
and leaf angle are related geometrically through dimensions a, b, and c according to
2b + a = 1, (3.49)
In sparse canopies (i.e., low L t ), NIR may best be shed by low which reflects to the forest floor,
while for dense canopies (i.e., high L t ) high are needed to reflect NIR back to space from the
canopy top.
Additional evidence is presented in Fig. 8.4.
This is defined as the horizontal projection of crown area on the forest floor.
Because of the low leaf reflection coefficient for PAR (cf. Fig. 3.7), we cannot rely upon multiple
reflections to supply the lowest leaf with even the compensation PAR under these average
geometrical conditions. Note however that Fig. 3.20 is for the time-averaged daylight-hour solar
altitude so that for some hours near midday, particularly near the summer solstice, deep crown
penetration of the incident radiation will occur.
Geometrical considerations show that for downward scattering, the leaf spacing drawn in
Fig. 3.20 is optimal by providing maximum insolation absorption per unit of basal area through
minimum leaf spacing at maximum energy absorption efficiency.
This weak dependence of light direction makes our argument approximately applicable also to
diffuse radiation making d = as well.
74 Radiant uxes
L
h
b
Fu
ll
1
illu
b
mi
Leaf c Leaf
nat
ion
a
h L
Fig. 3.20. Proposed optically optimal geometry for opaque leaves (crown-average and
daylight hour average conditions).
in which
sin h
a= = ; b = c tan h + L ; and
b(1) sin ( h L ) 2
c = b(1) sin L (3.50)
where, as earlier, b(1)
the one-dimensional shadow of a solitary leaf. Equations 3.49
and 3.50 give the idealized geometrical constraint on optical optimality
sin( h L ) sin h
2 sin L tan h + L + = 1, (3.51)
sin h 2 sin( h L )
which is plotted over the practical range of h as Curve (b) in Fig. 3.15.
In evaluating the mechanisms leading to Eq. 3.48 as the source of the observed h
vs. correlation, remember that the natural translucent case will include transmission,
additional specular reflection from an array of neighboring leaves, and scattering in a
variety of directions from irregularities in the reflecting surfaces.
We now postulate that through natural selection, whenever light limits productivity,
leaf surface textures and the spatial arrangements of foliage elements of closed canopies
scatter the incident light so as to put leaf surfaces in full light and the crown basal area
in full shadow and thereby to reduce to the value
= , (3.52)
while at extreme temperatures, where photosynthetic temperature limits productivity,
Eq. 3.51 is biased by a need to control heat through regulation of NIR . For open
canopies we expect > as is shown for individual leaves in Fig. 3.14.
For cases in which plant productivity is limited by carbon or nutrient supply rather than light, the
condition of Eq. 3.52 is relaxed. For cases in which excess heat is limiting plant productivity
rather than insufficient light, the selection pressure leading to Eq. 3.52 is also relaxed and will be
driven by the needs of leaf temperature control as determined by the coupled
landcanopyatmosphere heat balance. We do not consider this case here.
Optical optimality 75
(2) (3)
(1) Compensation point Iko Lt
Species (% of full sunlight a ) (from column 2 by Eq. 3.53)
a
Full sunlight is interpreted here as the average growing-season daylight-hour
insolation at canopy top. See discussion following Eq. 3.53.
Source: Baker, F.S., Principles of Silviculture, Table 12, p. 143, Copyright c
1950 McGraw-Hill Book Co. Inc., with kind permission from the McGraw-Hill
Companies.
76 Radiant uxes
Plate 3.5. Tree height often wins the battle. Elevation of the leaf surfaces above those
of competing species captures the light at the metabolic cost of additional supporting
trunk and branches. This higher elevation in the boundary layer feeds back positively
into lower canopy resistance and thus to the higher nutrient fluxes needed to support
this structure. (Photographed by Peter S. Eagleson in the Borghese Gardens, Rome.)
of a constant full sunlight. However, using the Gates (1980, p. 94) value of full
sunlight, Bakers percentages yield absolute values of Ik which are as much as two
orders of magnitude larger than those listed by Larcher (1983, Table 3.6, p. 104); for
example Ik = 18.3 klx (Baker) vs. 0.2 klx (Larcher) for the shade leaves of conifers,
and Ik = 2.15 klx (Baker) vs. 0.45 klx (Larcher) for the shade leaves of deciduous trees.
The value of Ik is known to reflect habitat light conditions (e.g., Larcher, 1983,
p. 103), thus we interpret the Baker values in column 2 of Table 3.9 as Iko Ik /I0 ,
where I0 is the daylight-hour, growing-season average, canopy-top insolation of the
particular habitat.
We conclude that solar elevation plays an important and perhaps dominating role
in selecting the leaf angle in a given climate.
Crown shape
To absorb maximum solar energy the cold-climate conifer has darkly pigmented
needles which we have just seen to be oriented on average nearly normal to the
radiation. To conserve this heat, the conifer may restrict sensible heat diffusion by
packing the needles in a dense monolayer shell at the surface of the crown. This
strategy is particularly successful in situations with large amounts of scattered light
(e.g., high latitudes, overcast skies, or snow-covered surfaces) since light can then reach
the side of the tree which is shaded by the monolayer in direct light. As a limiting case
In contrast to warm-climate conifers such as loblolly pine.
78 Radiant uxes
of this monolayer we can imagine the needles in a single planar layer coincident with
a crown surface that is perpendicular to the beam radiation. In such an idealized case
we can derive the relationships between:
(1) solar altitude and the proportions of the conical crown, and
(2) solar altitude and leaf area index.
but the (single-sided) surface area, Al , of the needles lying in and comprising the
conical surface is larger than given by Eq. 3.56. Lang (1991) gives this as
A1 = As . (3.57)
2
By definition of the leaf area index, we can now use the last two equations to write
2 1/2
A1 As hc
Lt = = 1+ . (3.58)
ro2 2 ro 2 2 ro
Jarvis et al. (1976, Table XVI, pp. 234235) give the latitudinal locations of leaf area
indices measured for various conifer species. Selecting the pines because of the many
varieties measured over a wide range of latitudes, we present the observations of L t vs.
in Fig. 3.21. Combining Eqs. 3.55 and 3.58 and using Fig. 3.9 as before to translate
h into , we can plot Curve (a) in Fig. 3.21 in comparison with the observations.
Except for the Japanese red pine the agreement is good and supports the notion that
the aspect ratio of the conical cold-climate conifers reflects the solar altitude.
For three varieties of pine (Scots, ponderosa, and lodgepole) the averaged values
of Iko reported by Baker (cf. Table 3.9, column 2) give Ik /I0 = 0.135 which makes
Eq. 3.53
1 1
L t = ln . (3.59)
0.135
Structural response to radiation regime 79
10
Pine
8
6
Lt (b)
Scots
4 Lodgepole
(a)
2 Loblolly Norway
Japanese red
0
30 40 50 60 70
Latitude, (N)
Fig. 3.21. Sensitivity of crown shape to solar altitude. Data from Jarvis et al. (1976,
Table XVI, pp. 234235). For explanation of Curves (a) and (b) see text.
With the needle surface perpendicular to the solar beam, L = 2
h , and Eq. 3.59
is, for pine
L t = 2 csc(h ), (3.60)
which, with the help of Fig. 3.9, is plotted as Curve (b) in Fig. 3.21. It serves both as
a verification of Bakers (1950) compensation light intensity (as interpreted here) for
these pine varieties, and as confirmation of the observed slight increase of leaf area
index with latitude.
4
Turbulent uxes
A Background
Under the assumptions of this work, the essential controls on forest growth are the
vertical fluxes within the canopy of light, carbon dioxide, water vapor, and heat. We
have considered the first of these in the previous chapter. Here we seek analytical
expression of the turbulent fluxes of mass and heat in terms of the canopy structure
and climate. To reach this goal we represent the three-dimensional canopy by a two-
dimensional approximation.
We begin by reviewing the turbulent transfer of any conservative fluid quantity
which does not change by virtue of its displacement in the direction of transfer. We let
s be the instantaneous amount of this quantity per unit mass of fluid, being the sum
of a temporal mean value, s , and a turbulent fluctuation therefrom, s , and we assume
that s varies in the z-direction only as is shown in Fig. 4.1.
Because of zero-mean vertical velocities, w = w , a vertical exchange of fluid pack-
ets takes place through any horizontal area, A, even though the mean fluid velocity is
in the horizontal direction, x. The vertical distance traveled by these packets before
taking on the character of their new surroundings is assumed to be a random quantity,
80
Background 81
z
Fig. 4.1. Definition sketch
for turbulent transport.
(After Eagleson, 1970,
+l +l Fig. 3-2, p. 17.)
s Area A
s, x
0
Transverse
S' motions
dm
l l
the mean value of which we shall call the mixing length, l; we will return to this
quantity later in the chapter. The amount of the quantity s carried by a parcel of mass
dm and passing through A will depend upon the point of origin of motion of the parcel.
This is due to the assumed gradient of the quantity s in the direction of the random
motion. Statistically speaking, no particle lying outside the limits +l > z > l will
pass through A, and the average time needed for the parcel to travel the distance l will
be called t. The average rate of net upward transport of the quantity s per unit area
is called Sz and may be written
0 +l
l
Sz = s(z) dz dA s(z) dz dA , (4.1)
At A l A 0
in which is the fluid mass density (g cm3 ). Expanding s(z) in a Taylor series about
z = 0, retaining only the linear terms, and integrating gives the general form of the
so-called fluxgradient relationship
l 2 ds ds
Sz = = K , (4.2)
t dz z=0 dz z=0
in which K is a turbulent transfer coefficient (cm2 s1 ) dependent upon the flow
conditions. Alternatively, we can write this vertical transport as
Sz = ws = w s . (4.3)
Introduced by Prandtl (1926) as the mixture length, l is the average distance traversed by a
fluctuating fluid element before it aquires the velocity of its new region.
82 Turbulent uxes
Finnigan and Raupach, 1987; Raupach, 1988; Raupach and Finnigan, 1988; Shuttle-
worth, 1989) on the basis of frequent and pronounced countergradient fluxes of heat,
water vapor, and carbon dioxide observed at times in certain canopies (e.g., Denmead
and Bradley, 1985, 1987). The countergradient fluxes are attributed to length scales
of the vertical mixing (i.e., mixing lengths, l) that are related to the crown dimension,
probably through eddies shed in the wake of individual crowns in a canopy with an
open top (e.g., conical crowns), and thus violate the fine grain mixing implied
in the linearization of Eq. 4.1. A more correct, higher-order formulation has been
proposed by Wilson and Shaw (1977), and by Raupach and Shaw (1982), but its com-
plexity is mismatched to the other gross simplifications of this work. In spite of the
demonstrated inadequacies of the fluxgradient approach, Raupach and Thom (1981,
p. 98) acknowledge that the fluxgradient methods have given approximate but useful
insight into the way in which physical and biological factors combine to govern the
transpiration and photosynthesis rates of a plant canopy. . ..
General insight is the goal of this work rather than precision of detail, so we will use
the fluxgradient approach with all its flaws as a suitable approximation. To reduce
the resulting error, we will consider primarily dense canopies in which l is most
likely to have the scale of the interleaf distance, and will concern ourselves with only
the long-term average state of atmospherecanopysoil systems, a state under which
(continuous) countergradient transport within the crown will be minimized.
B Introduction to momentum ux
The canopy and its substrate constitute a sink for the horizontal momentum that is
characteristic of the ambient mean airflow. Turbulent diffusion carries horizontal mo-
mentum vertically downward within the canopy and at each level the resistance offered
to this flow (i.e., the drag) of the solid canopy elements extracts a portion of that mo-
mentum and converts it into turbulence and thence into heat. This continuous supply
from the ambient atmosphere to the drag on the canopy elements defines what is called
the vertical flux of momentum. Its formulation is the beginning point of our analysis.
During the second half of the twentieth century agriculturalists and later foresters
studied momentum flux in various types of plant canopies; the literature is vast, but a
useful summary is provided by Brutsaert (1982). A mixture of theory and empiricism
has led the field to expressions of the form
u( )
= exp (aw ), (4.4)
u0
f ( )
= exp (ad ), (4.5)
0
and
K m ( )
= exp (am ), (4.6)
K m (0)
Introduction to momentum ux 83
u(z) dynamic
sublayer
( ~
~ constant)
0 (logarithmic profile)
u0 s f (0)
z0 f ()
exponential canopy
h d0 profiles sublayer
z
s assumed
hs uniform
Fig. 4.2. Definition sketch for canopy kinematics. (After Brutsaert, W.H., Evaporation
into the Atmosphere, Fig. 3.1, p. 54, Copyright
c 1982 D. Reidel Publishing Company,
with kind permission from Kluwer Academic Publishers, Dordrecht, The Netherlands.)
in which u, f , and K m are, respectively, the horizontal wind velocity, the horizontal
shear stress on the foliage, and the kinematic eddy viscosity (see Eq. 4.2), while as
before (see Fig. 4.2)
hz
. (4.7)
h hs
The reference values, u 0 , 0 , and K m (0), are those in the free stream at = 0.
The coefficients aw , ad , and am are called canopy indices or extinction parameters
and are estimated empirically to be 0.4 < aw < 4 (Cionco, 1972), ad 3 (Kaimal and
Finnigan, 1994, p. 78), and 2.2 < am < 4.25 (Brutsaert, 1982, p. 106). To distinguish
the behavior of different plant types, it is important that we understand the physical
reasons for the variation in these indices. We will attempt to derive them from first
principles of fluid dynamics, keeping in mind the cautionary words of Landsberg and
Jarvis (1973, p. 645):
df
[f + s ] Ab
d
w
hs v
u
z
y
x
Table 4.1. Discrete drag regimes for idealized foliage elements
103 RD
and all practical Rx
RD < 101 101 RD < 100 100 RD < 10 10 RD < 103 (two-dimensional
(two-dimensional (two-dimensional (two-dimensional (two-dimensional circular cylinders
circular cylinders) circular cylinders) circular cylinders) circular cylinders) and flat plates)
CD u
0 =1 = 2/3 = 1/2 = 1/4 =0
u 2
0 2 =1 2 = 4/3 2 = 3/2 2 = 7/4 2 =2
u0 m m=1 m = 3/4 m = 2/3 m = 4/7 m = 1/2
Wake separation None None Laminar Laminar Laminar
Plant regimes Inoue and Uchijima (1979) Thom (1971) (beans) Landsberg and Jarvis (1973) Amiro (1990b)
(based upon CD ) (rice and maize) (spruce shoots) (pine and aspen)
Jarvis et al. (1976)
(coniferous forest)
General formulation of momentum ux 87
In this chapter we will work primarily with closed canopies where we will assume
the boundary shear stress to be negligible. In the manner of Brutsaert (1982, p. 97),
we assume the foliage resistance to be diffusely distributed throughout the flow
field. That is, we assume there to be an infinite number of small, unconnected resistive
elements behaving as a continuum-like sink for momentum. However, for illustrative
purposes, in the definition sketches of Fig. 4.3 we have lumped this distributed resis-
tance so that the isolated fluid element contains a single leaf. The horizontal drag force,
f D , which this leaf exerts on the fluid is balanced (in the absence of both a horizontal
pressure gradient and horizontal velocity gradients) by the forces due to fluid shear
stress, zx = f + s , on the top and bottom horizontal surfaces of the fluid element.
The difference in these shearing forces, f Ab , is the increment of horizontal fluid
momentum extinguished by the solid matter per unit of time. The horizontal equation
of motion is then written in the form
df
+ f D = 0, (4.8)
dz
in which, using the meteorological convention,
horizontal component of foliage resistance CD u 2 AD (z)
fD = = . (4.9)
unit volume of canopy Ab z
In Eq. 4.9, CD is the fluid dynamic drag coefficient of the canopy solid matter,
is the air density, AD is the horizontal component of the effective drag-producing
area of the solid matter, and Ab is the cross-sectional area of the right circular cylinder
that just circumscribes the crown.
Drag coefcients
In the general case illustrated in Fig. 4.3, the drag coefficient of the foliage elements
may be influenced by one or more of three factors:
1. Shape (C D1 )
Here we are concerned with the surface and form resistance of the foliage element shape
in an infinite flow field. This is the principal component of the resistance and is given by
the familiar empirically determined drag coefficient (CD or Cf ) vs. Reynolds number
(R) curves. These as are given for flat plates (leaves) by Schlichting (1955, Fig. 21.9,
p. 448), and for two-dimensional circular cylinders (stems, twigs, and shoots) by
the generic curve found in most fluid dynamics texts (e.g., Rouse, 1946; Daily and
Harleman, 1966) and presented here as Fig. 4.4, all for steady flow and stationary
bodies. We will represent the latter curve, in the Reynolds number range of interest, by
CD0
CD1 = , (4.10)
RD
as is shown by the straight line segments in Fig. 4.4. The values of are summarized
with their Reynolds number ranges in Table 4.1.
88 Turbulent uxes
2
1
1 4
0.1
1 0 2 3 4 5 6
10 10 10 10 10 10 10 10
u( )D
RD =
2. Interference (C D2 )
Here we are interested in the fact that proximity of multiple, two-dimensional drag-
producing elements reduces the total drag on the set of elements to less than that on the
sum of individuals each governed by the infinite fluid observations of Fig. 4.4. This
effect is due to reduction in the form drag of body (a) due to the stagnation pressure
bulb upstream of body (b) in the wake of body (a). This is illustrated in Fig. 4.5
for stationary two-dimensional circular cylinders and for stationary streamlined struts
as given by Hoerner (1958). We will assume that the shape and two-dimensional
interference effects are separable so that
3. Roughness
In the case of canopy-scale non-homogeneity, the open canopy top causes a stream-
line pattern which is variable with free stream velocity. In the idealized case of two-
dimensional roughness strips illustrated in Fig. 4.6, we show the free stream tending
to follow the crown contours at low u 0 , and to skim over them at high u 0 . This causes
General formulation of momentum ux 89
0.1
In free flow
x
0 c
2 4 6 8
0.1
First strut
(b)
(a)
u0
(b)
90 Turbulent uxes
a decrease in the aerodynamic roughness of the canopy with increasing free stream
velocity and is expressed through the canopy surface resistance coefficient, Cf , that is
0 = Cf (u 0 )u 20 . (4.12)
In any particular case, one or more of these influences may be present depend-
ing upon the crown shape, foliage volume density, leaf area density, and leaf area
distribution; leafy plants may be areally dense yet volumetrically sparse.
CD (u, z) u 2 AD (z)
fD = = CD (u, z) u 2 aD (z) = f aD (z). (4.13)
Ab z
In sparse canopies where the substrate shear stress cannot be neglected the following
development is inexact.
Momentum ux in homogeneous canopies 91
Using Eqs. 4.8, 4.13, and 4.14, and noting that by definition
The variation of fluid velocity is then obtained from Eqs. 4.13 and 4.17 depending
upon the drag coefficient vs. Reynolds number relationship.
We will now carry out the detailed calculations for a number of canopies beginning
with the limiting cases posed by low volume density homogeneous canopies of
(1) leafy plants such as deciduous and needle-leafed trees (where the surfaces
have a significant horizontal projection and the drag is primarily skin-friction); using
the terminology of Horn (1971) we refer to such plants as multilayer plants due to
the layered appearance of their leaves resulting from the branching structure, and
(2) stemmy plants such as rice and wheat (where the surfaces are largely vertical
and it seems a priori that form drag may predominate).
h 12
10
Homogeneous
8 a t( )
at( ) = 1
1
z (m)
6
hs
as aL
4
2 z
0 (aL+ as )dz = L t
0
0.05 1 3 5 7 9
Lt
Fig. 4.7. Typical foliage area distribution of leafy canopy and its homogeneity
approximation. as = non-leaf area (one-sided); aL = leaf area (one-sided). (From
Rauner, J.L., Deciduous forests, in Vegetation and the Atmosphere, Vol. 2, Case
Studies, edited by J.L. Monteith, Fig. 1, p. 243; Copyright c 1976 Academic Press
Ltd, with kind permission from Academic Press Ltd, London.)
3
e
1.0
0 0.5 1.0 1.5
f ()/0
This simple exponential nature of the momentum flux decay is supported by obser-
vations of, among others, Hicks and Sheih (1977) in maize, Seginer et al. (1976) in
an artificial canopy, and Kaimal and Finnigan (1994, p. 78) in a variety of natural
and artificial canopies. The Kaimal and Finnigan (1994, Fig. 3.3b, p. 78) observations
embrace artificial canopies (rods), as well as crops (wheat, corn), and forest (eucalyp-
tus, pine), and they all cluster around an extinction parameter, ad = n L t = 3 (see
Fig. 4.8), which we take to represent the typical leafy canopy.
u fm , (4.19)
1
m= . (4.20)
2
Strictly speaking, the proportionality of Eq. 4.19 is variable with elevation even in
homogeneous canopies due to increasing with as the velocity variation leads into
a new Reynolds number regime (see Table 4.1). However, this variation is confined to
the lower part of the canopy which contributes little to the dynamics. Consequently,
we will assume the proportionality of Eq. 4.19 to be invariant and use it with Eq. 4.18
to write the general relation for horizontal windspeed
u( )
= exp (mn L t ), (4.21)
u0
where u 0 is the horizontal wind speed at the top of the canopy, z = h, or = 0. This
exponential form of the velocity extinction is confirmed by observations over a wide
range of leafy plants (e.g., Rauner, 1976, p. 256; Cionco, 1978, p. 83; and Kaimal and
Finnigan, 1994, p. 78). We reproduce the Rauner (1976, Fig. 11, p. 256) observations
here in Fig. 4.9. We will continue to separate the effects of solid density (i.e., L t ), from
those of geometry (i.e., mn) by writing Eq. 4.21
u( )
= exp ( L t ), (4.22)
u0
in which
mn. (4.23)
= . (4.24)
Thus the extinction parameters for both horizontal wind velocity and light are related
by
= mn. (4.25)
in which u(z) is the temporal mean wind velocity at elevation z, k is von Karmans
constant, u is the shear velocity defined by
1/2
u 0 / = Cf u 0 , (4.27)
where d0 is the zero-plane displacement height, z 0 is the surface roughness length, and
0 is the shear stress at z = h. Various derivations of Eq. 4.26 are given by Monin and
Yaglom (1971, pp. 274 et seq.), and a definition sketch is given here in Fig. 4.10. At
the top of the canopy Eq. 4.26 gives
u h d0
u0 = ln , z = h. (4.28)
k z0
du(z) u
= , (4.29)
dz k(z d0 )
Momentum ux in homogeneous canopies 95
z Canopy top
1
h
z0 /h d 0 /h
0.5
0 1 2 3
u
u0
du(z) u0
= . (4.30)
dz z d0
(h d0 ) ln
z0
Within the canopy sublayer we have found (Eqs. 4.7 and 4.22)
hz
u(z) = u 0 exp L t , h s z h. (4.31)
h hs
For 0 z h s we will assume the velocity constant at u(z) = u(h s ). From Eq. 4.31
this is
Finally, setting Eq. 4.30 equal to Eq. 4.33, gives, at the canopy top z = h, the
velocity gradient matching condition:
d0 d0
1 1
h ln h = 1,
Lt
hs z0
(4.34)
1 h
h
which was first presented by Kondo (1971) for the special case h s = 0. Using Eq. 4.28,
this velocity gradient matching condition can be written in the useful alternative form
u h d0
= L t. (4.35)
ku 0 h hs
In Eq. 4.35, the left-hand side is called the bulk drag parameter, while the right-hand
side is the normalized canopy drag area density, or more simply, the crown density.
Eddy viscosity
Introducing the usual assumption that the shear stress is proportional to the velocity
gradient through the kinematic eddy viscosity, K m , we have, as a special case of
Eq. 4.2:
/ (f + s )/
Km = . (4.37)
du/dz du/dz
Above the canopy, Eqs. 4.29 and 4.36 allow us to evaluate this quantity as
K m (z) = ku (z d0 ), (4.38)
while within the areally dense canopy, s is negligible, and Eqs. 4.07, 4.18, 4.27, and
4.33 give
u 2 (h h s ) hz
K m (z) = exp n(1 m) L t . (4.39)
u0 L t h hs
Since K m is defined in terms of quantities that we are matching at z = h, K m will
automatically match there also without generating an independent matching condition.
From Eq. 4.38 at z = h, K m (h) = ku (h d0 ), and with Eqs. 4.07, 4.23, 4.35, and
4.40, Eq. 4.39 becomes finally
K m ( ) (1 m)
Km
o
= exp L t . (4.40)
K m (0) m
Equation 4.40 has been found to match observed eddy viscosity decay in maize (e.g.,
Hicks and Sheih, 1977), pine forest (Lemon, 1965), rice (Uchijima, 1962), and wheat
(Denmead, 1976). For aspen, Rauner (1976, Fig. 10, p. 255) finds a variation somewhat
different from Eq. 4.40 as is shown here in Fig. 4.11. Such a variation is more like that
which we will derive later for non-homogeneous canopies.
Alternatively we may refer to K m as the eddy momentum diffusivity.
98 Turbulent uxes
Copyright c 1976
hz
6 Km (0) exp (1.5 hh ) Academic Press Ltd, with
s
kind permission from
hs Academic Press Ltd,
4
London.)
2
0
0 0.2 0.4 0.6 0.8 1.0
Km (m s1)
2
Averaging Eq. 4.40 over the depth of the homogeneous cylindrical crown, we obtain
(1 m)
1 1 exp Lt
mo K m ( ) m
K d = . (4.41)
0 K m (0) 1m
Lt
m
mo is derived for non-homogeneous hemispherical and conical crowns
The function K
in Appendix A.
Mixing length
In describing turbulent transport in fluids a length scale is needed, and to meet this need
Prandtl (1926, cf. 1952, pp. 117118) introduced the mixing length, l. Conceptually,
l is the average distance traversed by a fluctuating fluid element before it acquires
the velocity of its new region. He assumed the mixing length to be dominated by the
geometrical characteristics of the flow; namely, in the presence of a solid boundary he
took the size of the turbulent eddies to be proportional to the distance of the eddy from
that boundary. This led him to relate the shear stress to the mean velocity gradient in
turbulent flow in a manner similar to that for molecular shear but with a proportionality
coefficient dependent upon the flow features. Subscripting l with Prandtls initial we
then write
2 du du
= lP , (4.42)
dz dz
from which, using Eq. 4.38, we see that the kinematic eddy viscosity is defined by
2 du
K m = lP . (4.43)
dz
Momentum ux in homogeneous canopies 99
Using Eqs. 4.33 and 4.39, Eq. 4.43 gives, within the canopy,
u h hs 1 2m hz
lP (z) = exp Lt , z h. (4.44)
u0 Lt 2m h hs
Using Eqs. 4.29 and 4.38, Eq. 4.43 gives, above the canopy,
Von Karman (1934) approached the need for a length scale differently than did
Prandtl, basing his development upon a physical picture of the turbulent mechanisms.
He assumed the mixing length to be governed by a local turbulence structure which is
similar from point-to-point differing only in its length and time-scales and without any
geometrical reference to the source of the turbulence. This led him to the expression
for the (Karman) mixing length
du
dz
lK = k 2 .
(4.46)
d u
2
dz
With the velocity distribution of Eq. 4.31, Eq. 4.46 gives the constant mixing length
k(h h s )
lK = , z h, (4.47)
Lt
while for z > h, Eq. 4.45 again results.
Zero-plane displacement
In experiments on an artificial canopy composed of a dense array of thin vertical
rods, Thom (1971) found the zero-plane displacement to be indistinguishable from
the level of action of the drag on the component parts of the canopy. Hicks and
Sheih (1977) confirmed this finding for a maize canopy. If we accept this to be a
general result as Thom proposes, we can calculate the zero-plane displacement in
terms of the drag-producing characteristics of the canopy. Referring to Fig. 4.12, we
take moments about an axis perpendicular to the flow and in the plane of the surface to
write
h
d0 FT = z f D (z) dz, (4.48)
hs
in which
d0
z = hs
Using Eq. 4.18 with Eq. 4.14 in Eq. 4.49, we can integrate Eq. 4.48 with Eq. 4.50
for areally dense homogeneous canopies (i.e., a( ) = 1; see Eq. 4.17) to obtain
h d0 1 exp (n L t )
= , (4.51)
h hs n L t 1 exp (n L t )
which is plotted in Fig. 4.13 in comparison with observations from the literature.
Most observations in this figure do not include h s , hence we assume h s / h is negli-
gible for purposes of the comparison; this certainly accounts for some of the scatter.
Nevertheless, from the limited data available, Eq. 4.51 seems for trees to be a de-
cided improvement over the constant value of d0 / h = 2/3 (dashed line) proposed by
0.6
h d0 1 exp (nLt )
= -
h hs n Lt 1 exp (n Lt )
0.5
0.4 3 1
n Lt
Brutsaert (1982)
h d0
h hs
0.2
4
6
7
0.1 2
hs = 0
8
hs assumed zero
0
0 2 4 6 8 10
n Lt
Broadleaved plants
The characteristic dimension of the leaves of broadleaved plants is about x = 4 cm
),
(e.g., Murphy and Knoerr, 1977), and a nominal average wind speed in the canopy, u(
will be taken as 2 m s1 . The nominal drag-defining Reynolds number for broadleaved
plants is then:
)x
u( (2 m s1 )(4 102 m)
Rx = = = O(103 ).
1.5 105 m2 s1
It seems reasonable to assume that the flat leaves of broadleaved plants offer surface
resistance over at least a portion of both sides of the leaves, thus setting n = 2 in Eq. 4.25.
At Reynolds numbers Rx = O(103 ) such as given above, and with an assumed leaf
surface absolute roughness, 1 mm, it is assumed that the surface is aerodynamically
An obvious exception is the aspen (Populus tremuloides) the leaves of which vibrate rapidly in the
wind.
Homogeneous leafy canopies: M = 1 103
mn = 1. (4.52)
Needle-leaved plants
There is a wide range of drag geometries within the group of needle-leaved plants, but
we will identify and treat only the two extremes.
1. Individual-needle varieties
Most Pinus needles emerge from the stem in a tight circular bundle (called a fascicle)
of multi-sided individuals which opens progressively along the needle axis, thereby
providing a varying degree of mutual drag interference. Those varieties with very long
needles in which the individuals become separated to the point of relative independence
with respect to drag we call individual needle varieties. As is pictured by Zim and
Martin (1987), these are exemplified by the longleaf pine (Pinus palustris) and the
loblolly pine (Pinus taeda), and (to a somewhat lesser degree) include others such
as sugar pine (Pinus lambertiana), pitch pine (Pinus rigida), ponderosa pine (Pinus
ponderosa), and Eastern white pine (Pinus strobus).
There are three needles to the circular bundle of Pinus taeda (Kozlowski and
Schumacher, 1943) and of Pinus radiata (Raison et al., 1992), so each needle of
these varieties has a cross-section consisting of a 120 circular sector. The important
ratio of total needle perimeter (P) to arc-chord (Ca ) of these varieties is P/Ca = 2.36
(Raison et al., 1992). All three sides of these needles are stomated.
The cross-section of Pinus strobus needles is a 72 circular sector (i.e., five to the
bundle) with stomata only on the straight sides (Kozlowski and Schumacher, 1943), and
Scots pine (Pinus sylvestris) needles have a semicircular cross-section with stomates
on only the straight side. Over a range of needle varieties, Waring (1983) found the
average perimeter-to-chord ratio, P/Ca = 2.5.
Realizing that varying needle orientation exposes an average silhouette area for
either drag or light interception, Lang (1991) used one of the Cauchy (1832) theorems
to find P/C = where C is the average chord of the cross-section when the needle is
rotated about its longitudinal axis. Finally, Chen and Black (1992) averaged over all
needle positions in three dimensions to find P/Ca = 4.0.
Pinus needle diameters are about 2 millimeters, and assuming a nominal average
wind speed, u( ), equal to 2 m s1 , the nominal drag-defining Reynolds number for
the individual needle varieties is
)x
u( (2 m s1 )(2 103 m)
Rx = = = O(102 ).
1.5 105 m2 s1
104 Turbulent uxes
With this Reynolds number Table 4.1 gives m = 4/7, and assuming surface resistance
over the full perimeter, n = 2. Thus for the individual needle varieties, m = 4/7,
and
mn = 1. (4.54)
Finally, in conjunction with Eqs. 4.24 and 4.25, Eqs. 4.524.54 give:
for leafy and clustered needle plants: = =
. (4.55)
and for individual needle plants: = 1.15 = 1.15
This tie between the vertical extinction of solar radiation and of horizontal momen-
tum flux within the canopy was derived and verified, apparently for the first time, in
Chapter 3. We will show it to be a critical link in climatic determination of canopy
structure.
Observational support for Eq. 4.55 is obtained from our collection of field estimates
of various parameters of leafy canopies which we have attempted to convert to common
definition and have included in Tables 2.2 (evergreen forest), 2.3 (deciduous forest),
and 2.4 (crops and grasses). Wherever, in these tables, estimates of both and are
available for the same species, these values are plotted in Fig. 4.14. In these coordinates,
the slope of the line from the origin to the plotted point is the value of the product
mn. For leafy canopies and for spruce, the limited data clearly cluster around the line
Homogeneous leafy canopies: M = 1 105
2
n=
s)
rod
1
hin
at)
m=
he
(T
(W
1.2
Sunflower
)
ze
y
1.0
mm
ns ai
ea (M
Ste
)
2
=
(B
n
ne
Pi
0.8
2
1/
=
m
Wheat
Rice y
af
Grass
Le
0.6 Maize
0.2
Vertical
rods
0
0 0.2 0.4 0.6 0.8 1.0
or
mn = 1 which represents both broad-leaved and clustered needle trees. For the various
Pinus varieties given in Table 2.2, the averaged data give mn = 1.14 which agrees
well with that for individual needled plants.
Other data points on Fig. 4.14 are for those crops in which there is a significant
component of canopy drag from the vertical stems which do not obey the relation
= . We will return to these stemmy plants shortly.
For very low Reynolds numbers, CD u 1 (Lamb, 1932, p. 616), whereupon
m = 1 (see Table 4.1), and with n = 2, mn = 2. This is the upper limit of mn shown
by the upper dashed line in Fig. 4.14. The lower limit of mn = 1/2 is given by
high Reynolds number flow with m = 1/2 over ground-hugging leafy plants (such as
plantain) for which n = 1. It is shown as the lower dashed line in Fig. 4.14.
Mixing length
Considering its conceptual definition as the average distance traversed by a fluctu-
ating fluid element before it acquires the velocity of its new region, we expect the
Prandtl mixing length to be related to the vertical spacing of the solid elements in
the canopy. Thus for the uniform vertical spacing of the solid elements accompanying
the homogeneous leafy canopies being considered here, it is reasonable to assume that
within the crown
lP = constant, h s z h, (4.56)
106 Turbulent uxes
as was proposed by Inoue (1963). From Eq. 4.44, this can be seen to require
m = 1/2, (4.57)
as was reasoned above from classical CD vs. R behavior. With the assumption that
n = 2, our earlier observational evidence that mn = 1 for leafy trees is a demonstration
that the mixing length really is constant within these canopies. From Eq. 4.44 this
constant value within the crown is
u h hs
lP = , h s z h. (4.58)
u0 Lt
When the eddy-generating solid surfaces are parallel to the flow as for the leafy
plants now under discussion, the Prandtl and von Karman definitions of mixing length
are equivalent, and we equate Eqs. 4.47 and 4.58 to obtain the theoretical value of the
bulk drag parameter
u
= 1, (4.59)
ku 0
which can be shown to require a reversal in curvature of the velocity profile at = 0.
The bulk drag parameter was suggested as a constant for tall, dense plant communi-
ties by Kondo (1971) and has been shown empirically by Kondo (1972) to have the
value of unity through collected observations in pine and Japanese larch canopies. The
generally accepted value of the Karman constant is k = 0.4 from which Eqs. 4.12,
4.27, and 4.59 imply
u 1/2
= Cf 0.4. (4.60)
u0
Kondo and Akashi (1976, Fig. 11) show, both analytically and from wind tun-
nel experiments, that Eq. 4.60 is satisfied in an areally dense canopy. Together,
Eqs. 4.584.60 confirm our major initial assumption (see discussion at end of
Section 4A):
mixing length lP
= = k < 1, z h.
average vertical interleaf distance (h h s )/( L t )
Using Eqs. 4.54 and 4.59 in Eq. 4.35, we have for areally dense, homogeneous,
leafy canopies
h d0 h d0
Lt Lt = 1, (4.61)
h hs h hs
or
d0 z0
1 = 2.718 . (4.63)
h h
In Fig. 4.15, Eq. 4.63 is compared with observations compiled by Jarvis et al. (1976,
Table V, pp. 188189). When it is realized that the plotted points numbered 10 to 14
are for trees with conical crowns and thus non-homogeneous canopies, we can say
that Eq. 4.63 represents homogeneous canopies quite satisfactorily. Note that Eq. 4.63
0.3
Range
Homogeneous canopy
u 1
* >1 h
0.2 ku0 11 ln =1
0
h
z0
10 Paeschke (1937), grasses, beets
h
0.1 1 12
8
13 2
u 7
* <1
0 3 9
4
= Non-homogeneous canopy 14
6 5
0
0.6 0.7 0.8 0.9
d0
h
Fig. 4.15. Boundary layer parameters for needle-leaved canopies (assumed
(z = h s ) = 0).
Number Common Name h (meters) Lt z0 / h d0 / h
Data from Jarvis, P.G., G.B. James and J.J. Landsberg, Coniferous forest, in
Vegetation and the Atmosphere, Vol. 2, Case Studies, edited by J.L. Monteith,
Table V, pp. 188189; Copyright c 1976 Academic Press Ltd, with kind
permission from Academic Press Ltd, London.
108 Turbulent uxes
follows directly from the turbulent velocity distribution, Eq. 4.28, as soon as Eq. 4.59
is established.
3.0
1.0
Normal pressure drag
1/2 Skin-friction drag (difference of above)
4R
Skin-friction drag
(boundary layer theory)
0
1 2 3 4 5
Log R
aw = mn L t = 3.64. (4.64)
Assuming the rods to be tall enough and their density to be high enough that no
significant shear reaches the substrate, we can consider the individual rods to be foliage
110 Turbulent uxes
Canopy top
10
mn L t = 1.63 h = 19 cm
Lt = 2 D = 2.38 mm
3 cm centers
0
1 2 5 10 15
(b) 1
Mean velocity, U (m s )
elements in a dense canopy rather than entire plants in a sparse canopy. The basal area
is then given by the rod spacing rather than by its diameter, and the single-sided foliage
area index is
Dh/2
Lt = = (0.1)(14.3) = 4.50, (4.65)
0.5 cm2
and for vertical cylindrical friction surfaces in a horizontal wind
2
= . (4.66)
Equations 4.644.66 give mn = 1.27, and with no separation, n = 2; thus m = 0.73.
Homogeneous stemmy canopies 111
aw = mn L t = 1.63. (4.67)
With D and h in centimeters, the single-sided foliage area index for these rods is
Dh/2
Lt = = (0.238)(19)/9
= . (4.68)
4.5 2
Once again = 2 , and n = 2, giving m = 0.82.
Averaging the results of the two studies, we have for arrays of vertical cylindrical
rods
which is almost identical to the value m = 0.75 (Table 4.1) found on the basis of
drag measurements for rice and maize. We conclude that m = 0.75 and n = 2 are the
nominal values for thin, vertical rods and hence for the stems of stemmy plants. Their
product, mn = 1.5, is drawn on Fig. 4.14, where along with the line mn = 1 for leafy
plants, it provides a limit for those plants whose drag has significant components from
both stems and leaves.
Stemmy crops
We can now use the above bounding relations:
and
to estimate mn for stemmy crops where both stems and leaves are important to drag.
The relative proportion of the one-sided area of leaves and stems for wheat, maize,
and beans are listed in Table 2.1. These can be used to estimate the average value of
mn by simple proportion
%stems %leaves
mn = (mn)stems + (mn)leaves . (4.72)
100 100
Stems are assumed to be photosynthetically active.
112 Turbulent uxes
1. Wheat
The relative foliage areas for wheat (Table 2.1) are 74% stems and 26% leaves. From
Eq. 4.72
3 1
mn = 0.74 2 + 0.26 2 = 1.37, (4.73)
4 2
which is plotted on Fig. 4.14. The values of and obtained from the literature
(Table 2.4) for wheat and for sunflower, rice, and grasses as well agree fairly well with
Eq. 4.73.
2. Maize
The average relative foliage areas for maize (Table 2.1) are 24% stems and 76% leaves.
From Eq. 4.72
3 1
mn = 0.24 2 + 0.76 2 = 1.12, (4.74)
2 2
which is also plotted on Fig. 4.14. The values of and obtained from the literature
(Table 2.5) agree very well with Eq. 4.74.
Ross (1981, Fig. 28, p. 103) presents the probability density function for leaf in-
clination angle, L = cos1 (), for various elevations in the maize canopy and at two
stages of its growth. From these we conclude that for maize
0.5. (4.75)
The foliage area of maize is uniformly distributed over the typical mature plant height
(see Fig. 4.19 taken from Allen et al., 1964,) and thus our earlier analysis of homo-
geneous canopies is again applicable. However, the foliage area index, L t , is highly
variable both varietally and seasonally; the observations collected in Table 2.4 show
1.05 < L t < 5.6 with an average value of L t = 3.72. These parameters allow us to
4
Cumulative Foliage
Observed
Area Index
2 at ( ) = 1
0
100 200 300
Depth (cm)
Fig. 4.19. Homogeneity of typical maize canopy. (From Allen, L.H., C.S. Yocum and
E.R. Lemon, Photosynthesis under field conditions, VII, Radient energy exchanges
within a corn crop canopy and implications in water use efficiency, Fig. 5, p. 255,
Agronomy Journal, 56; Copyright c 1964 American Society of Agronomy; with kind
permission from the American Society of Agronomy, Madison, WI.)
Non-homogeneous leafy canopies: M = 1 113
estimate the average extinction coefficients for maize in the manner we did for mn in
Eq. 4.72:
which can be compared with the value, ad = 5.6, reported by Uchijima and Wright
(1964) from drag experiments on a particular maize crop. Note that the range of
observed variation in L t gives a range of estimated ad , 1.30 < ad < 6.94, which easily
includes Uchijima and Wrights observation.
which is somewhat smaller than the value, am = 2.88, determined by Wright and
Brown (1967) from experimentally based momentum balance and energy balance anal-
yses. In this case however, the range of observed variation in the leaf area index gives
an estimated range 0.53 < am < 2.80 which does not quite include the observation of
Wright and Brown (1967) unless the uncertainty in (for the leaves) is also included.
We conclude that with reasonable estimates for and for L t along with the pro-
portion of stem area to total area, we can make credible estimates of the extinction
coefficients for stemmy crops.
compensation light intensity at the lowest leaf for all radii, the crown foliage area
density at any elevation must increase with radial distance from the trunk. To compare
this reasoning with observation we refer to Fig. 4.20b where the plotted points are
observations from the upper, photosynthetically active region of Sitka spruce crowns.
At the base of this region the leaf area density is represented fairly well by a homoge-
neous crown foliage area density. Light transmitted below this elevation (z > 8 m) is
used to support thinning foliage in either an understory, or on dying lower branches as
is shown in Fig. 4.20a. The foliage requirements of compensation light intensity are
satisfied fairly well in this region (again see Fig. 4.20b). Therefore, for the purposes
of estimating canopy resistance for tapered crowns we not only assume the crown
to be homogeneous in as we did in Chapter 3, but also we take the crown foliage
11
10
z (m)
8
Assumed
7 inactive
6
0 1 2 3
a (z) (m1)
L
(a) Leaf area spatial density
1.0
0.8
0.6
0.4
0.5
1 1
0.4 n
0.2 row
al c
conic
s
eou
gen
mo
Ho
0.1
0.1 1 10
(b) a L()
Fig. 4.20. Non-homogeneous Sitka spruce canopy. Observed distribution of leaf area
density. (From Landsberg, J.J., and P.G. Jarvis, A numerical investigation of the
momentum balance of a spruce forest, Fig. 1, p. 648, Journal of Applied Ecology, 10;
Copyright c 1973 Blackwell Science Ltd, with kind permission from Blackwell
Science Ltd. (b) Fitted leaf area distribution. (Data from Landsberg and Jarvis, 1973.)
Non-homogeneous leafy canopies: M = 1 115
AD (z)
aD (z) = n c
os L at ( ) n L t a( ), (4.79)
Ab z
where a( ) is again the homogeneity function which is equal to unity in the ho-
mogeneous case. Here we drop the restriction that the canopies be areally dense in
recognition of the fact that non-homogeneity of the foliage area density is a common
cause of higher velocity penetration of the canopy and hence of significant substrate
shear. We incorporate the substrate shear only to the first approximation by making
the approximation
f ( ) f ( )
, (4.80)
0 f (0)
which, in conjunction with Eq. 4.17, introduces the total downward flux of momentum
in the manner
f ( )
= exp n L t a( ) d( ) . (4.81)
0 0
We will consider only one non-homogeneous case, conical crowns such as are
displayed by many conifers (Landsberg and Jarvis, 1973), and which are non-
homogeneous both because a( ) = 1 (cf. Eq. 4.17), and because their crown shape
does not fill the canopy space.
foliage area density, at ( ) and will assume the two are equal. The observed average
leaf area density at each elevation is fitted through the canopy in Fig. 4.20b by the
solid line
confirming by its departure from unity that the canopy has a non-homogeneous leaf
area density, and by the departure of the exponent from 2 that the crown is non-
homogeneous as well (cf. Appendix A).
No information is available on the volume density of the other foliage elements. We
will assume it to be low and that the momentum absorption is as for the leafy plants
studied so far.
Momentum ux
Using Eq. 4.85 in Eq. 4.17 yields the momentum flux variation
f ( )
= exp (n L t 3.5 ). (4.86)
0
With the high degree of crown-scale non-homogeneity evidenced in Fig. 4.20 and in the
resulting Eq. 4.85, we expect the foliage element drag coefficient to vary with elevation
due to changing foliage geometry and density (i.e., changing interference), as well as
to changing Reynolds number. The true form of the interference function, s( ), is
probably quite complicated and is certainly unclear. However, expressing interference
by the exponential of Eq. 4.11, s( ) must be linear in in order to match velocity
gradients at the canopy top. To maintain reasonable agreement with Fig. 4.05a, we
assume the simple linear form
s( ) = B, (4.87)
and use Eqs. 4.104.13 to relate the canopy wind velocity and the foliage element
drag according to
f ( ) u 2 exp (B ), (4.88)
The difference is on the order of 10% for most trees as we see in Table 2.1.
Non-homogeneous leafy canopies: M = 1 117
u( ) fm exp (m B ). (4.89)
u( )
= exp [ L t ( 3.5 B0 )]. (4.90)
u0
Estimation of coefcients
Using Eq. 4.7, we differentiate Eq. 4.90 to obtain
du u0
= (3.5 L t 2.5 + B0 L t ) exp[ L t ( 3.5 B0 )], (4.91)
dz h hs
which at the canopy top ( = 0) is
du u0
= (B0 L t ). (4.92)
dz = 0 h hs
Using Eq. 4.92 with Eq. 4.29 (at z = h), we match the velocity gradients to get
u
ku
B0 = 0 . (4.93)
h d0
Lt
h hs
To evaluate B0 we need estimates of mn, , and the bulk drag parameter, u /ku 0 .
We have seen in Fig. 4.14 that for leafy plants (both needle-leaved and broadleaved),
m = 1/2 and n = 2, while at the latitude of Fetteresso (about 57 N), Figs. 3.9 and
3.18 give 0.50. With these parameters
L t mn L t = 1.66, (4.94)
h d0
= 0.53. (4.95)
h hs
We estimate the bulk drag parameter, u /ku 0 , by taking its average value over the
canopy depth under the assumption that the von Karman and Prandtl mixing lengths
are equal at each depth. Using Eqs. 4.42, 4.86, and 4.91, the dimensionless Prandtl
mixing length is
lP u (k) exp ( L t B0 )
= , (4.96)
h hs ku 0 |3.5 L t 2.5 + B0 L t |
118 Turbulent uxes
while from Eqs. 4.46 and 4.91, the dimensionless von Karman mixing length is
lK 3.5 L t 2.5 + B0 L t
= k . (4.97)
h hs (3.5 L t + B0 L t ) 8.75 L t
2.5 2 1.5
Equating these two mixing lengths and using Eq. 4.93 gives
h hs u
exp
u h d0 ku 0
= . (4.98)
ku 0
8.75 L t 1.5
1 2
h hs u
3.5 L t 2.5 +
h d0 ku 0
We solve Eq. 4.98 along with Eqs. 4.94 and 4.95 for u /ku 0 at various values of
to get the function shown in Fig. 4.21. From this figure, the average of the bulk drag
parameter over the crown depth is
!
u
= 0.89, (4.99)
ku 0
which with Eqs. 4.93, 4.94 and 4.95 gives
B0 = 1. (4.100)
0.4
0.6
Sitka spruce
0.8
1.0
0 0.2 0.4 0.6 0.8 1.0
u
*
k u0
Non-homogeneous leafy canopies: M = 1 119
2.4 1.23
3.19
4.26
2.0
Decreasing effective
u() roughness with
ln
u(0) 1.6 increasing wind speed
0.8
0.4 3.5
Lt
0
0 0.2 0.4 0.6 0.8 1.0
Eddy viscosity
With the definition of Eq. 4.37 and using Eqs. 4.40, 4.86, and 4.90, the theoretical
spatial variation of the atmospheric eddy viscosity within the crown is
u 1 m 3.5
exp L t B0
K m ( ) ku m
K mo = 0 . (4.101)
K m (0) h d0 1 + 3.5 L t 2.5
Lt
h hs
With m = 1/2, and using Eqs. 4.93, 4.94 and 4.100, this particular case becomes
which is plotted in Fig. 4.23 in comparison with values calculated by Landsberg and
Jarvis (1973) using observed velocities and laboratory-determined drag coefficients.
The agreement is excellent lending strong support to the observation-free method of
analysis introduced here. We have confidence now in our ability to estimate the effect
of crown shape on the eddy viscosity distribution. The shape of the eddy viscosity dis-
tribution for this particular non-homogeneous conical crown is qualitatively different
from the exponential decay found for the homogeneous circular cylinder in Eq. 4.40.
120 Turbulent uxes
Sitka Spruce
0.2
0.4
0.6
02
4.1
Eq.
0.8
1.0
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
o K m( )
Km =
K m(0)
Fig. 4.23. Test of theoretical eddy viscocity distribution for Sitka spruce. (Data from
Landsberg and Jarvis, 1973 for u 0 = 1 m s1 .)
h = 118 cm
h s = 18 cm
d0 = 88.5 cm
z 0 = 7.67 cm
L t = 3.55.
Unpublished analysis of the bean data by the author shows, in a manner similar to
that used above for the Sitka spruce, that = 0.53, ku u
0
= 0.59, and B0 = 0.20.
However, in this case our predicted eddy viscosity distribution is a poor representation
of that found from observations.
Strictly speaking this is an infeasible foliage arrangement since it precludes satisfaction of the
compensation light intensity for all radii at the crown base.
Non-homogeneous leafy canopies: M = 1 121
Avg. = 0.52
multilayer crowns. (b)
Homogeneous monolayer
Avg. = 0.88
0.4
crowns.
0
Cyl.:
1
0.6 =1
Cone:
0.8
1.0
0 0.2 0.4 0.6 0.8 1.0 1.2
o K m( )
Km =
(a) K m(0)
0
Low density
Avg. = 0.52
Avg. = 0.59
0.2
Avg. = 0.70
0.4
High density
0.6 cone
Lt = 1.5
Hemisphere:
0.8
Cone:
Cyl.:
1.0
0 0.2 0.4 0.6 0.8 1.0 1.2
o K m ( )
Km =
(b) K m (0)
in Eq. 4.41 where it is evaluated for the homogeneous right circular cylindrical crown.
In Appendix A we present the numerically integrated results for the cone and hemi-
spherical segment in Eqs. A.25A.27. They are compared graphically in Fig. 4.25a
where we see again that for the homogeneous canopy of cylindrical crowns the
average eddy viscosity declines with increasing horizontal leaf area, L t , while for
the non-homogeneous canopies of conical and hemispherical crowns the average eddy
viscosity increases with L t .
Generally speaking, non-homogeneities which produce open canopy tops ad-
mit higher wind velocities at a given depth than do homogeneous canopies or
non-homogeneous canopies which become less dense with depth. The open canopy
tops produce canopy-average eddy viscosities which increase with the degree of non-
homogeneity at a given L t (greater than some minimum), and which increase with
increasing L t for a given non-homogeneous crown shape. This is in marked con-
trast to the canopy-average eddy viscosities for homogeneous cylindrical crowns
which decrease with increasing L t . In the following chapter we will show that these
Non-homogeneous leafy canopies: M = 1 123
o >> 1
K m 0.8
Hemisphere
0.6
0.4
Cylinder
0.2
0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
(a) Lt
1.2
Cone and multilayer
hemispherical segment
1.0 with >>1
Low density
0.8
Hemisphere
o
K m 0.6
cone
High density
0.4
Cylinder
0.2
0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
(b) L t
differences in turbulent transport within the canopy may help to explain the observed
prevalence of certain crown shapes in particular climatic regimes. Suppose for exam-
ple that survival favors those trees that have the highest ability to transfer heat and
mass into and/or out of the canopy. Then Fig. 4.25a suggests that trees with low L t
should be homogeneous while those with high L t should be non-homogeneous. The
range 0.75 < L t < 1.0 would seem to be more or less indifferent to crown shape.
Of course the reality is not this simple as we will see!
the closeness of their foliage elements brings about a fundamental change in the dom-
inant mechanism of momentum absorption by the canopy, and hence in the sensitivity
of the eddy viscosity to increasing leaf area.
Definition sketches for monolayer hemispherical and conical crowns are given in
Appendix A along with derivations of the distribution and spatial average of the eddy
viscosity for the cylinder, cone, and hemispherical segment. Graphical comparison of
the crown shapes is presented in Figs. 4.24b and 4.25b, while multilayer and monolayer
are compared between Figs. 4.24 and 4.25. Distributions of monolayer eddy viscosity
(cf. Fig. 4.24b) are qualitatively similar to those of the multilayer but less variable over
the crown depth.
The monolayer values of K mo (cf. Fig. 4.25b) display only quantitative differences
with the multilayer curves (cf. Fig. 4.25a); sensitivity to L t is reduced in the monolayer
case. The magnitude of K mo for the high-density conical monolayer is the smallest of all
the shapes tested. It seems clear from this finding that the occurrence of high-density
conical monolayers in nature is not due to any capability for high vertical transport
rates. On the contrary, it is likely that they have adapted to situations in which low
transport, probably of sensible heat, is the key to survival.
H Drag partition
Raupach (1992) presented an analysis of the drag on rough surfaces composed of
differing densities of surface-mounted roughness elements, on the one hand, and of
natural forests of differing canopy cover, M, on the other; his results are of great interest
here. In Fig. 4.26a we reproduce his findings for k times the bulk drag parameter as a
function of the roughness density or frontal area index, , defined as
bh 4M
= = , (4.106)
D2 (b/ h)
in which D is the roughness element spacing, while b and h are the roughness element
breadth and height respectively. In this figure, the data (closed circles) of Raupach
et al. (1980) and (open circles) of OLoughlin (1965) are for geometrical roughness
elements in a wind tunnel, while the data (open triangles) of Garratt (1977), Jarvis
et al. (1976), and Raupach et al. (1991) are for natural vegetation canopies and for
vegetation canopies modeled in a wind tunnel.
The Raupach (1992) work deals with what we have called canopy-scale non-
homogeneity due to open stand structure. For small roughness density, the drag
1/2
parameter is small, approaching its minimum value, uu 0 Cf = 0.055, or
u
= 0.137, (4.107)
ku 0
which represents the drag on the average rough substrate as determined empirically
(Raupach, 1992). For large roughness density (i.e., M = 1), this drag parameter
Drag partition 125
Sitka
k u0
spruce
*
0.8
u
Bulk drag parameter
0.6 Beans
Substrate
drag
0.4
Foliage drag
0.2
Forest substrate
(Raupach, 1992)
0
0 0.2 0.4 0.6 0.8 1.0
h d0
Crown density Lt
(b) h hs
and putting Eq. 4.109 back into the left-hand side of Eq. 4.96 gives the second approx-
imation
u h d0
Lt . (4.110)
ku 0 h hs
Lt
Remembering that hh s
represents the average density of drag-producing foliage area
in the crown (m1 ), and using the quantity, h d0 , as a normalizing factor, the right-
hand side of Eq. 4.110 is a dimensionless measure of canopy sparseness that we call
crown density. For closed but sparse canopies crown density apparently governs the
partition of drag between the foliage elements and the substrate as the roughness
density, , does for the dense but open canopies studied by Raupach (1992).
We test this conclusion by comparing our limited data with Eq. 4.110 as is shown
in Fig. 4.26b. Equation 4.110 was derived from relations containing only the foliage
drag (i.e., with substrate shear neglected), so when the crown density is zero, Eq. 4.110
shows the bulk drag parameter to be zero. At the other end of the scale, we showed
for homogeneous, leafy canopies with zero substrate shear stress, that the bulk drag
parameter was unity. The dashed line in Fig. 4.26b connects these two points. To get
the bulk drag parameter for the case in which substrate shear contributes to the total
Plate 4.4. Dense branching structure. Closeup view of the branching structure of a
mulberry tree. (Photographed by William D. Rich in Arnold Arboretum, Boston;
Copyright c 2001 William D. Rich.)
Flux of mass and of sensible heat 127
drag, we must add the value given by Eq. 4.107 at the low-density end, and add nothing
at the high-density end. We will assume the variation of the substrate component to
be linear between these two extremes.
The field observations for Sitka spruce and for beans (Thom, 1971) validate the
above reasoning. From these observations we note also that the proportionality coef-
ficient,
u
ku 0
" , (4.111)
h d0
Lt
h hs
apparently will be about unity (dashed line in Fig. 4.26b) over the practical range of
crown density.
density. However, for the purposes of this book, which concern optimal forest structures
in a given resource environment, we need to estimate the vertical fluxes of mass (water
vapor and carbon dioxide) and of sensible heat. These are given by the following
analogous variants of the general fluxgradient relation derived in Eq. 4.2:
water vapor
dqv
E = Kv , (4.115)
dz
carbon dioxide
dc
Q = Kc , (4.116)
dz
in which E and Q are the vertical flux densities (mass area1 time1 ) of water vapor
(i.e., evaporation) and carbon dioxide respectively; is the moist air mass per unit
volume; K v and K c are the respective eddy diffusivities (cm2 s1 ), and qv and c are
the masses of water and carbon dioxide per unit mass of moist air respectively. The
quantity qv is commonly called the specific humidity.
sensible heat
dT
H = cp K h , (4.117)
dz
in which H is the vertical flux density (cal cm2 s1 ) of heat, cp is the specific heat
of air at constant pressure (cal g1 K1 ), T is the temperature (K), and K h is the eddy
diffusivity for heat (cm2 s1 ).
It has been shown by Monin and Obukhov (1954), and verified since by many others
(see Brutsaert, 1982, for a review), that
Kv
= a1 (1 b1 Ri ), (4.118)
Km
and
Kh
= a2 (1 b2 Ri ), (4.119)
Km
in which a1 and a2 are approximately unity, b1 and b2 are of order 102 , Ri is the
Richardson number representing the relative importance of buoyancy forces, and
the + or sign designates the direction of the buoyant instability. The Richardson
number is defined by
dT
g
Ri dz 2 . (4.120)
du
T
dz
in which g is the gravitational constant (m s2 ).
Flux of mass and of sensible heat 129
It is consistent with the accuracy of this work to endow our climates with atmo-
spheres which are, on the average, neutrally stable with dT /dz 0. This eliminates
consideration of buoyancy (i.e., thermal convection) and justifies our invoking the
Reynolds analogy which states
Kh = Km = Kv = Kc. (4.121)
Among the many observational confirmations of Eq. 4.121, Wright and Brown (1967)
show that K h = K m in maize, Denmead and McIlroy (1970) demonstrate that K h = K v
in wheat, and Denmead (1976) shows that K h = K v = K m in temperate cereals under
near-neutral stability conditions. Of course there are ample demonstrations that
Eq. 4.121 breaks down in the presence of instability (e.g. Daily and Harleman, 1966;
Businger et al., 1971; Pruitt et al. 1973).
Using Eq. 4.121, flux estimation now depends upon our ability to estimate the
relevant concentration gradient. In the case of evaporation, it is common practice to
circumvent this difficult task through joint use of the fluxgradient and thermal energy
conservation relations to derive the rate of evaporation from a saturated surface for
use as a reference maximum (i.e., potential) rate. The details of this calculation are
given here from two points of view: (1) In the next section we use the traditional flux
gradient approach introduced by Penman (1948), refined by Van Bavel (1966), and
common to the literature of hydrology; (2) in Chapter 5 we revisit this question from
the statespace (i.e., vapor pressure vs. temperature) viewpoint pioneered by Monteith
(1965) and most familiar to plant scientists.
in which
de
E = 0.622K v . (4.123)
p dz
0 = u 2 , (4.124)
u 2
Km = . (4.125)
du/dz
K v u 2 (e1 e2 )
E = 0.622 , (4.126)
K m p (u 2 u 1 )
E = Bo (e1 e2 ), (4.127)
Evaporation from a saturated surface 131
in which
K v k 2 u2
Bo = 0.622 . (4.128)
Km p 2 z 2 d0
ln
z0
The flux of latent heat of vaporization (i.e., the heat equivalent of the evaporation
rate) is written
in which
Q v = Bo (e1 e2 ). (4.130)
In a similar manner we use Eqs. 4.117, 4.123, and 4.129 to write the Bowen Ratio,
Rb , as
H Ts T2
Rb = o , (4.131)
Qv es e2
which assumes saturation conditions (subscript s) at elevation 1 and defines
cp K h
o p (4.132)
0.622 K v
as the psychrometric constant. Equations 4.130 and 4.131 then give
We now introduce the slope of the saturation vapor pressure vs. temperature curve
es es2
, (4.134)
Ts T2
which allows us to eliminate the troublesome need to know the surface temperature,
Ts . Equation 4.133 then becomes
o o o
H= Bo (es es2 ) = Bo (es e2 ) Bo (es2 e2 ), (4.135)
or
o o
H= Q v Bo (es2 e2 ). (4.136)
Neglecting horizontal flux divergences, storages within canopy and substrate, and
biological metabolism, conservation of energy gives
Q v = Rn H, (4.137)
These assumptions are discussed in Chapter 5.
132 Turbulent uxes
Remember that the units of E are g cm2 s1 while the conventional measure of
evapotranspiration rate, E T , has the units cm s1 . To make the conversion we use the
standard unit mass density of fresh water, w (g cm3 ) = 1, to write
Using Eqs. 4.129 and 4.137, Eq. 4.136 becomes the standard combination form (i.e.,
diffusion and energy balance) of the Penman equation for evaporation from a saturated
surface
Rn + o Bo (es2 e2 )
E T = . (4.138)
+ o
in which the first term on the right-hand side is the so-called available energy forcing,
and the second term is the air dryness forcing (Kim and Entekhabi, 1997). Elevation
2 is usually taken at screen height (normally 2 m) above the evaporation surface,
making u 2 u 0 , the free-stream velocity, and z 2 = h + 2, in Eq. 4.141. Note that
when h 2 m, Eq. 4.62 reduces Eq. 4.141 to the estimator
ra (k 2 u 0 )1 . (4.143)
134 Turbulent uxes
ro rs + rla . (4.145)
In addition, for homogeneous multilayers at least, the leaf boundary layer resistance,
rla , is small when compared with the stomatal resistance (Gates, 1980, p. 329), at least
when the wind speed u 0 > 2 m s1 (Larcher, 1983, p. 224). Then Eq. 4.145 becomes
ro rs , (4.146)
Transpiration from saturated stomatal surfaces 135
H 2O
rla
rcu Light
Upper epidermis
Palisade cells rr
Spongy cells
ric
rs
Lower epidermis
rc
rla
Leaf Free
cells atmosphere
(b) rr ric rs
Fig. 4.27. Stomatal cavity resistances during transpiration. (a) The structure of a
stomatal cavity. (From Larcher, W., Physiological Plant Ecology, corrected printing of
2nd Edition, Fig. 3.11, p. 88, Copyright
c 1983 Springer-Verlag GmbH & Co. KG,
with kind permission from both Springer-Verlag GmbH & Co. KG, Heidelberg, and
W. Larcher. (b) The equivalent resistance of a stomatal cavity.
136 Turbulent uxes
and we see that in this common case the total evaporative resistance of the stomatal
cavity is approximated by the stomatal resistance alone.
To transfer moisture from the leaf boundary layer out to the ambient free stream
requires the equivalent atmospheric resistance, ra , to be placed in series with ro .
while including the leaf boundary layer resistance gives the equivalent leaf resistance,
rl , as
Once again, to transfer moisture from the leaf boundary layer out to screen height in
the ambient atmosphere requires that we add the equivalent atmospheric resistance,
ra , in series with rla as is shown in Fig. 4.28.
Typical measured values of the important stomatal and leaf properties are given in
Tables 4.2, and 4.3 for various vegetation classes as taken from the literature (Larcher,
1983; Gates, 1980; Kozlowski, 1968). For the moment consider Table 4.3 where we
see that to within an average of 15% we may simplify Eq. 4.148 to
rl rls . (4.149)
ra
rls rla
Leaf
Transpiration from a dry leaf surface 137
a
Open stomata.
b
As modified in Larcher (1995).
Source: Larcher, W., Physiological Plant Ecology, corrected printing
of 2nd Edition, Table 3.2, p. 91; Copyright
c 1983, Springer-Verlag
GmbH & Co. KG, with kind permission from Springer-Verlag GmbH
& Co. KG, Heidelberg, and from W. Larcher.
Table 4.3. Typical tree leaf properties
Minimum rls c
Leaves per Single leaf 104 Stomata (s m1 )
a
Species h h s (m) crown,a n l area,a Al (cm2 ) per leaf b (per unit area) rcu d (s m1 ) rla d (s m1 )
a b
Larcher (1983, Table 5.4, p. 227). Using 104 stomata per cm2 (Larcher, 1995, Table 2.2, p. 81).
c
Rutter (1968, Table XIV, p. 65) (from ratio of vapor pressure gradient to transpiration rate except where noted), all values are per unit of leaf area.
d
Holmgren et al. (1965), as given by Gates (1980, Table 10.2, p. 329).
e
From geometric measurement (cf. Gates, 1980, pp. 333335).
Transpiration from a dry forest canopy 139
where the internal resistance, rcs = equivalent canopy stomatal resistance, represents
the effect of variable plant stress in all the leaves of the crown, and the external
resistance, rci = equivalent interleaf resistance, represents the effective labyrinthine
atmospheric pathway for fluxes through the crown. The pathway on the right-hand
McNaughton and Jarvis (1983) write E v = #E r + (1 #)E a in which E r depends mainly upon
radiation and E a upon the atmosphere. Their decoupling factor, #, is controlled by the resistance
ratio.
140 Turbulent uxes
u0 , T0
Screen height
above canopy top
ra 2m ra
Sensible
(Mass) heat
(rrcci ) r
CH
(rrcs )
rc Tl
Fig. 4.29. Big leaf model of dry canopy. (Adapted from Dickinson, R.E., A.
Henderson-Sellers, C. Rosenzweig and P.J. Sellers, Evapotranspiration models with
canopy resistance for use in climate models: A review, Fig. 3, p. 377, Agricultural and
Forest Meteorology, 54; Copyright c 1991 Elsevier Science, with kind permission
from Elsevier Science.)
side of Fig. 4.29 is that for sensible heat in which the canopy resistance, rcH , is fixed.
Invoking the Reynolds analogy, the effective interleaf resistance will be the same for
mass and heat giving
The atmospheric resistance, ra , regulates the flux from canopy top out to screen height
(above the canopy), and is given by Eq. 4.141 (or 4.143) independent of our canopy
model.
By analogy with Eq. 4.155, the governing PenmanMonteith equation for transpi-
ration from a square centimeter of dry canopy is then
cp
Rn + [es (T2 ) e2 ]
ra
E v = , (4.158)
rc
+ o 1 +
ra
where elevation 2 is now taken at screen height above the canopy top. The canopy
resistance, rc , is the primary flux-controlling plantatmosphere property, and its esti-
mation poses the greatest challenge to successful use of Eq. 4.158. Monteith (1963,
discussion) considered rc (rs in his notation) to be strongly dependent upon stomatal
resistance, whereas Philip (1966, p. 264) suggested that further work would show
rc (again rs in his notation) is subject to great variability and is relatively weakly
Transpiration from a dry forest canopy 141
linked with stomatal resistance. As we show below, our model of the big leaf canopy
provides theoretical support for Philips conjecture.
As a canopy grows from a single leaf where rc rl = rls +rla (Eq. 4.148), to many,
many leaves where rc = rcs + rci , the relative importance of the internal and external
resistances reverses.
zL
t 1
inactive z = hs , = 1
142 Turbulent uxes
ra
z=h =0
r
i
r r n=1 r
i i i
rlls
r
i
r n=2 r
i i
rlls
r
i
n=3 r
i
rlls
r
i
n = Lt
rlls
z = hs = h hs
diagram shown in Fig. 4.31. In this model, the water vapor emanating from the nth leaf
layer down from the canopy top must diffuse serially through n interleaf atmospheric
resistances, while all of these n separate diffusion paths are in parallel. From this
seriesparallel circuit we have
1 1 1 1
= + + + . (4.160)
rc rlls + ri rlls + 2ri rlls + (L t 1)ri
under no stress. A basic assumption of this work is that natural selection will set the
average operating state as close to the stressless state as possible. Therefore, using the
minimum single leaf resistances ( per unit of leaf area), rls , of Table 4.3 in Eq. 4.161
yields the mimimum equivalent leaf layer resistance ( per unit of leaf layer area),
L trls
rlls = . (4.162)
nl
L trls 5 500
rlls = = 0.42, (s m1 ). (4.163)
nl 6000
Using Eq. 4.143 with Eq. 4.163 an estimate of the dimensionless resistance ratio is
u (h h s )
ri = = , (4.165)
K m ( )
in which K m has been lumped at the scale of z L . Accuracy of Eq. 4.165 will be
inversely related to this lumping scale. Although the ri will actually vary with due to
the dependence of ri upon K m ( ), we will neglect this variation here, replacing K m ( )
by Km ( ). With this first order approximation, the ri become identical, and we use
Eqs. 4.38 and 4.159 to write Eq. 4.165 as
1 h hs 1
ri = . (4.166)
ku h d0 mo
(L t 1) K
For homogeneous cylindrical multilayers, we use Eqs. 4.35, 4.59, and 4.41 in Eq. 4.166,
to get the first order estimate of ri
1m
1 ( L t )2
ri = m , (4.167)
ku 1m
(L t 1) 1 exp Lt
m
and with Eqs. 4.59 and 4.143, the associated resistance ratio is
For the common case: m = 0.5, n = 2, = 0.5, L t = 5, Eq. 4.168 gives the
estimate
ri
1.7, (4.170)
ra
Plate 4.7. Moreton Bay fig. (Photographed by Peter S. Eagleson in the Royal Botanic
Gardens, Sydney. Identification kindly supplied by Eric and Daphne Laurenson.)
Transpiration from a dry forest canopy 145
1.0
0.9
Lt = 1 5 10
0.7
0.6
0.5
10 2 10 1 1
rlls / r
i
Fig. 4.32. Dominance of interleaf layer resistive component of big leaf canopy
model.
1 1 1 1
+ + + . (4.173)
rc ri 2ri (L t 1)ri
We can find the effect on rc of neglecting rlls by dividing Eq. 4.160 by Eq. 4.173
1 1 1
+ + +
1 + rlls /ri 2 + rlls /ri (L t 1) + rlls /ri
rc (ri )/rc (ri , rlls ) = . (4.174)
1 + 1/2 + 1/3 + + 1/(L t 1)
Equation 4.174 is plotted in Fig. 4.32 for various values of L t . We see there that the
estimate, rrllsi = O(101 ), leads to an error of only 6% in rc when we estimate the
latter using Eq. 4.173 in which the stomatal resistance is neglected. This is contrary to
conventional wisdom but provides a theoretical basis for the Philip (1966, p. 264)
conjecture mentioned earlier.
We will use the approximation rc = rc (ri ) to evaluate the resistance ratio, rc /ra , in
Chapter 7.
5
Within the control volume are the rates of change of stored energy density (W m2 ):
We restrict this analysis to surfaces which are closely homogeneous laterally in support of our
neglect of lateral energy fluxes.
146
Denitions and assumptions 147
Rn H Qv
Screen height
DH DQv
G
Fig. 5.1. Control volume for thermal energy balance.
With these definitions the thermal energy balance for the control volume is
Rn + D H = Q v + H + G + B + S + D Qv . (5.1)
Let us examine the relative importance of the individual terms of Eq. 5.1:
(1) There is general agreement in the literature (e.g., Gates, 1980, p. 26; Larcher,
1983, p. 20) that the biological rate of energy consumption, B, is negligible in
comparison with the other terms of this energy balance.
(2) The substrate heat storage, G, will have a high diurnal and seasonal fluctuation,
particularly in the presence of high soil moisture, but its long-term average rate
is zero. Similarly, if we confine our interest to the long-term average, we can
omit S as being identically zero.
(3) For a typical energy budget over vegetation, Thom (1975, Table I, p. 98)
shows that D H and D Qv are small enough to be neglected for the control
volume chosen. However, for non-zero H, there must be flux divergence
external to this control volume and fixed by conditions in the planetary
boundary layer. Such geographically and seasonally variable fluxes may
drastically alter the findings of this chapter. To promote generalization, we
make the simplifying assumption that these active connections to the external
system are severed and replaced by prescribed climatic boundary conditions,
T0 , S r , u 2 , etc.
D Qv = D H = G = S = B = 0, (5.2)
and, for an evaporating (as opposed to a condensing) surface, the long-term average
climatic thermal energy balance reduces to
Rn = Q v H, (5.3)
For example, does the boundary layer entrain dry air (cf. de Bruin, 1983)?
148 Thermal energy balance
in which we have introduced the possibility that sensible heat may be added to
(i.e., H ) the control volume of Fig. 5.1 rather than being rejected by it (i.e., +H ) as
the figure shows. Using Eq. 4.129, we introduce the evapotranspiration, E T , to write
Eq. 5.3
Rn = E T H, (5.4)
Plate 5.1. Royal palm. Elevation of leaf surface is one mechanism for temperature
control as employed by this palm in India. (Photograph by Peter S. Eagleson.)
Evaporation in state space 149
in which D12 is called the wet-bulb depression, that is the difference between the
actual temperature and the saturation temperature at constant heat. With = angle of
inclination of the adiabatic line 12 in Fig. 5.2, we thus have
tan = o . (5.7)
150 Thermal energy balance
T)
e s(
Liquid Vapor
phase phase
cp D43
es(T3) ET =
3 Adiabat ra
Vapor pressure, e (mb)
e4 4
Rn
e ET =
2 + o
1
es(T2) 2 tan
cp D12
Adiabat ET =
e1 1 1 ra
Lines of T
constant
relative
humidity T2 T1 T3 T4
Temperature, T (C)
Fig. 5.2. The change of moisture state. (After Monteith, J.L., Evaporation and
environment, in The State and Movement of Water in Living Organisms, edited by
G.E. Fogg, Part II, Water in the Plant, Fig. 1, p. 207; Copyright
c 1965 The Society
for Experimental Biology; with kind permission from J.L. Monteith.)
To move along the saturated path from point no. 2 to point no. 3, we must increase
both the latent heat (higher vapor pressure) and the sensible heat (higher temperature)
of the parcel. This requires forcing from outside the parcel through the addition of
heat at the long term average rate, Q. In our simplified case of negligible advection
and storages this added energy is the net solar radiation, and hence Q Rn . The
flux of this driving energy is partitioned between increases in latent heat ( Q v ) and
sensible heat ( H ) along the saturated path 23. If the changes e and T are small
we can replace the curve es (T ) by its chord having slope, , as given by Eq. 4.134.
The Bowen ratio definition (cf. Eq. 4.131) then gives
H o
= . (5.8)
Qv
Equations 5.8 and 5.3 yield
o Rn
H = , (5.9)
o
and, for either direction of H,
Rn
Q v E T2 = . (5.10)
+ o
The saturated energy partition parameter, /o , is a function (through ) of the
Evaporation in state space 151
and Eq. 5.7 again applies. In this last condensation, 1 cm3 of saturated air will give up
cp (T4 T3 ) calories of latent heat which will be absorbed by the high sensible heat
capacity of the liquid water.
With the parcel being adjacent to a saturated surface, the source and depository
of the parcels liquid water is this surface. The time required for the exchange of the
released sensible heat with 1 cm2 of the saturated surface is called the atmospheric
resistance, ra (s cm1 ), and the rate at which the latent heat is removed from the parcel
in this step becomes
cp D43
E T3 = . (5.12)
ra
0.0 0.67 10.0 1.23 20.0 2.14 30.0 3.57 40.0 5.70 50.0 8.77
0.5 0.69 10.5 1.27 20.5 2.20 30.5 3.66 40.5 5.83 50.5 8.96
1.0 0.72 11.0 1.30 21.0 2.26 31.0 3.75 41.0 5.96 51.0 9.14
1.5 0.74 11.5 1.34 21.5 2.32 31.5 3.84 41.5 6.09 51.5 9.33
2.0 0.76 12.0 1.38 22.0 2.38 32.0 3.93 42.0 6.23 52.0 9.52
2.5 0.79 12.5 1.42 22.5 2.45 32.5 4.03 42.5 6.37 52.5 9.72
3.0 0.81 13.0 1.46 23.0 2.51 33.0 4.12 43.0 6.51 53.0 9.92
3.5 0.84 13.5 1.50 23.5 2.58 33.5 4.22 43.5 6.65 53.5 10.1
4.0 0.86 14.0 1.55 24.0 2.64 34.0 4.32 44.0 6.80 54.0 10.3
4.5 0.89 14.5 1.59 24.5 2.71 34.5 4.43 44.5 6.95 54.5 10.5
5.0 0.92 15.0 1.64 25.0 2.78 35.0 4.53 45.0 7.10 55.0 10.8
5.5 0.94 15.5 1.68 25.5 2.85 35.5 4.64 45.5 7.26 55.5 11.0
6.0 0.97 16.0 1.73 26.0 2.92 36.0 4.75 46.0 7.41 56.0 11.2
6.5 1.00 16.5 1.78 26.5 3.00 36.5 4.86 46.5 7.57 56.5 11.4
7.0 1.03 17.0 1.82 27.0 3.08 37.0 4.97 47.0 7.73 57.0 11.6
7.5 1.06 17.5 1.88 27.5 3.15 37.5 5.09 47.5 7.90 57.5 11.9
8.0 1.10 18.0 1.93 28.0 3.23 38.0 5.20 48.0 8.07 58.0 12.1
8.5 1.13 18.5 1.98 28.5 3.31 38.5 5.32 48.5 8.24 58.5 12.3
9.0 1.16 19.0 2.03 29.0 3.40 39.0 5.45 49.0 8.42 59.0 12.6
9.5 1.2 19.5 2.09 29.5 3.48 39.5 5.57 49.5 8.60 59.5 12.8
10.0 1.23 20.0 2.14 30.0 3.57 40.0 5.70 50.0 8.77 60.0 13.1
Source: Van Bavel (1966), with permission of the American Geophysical Union.
152 Thermal energy balance
Similarly, for the first adiabatic step, latent heat was added to the parcel at the rate
cp D12
E T1 = . (5.13)
ra
The first term of Eq. 5.15 is commonly referred to as the radiational forcing,
and the second term as the (net) drying power of the air, that is the capacity of the
ambient atmosphere to augment (or diminish) the radiational forcing through adiabatic
transformation of excess sensible heat into additional latent heat.
It is important to remember that Eq. 5.15 is a rate equation which we have examined
over a finite interval of time, t. The thermodynamic state of a particular air mass is
therefore changing with time, and without continuous replacement of the moistened
air mass at location xi jk with a dry one at the same location, the evaporation rate
there will decline steadily. For homogeneous surfaces of infinite extent, horizontal air
movement alone will not replenish the evaporative capacity, and vertical transport of
heat and mass, primarily due to turbulence in the surface atmospheric boundary layer,
is required. We have discussed this transport at length in the preceeding chapter.
Rn
Q v E 1 = , (5.16)
+ o
Evaporation in state space 153
T)
e s(
( surface
Wet
) end
3
C
Rn
E1 =
+ o
1
2 tan Rn
D
1
tan o cp D12
E2 =
ra
1a
A +H
4 H
( Screen
height) start
1b
Temperature, T (C)
Fig. 5.3. Evaporation from a warm, wet surface.
while the drying power of the ambient atmosphere leads to the additional evaporation
cp D12
E 2 = , (5.17)
ra
which has different numerical values for the different starting-points no. 1a and no. 1b.
The total potential evaporation is then
Rn cp D12
E ps = (E 1 + E 2 ) = + , (5.18)
+ o ra
as was originally presented by Slatyer and McIlroy (1961).
The wet bulb depression, D12 , is related to the saturation vapor pressure deficit,
es (T ) e, that is the difference between saturated and actual vapor pressure at the
actual temperature, by
es (T ) e es (T1 ) e1 es (T1 ) es (T2 ) es (T2 ) e1
= + + o , (5.19)
D12 T1 T2 T1 T2 T1 T2
whereupon Eq. 5.18 is written
Rn + cp [es (T ) e]/ra
E ps = , (5.20)
+ o
where E ps = bare soil potential evaporation rate, and in which for practical appli-
cations all atmospheric variables are measured at screen height above the surface.
154 Thermal energy balance
Equation 5.20 is recognized as the classic Penman equation (Penman, 1948) for po-
tential evaporation from a saturated surface and was derived here in another fashion
in Chapter 4 (cf. Eq. 4.142). For later convenience, Eq. 5.20 is also written
1
E ps = (Rn + Dp ), (5.21)
1 + o /
in which the variable, Dp , is defined as
cp es
Dp (1 Sr ), (5.22)
ra
and is called the drying power, while Sr is the (constant temperature) saturation ratio,
Sr e/es , i.e., the (fractional) relative humidity.
Plate 5.3. Branched palm. Note the symmetry of the branching structure in this
unusual Kenyan species. (Photograph by Peter S. Eagleson.)
156 Thermal energy balance
the sensible heat rejection decreases. When point no. 1 coincides with A, the system
is isothermal, thus H 0 and E ps = Rn (cf. Eq. 5.4). Moving still further along this
adiabat to an even drier ambient atmosphere at point no. 1b, the evaporation potential
exceeds the available solar radiation and is met by the importation of additional energy
through a negative flux of sensible heat. In all these cases, the atmospheric resistance,
ra , plays a key role in determining H for a given vapor pressure deficit. As we will see
later, the role of this resistance is markedly increased when the surface layer contains
a high-impedance vegetal canopy.
This condition might be approached, but not equaled, over deeper water bodies where
mixing in both water and air prevents the maintainance of a vertical atmospheric
temperature gradient near the surface. In such cases Eq. 5.4 gives
E ps = Rn . (5.24)
When points no. 1 and no. 2 are coincident, the atmosphere is saturated to screen
height and the vertical humidity gradient is zero. In this case, D12 0 and, using
Eqs. 5.16 and 5.5, the Bowen ratio takes on the limiting value for positive H
o
(Rb )max = , (5.25)
which is less than unity for temperatures above 6.5 C (cf. Table 5.1). This condition
might be expected over smooth, wet surfaces such as mud flats where the near-surface
mixing may be low enough and thus the resistance high enough to make the second
term of Eq. 5.20 negligible with respect to the first term. The same condition may
occur over shallow, wave-free water where the temperature difference between water
and atmosphere (i.e., D12 in Eq. 5.20) is small, that is whenever water is evaporating
into saturated air. In these cases, Eq. 5.20 leads to
Rn
E ps = . (5.26)
+ o
It is important to note that for negative H there is no obvious upper bound to Rb .
Short-term observations of the resistances of a pine canopy reported by Monteith
(1965, Table 2, p. 218) give Rb = 0.42, while short-term flux observations over
Sitka spruce, Scots pine, and Douglas fir (see Jarvis et al., 1976, Fig. 11, p. 210),
Philip (1987) found a higher bound above very cold surfaces at the onset of supersaturation.
Transpiration 157
0.5
8
0.4 Ze
ro
(N dry
o ing
hu
5 mi po
0.3 dit we
y r
gra
die R
Eq. nt) b =
5.37 o /
R b 0.2 4
7
6 1 R =
b (1/k *)
v 1 with rc
0.1 ra = 0.5
3
2
Isothermal 9 Rb = 0
0
(No temperature gradient)
0.1
10 15 20 25 30 35 40 45
T0 (C)
Fig. 5.4. Observed Bowen ratios under wet surface conditions. No energy advection,
no energy storage, and stable atmosphere.
Observation Location Surface
a
Conditions selected to be advection-free.
b
Dormant season.
(After Priestley and Taylor, 1972, Fig. 2, p. 87, courtesy of the American
Meteorological Society.)
lead to Rb values of 1.32, 2.31, and 0.40 respectively. Being greater than unity, the
first two of these are undoubtedly for surfaces having negative sensible heat flux.
Day-long observations over a variety of wet surfaces reported by Priestley and Taylor
(1972) and selected for the apparent absence of energy advection (1972, Fig. 2,
p. 87), yield the much smaller measured Bowen ratios plotted here in Fig. 5.4 where
they are seen to fall within the limits provided by Eqs. 5.23 and 5.25.
D Transpiration
The PenmanMonteith equation
It was shown in Chapter 4 that by using the so-called big leaf approximation for
a complex canopy of simple surfaces having aggregate flux resistance rc , Eq. 5.20
158 Thermal energy balance
can be used to approximate the transpiration from the entire dry vegetated fraction.
Monteith (1965) pointed out that the effect of the canopy resistance is as though the
simple surface of Eq. 5.20 was immersed in a different fluid having the augmented
psychrometric constant
o o (1 + rc /ra ). (5.27)
That is, for the dry vegetated fraction we have the PenmanMonteith equation
For the special case of open stomates, E v will take on its maximum value, E v = E pv ,
where E pv = potential canopy transpiration, whereupon kv has its maximum value
(kv < 1), called the potential canopy conductance
E pv
k v . (5.30)
E ps
It is important to recognize the difference between Eq. 5.29 as derived here and the
same equation as it commonly appears in the literature:
1. Here, with the stomates assumed always fully open, rc is dominated by the
interleaf layer atmospheric resistance, ri (see Chapter 4), which doesnt vary
with canopy water supply. In particular, rc doesnt vanish identically when the
canopy water supply is unlimited. Therefore, in this work the canopy
conductance takes on the maximum, open stomates value, kv k v , for the
given species. Note in particular that with rc dominated by the interleaf
atmospheric resistance, the rate of canopy evapotranspiration is approximately
the same when the exterior leaf surfaces are wet (such as after rain) as it is
when the wet evaporating surface is the interior lining of the stomatal cavity
(dry leaf stressless transpiration). The condition k v < 1 reflects that canopy
atmosphere control of the energy exchange discussed in the previous section.
2. In the literature rc is assumed to be dominated by the effective canopy stomatal
resistance. Equation 5.30 is obtained in such cases by normalizing the E v of
Eq. 5.28 with a rate given by Eq. 5.28 in which the plant canopy has saturated
leaf surfaces so that this stomatal resistance and hence rc vanish. In such cases
kv 1 as the water supply increases and the canopy (i.e., stomatal) resistance
vanishes.
can be converted into latent heat. This fraction is unity when the atmosphere and the
surface are at the same temperature or when the atmosphere is warmer than the surface.
Equation 5.34 may be written, albeit less revealingly, as
Rn
E ps = , (5.35)
+ o
which was first proposed by Kim and Entekhabi (1997). But how do we estimate the
Bowen ratio a priori?
Returning to Fig. 5.3 we note again that whenever T1 < T3 some of the net solar
radiation is not used in evaporation. In the absence of energy storage and lateral
advection, the excess energy is shed back to the atmosphere through sensible heat
transfer. Given that there is unlimited liquid water at the surface, this departure from
isothermality must be due to flux impedance in the water vapor path between surface
and atmosphere (i.e., screen height). For bare soil surfaces this path involves only
the atmospheric boundary layer over the 2-meter screen height so the impedance is
small. However, for vegetal canopies the vapor flux impedance can be high as we
have seen by the departure of the maximum canopy conductance, k v , from unity
(Eq. 5.30). Invoking Reynolds similarity between the fluxes of mass and heat, we
expect the Bowen ratio to be greater than zero to a degree that is proportional to the
difference of k v from unity (see the discussion following Eq. 5.22). In a successive
Transpiration 163
approximation we use Eqs. 5.30 and 5.34 in Eq. 5.5 to get the estimators
H Rn E ps 1
Rb = 1 1 = 1, +H (5.38)
E v E v E pv kv
and
H Rn E ps 1
Rb =1 1 = 1 , H. (5.39)
E v E v E pv kv
Observations
Pine forest
Dry (Norfolk, England) 0.74 Shuttleworth and Calder (1979)
Wet (Norfolk, England) 9.69 Shuttleworth and Calder (1979)
Wet 1.50 Gash and Stewart (1975)
Sitka and Norway spruce
Dry (Central Wales) 0.64 Shuttleworth and Calder (1979)
Wet (Central Wales) 1.50 Shuttleworth and Calder (1979)
Bush (wet soil) 1.33 Priestley and Taylor (1972)
Pasture crop (wet soil) 1.34 Priestley and Taylor (1972)
Snap bean crop (wet soil) 1.30 Priestley and Taylor (1972)
Unstressed agricultural
crops and grasses 1.26 Davies and Allen (1973)
Plowed bare wet soil 1.08 Priestley and Taylor (1972)
Indian Ocean (CSIRO) 1.26 Priestley and Taylor (1972)
No advection 1.25 Priestley and Taylor (1972)
Indian Ocean (CSIRO) 1.30 Priestley and Taylor (1972)
No advection 1.31 Priestley and Taylor (1972)
Indian Ocean
(University of Washington) 1.20 Priestley and Taylor (1972)
Lake Eucumbene 1.25 Priestley and Taylor (1972)
Equatorial Atlantic Ocean 1.30 Priestley and Taylor (1972)
Theory
Forest 3.62 Lhomme (1997)
Grass 1.26 Lhomme (1997)
Unsaturated surface 1 McNaughton (1976), Perrier
Moderately dry surface <1 (1982), Brutsaert (1982)
Realistic atmospheric conditions,
growing convective boundary layer <1.46; >1.08 Culf (1994)
Moderate weather in the Netherlands slightly >1.26 McNaughton and Spriggs (1989)
164 Thermal energy balance
1.8
1.7
1.6
Eq
.5
.36
1.5
Wet pine
Wet spruce
Eq
1.4
s.
5.
29
, 5.
38
Iso
the
5 rm
al,
1.3 8
6 3 9 R
b =0
2
4
1.2 R =
1 b (1/k *
v )1
rc
ra = 0
.5
1.1
7
Zero drying power, R b = o /
1.0
10 12 14 16 18 20 22 24 26 28 30 32 34 36
T0 (C)
Note that Rb is a ratio of flux rates. Therefore in this calculation the observed gross annual
evaporation is doubled to obtain an estimate of the evaporation rate during daylight hours for
comparison with Rn .
Table 5.3. Estimation of potential transpiration for vegetated surfaces (negligible snow and ice)
f g j
I 0b E T a (Obs.) q i e qb Rn D ph E ps
o i
Site locationsa (N) (ly min1 ) Tc T 0d ( C) Nd Sr d (m y1 ) (ly min1 ) (ly min1 ) (ly min1 ) (ly min1 ) 1 +
(calc.) (m y1 ) k v k Rb l
10 Mesilla, NM 32.3 0.329 0.18 14.6 0.18 0.53 0.86 0.270 0.125 0.145 0.069 1.63 1.15 0.75 0.23
11 Pecos, NM 35.6 0.316 0.18 13.8 0.18 0.45 0.90 0.259 0.126 0.133 0.081 1.65 1.14 0.79 0.33
12 Green River, KY 37.9 0.307 0.18 13.1 0.46 0.73 0.80 0.252 0.093 0.159 0.029 1.68 0.98 0.82 0.10
13 Tallapoosa River,
GA 32.5 0.328 0.18 15.4 0.36 0.77 0.84 0.269 0.103 0.166 0.028 1.60 1.06 0.79 0.10
14 West Fork of the
White River, MO 37.0 0.309 0.18 13.4 0.33 0.72 0.79 0.253 0.109 0.144 0.037 1.67 0.95 0.83 0.17
15 Cyprus Creek, TX 32.0 0.329 0.18 17.9 0.20 0.75 0.92 0.270 0.113 0.157 0.033 1.52 1.09 0.84 0.22
a
Hamon (1961).
b
Average of observations, Fig. B.2.
c
Kondratyev (1969, Table 7.5, p. 425); no attempt made to distinguish variations: assumed for all sites: 50% grass @ T = 0.22 and 50% forest @ T = 0.14.
d
U.S. Weather Service Annual Climatological Summary, 194170 (at nearest station).
e
qi = (1 T )I0 .
f
Eq. B.12.
g
R n = qi qb .
h
Eq. B.18.
i
Table 5.1.
j
Eq. B.1.
k
Eq. 5.30 with the assumption that E pv E T .
l
Eq. 5.5, using twice the observed annual E T as an estimate of the daylight-hour rate.
Canopyatmosphere temperature 167
Table 5.3.
0.6
R
b
0.4 11
10 15
0.2 14
13
12
0
0.5 0.6 0.7 0.8 0.9 1.0
k *v
atmosphere
es
T 0.622 Sr Rb (5.42)
cp p
with
597.3 cal g1 ,
cp 0.24 cal g1 K1 ,
p O(1000 mb), and
es O(10 mb),
Under normal conditions, Sr = O(101 ), so with Rb = O(101 ) as was found in the
previous section, Eq. 5.43 gives T = O(101 ), and the system is closely isothermal.
In this work we will make this approximation, assuming the long-term average big
leaf canopy temperature, Tl , equals the ambient (i.e., screen height) atmospheric
temperature, T0 . That is
Tl T0 . (5.44)
Gates et al. (1968) show that Tl T0 23 C for many small leaves. Monteith (1973,
p. 183) notes that in bright sunshine and in a breeze, small leaves are expected to be
only 12 C hotter than the surrounding air. Greater excess temperatures are observed
on very large leaves in a light wind. . .. Measurements within a spruce canopy bear this
out (Larcher, 1983, Fig. 2.16, p. 27). They show a temperature difference, Tl T0
1 C with averages taken over a typical midsummer day, while finding for a single
168 Thermal energy balance
Canna leaf that this difference is about +1 C with high wind but rises to about +20 C
for low wind. (Note that compared to a single leaf, the long-term canopy average Tl
is much less sensitive to wind, and the canopy Tl is much smaller than for a single
leaf.)
Looking at Eq. 5.42 from another direction, we can use the observed small
temperature gradients over forests (cf. Jarvis et al., 1976, Table IX, p. 204) to obtain
T = O(101 ), and with Sr = O(101 ) as before, we again obtain Rb = O(101 ).
This is in contrast to the much higher values, Rb > 1, calculated by Jarvis et al. (1976,
Fig. 10, p. 209) using much larger values of rc /ra than those derived here in Chapter 7.
Plate 5.6. Siberian larch. Author alongside larch in Arnold Arboretum, Boston.
(Photograph by Beverly G. Eagleson.)
Relevance to nutrient and carbon ux 169
Water balance
A Introduction
At land surfaces the soil column responds dynamically to the climatic sequence of
precipitation and evapotranspiration events. It accepts part of the moisture during
periods of precipitation, pumps some of this back to the surface during evaporative
periods, and rejects the remainder to the water table more or less continuously. This
surface moisture exchange and the accompanying surface heat exchange, to a large
degree, depend critically upon the physical properties of the soil and vegetation as
well as upon the weather conditions during the alternate periods of precipitation and
evapotranspiration. The quantitative relation among the long-term averages of this
partition of precipitation is called the water balance.
We seek here to express this climateland surface coupling in a simplified form
which provides insight into the physical basis for the role played by water in the growth
Much of the material of this chapter is taken from Eagleson (1978a through 1978g) incorporating
the corrections and extensions provided in the MIT Master of Science thesis of Salvucci (1992).
The latter work is also available in an MIT Laboratory Report by Salvucci and Eagleson (1992).
This summary does not contain all the details of the assumptions and approximations which are
available in the original works (Eagleson, 1978ag; Salvucci and Eagleson, 1992).
170
Introduction 171
of vegetation communities. Such insight can come only through retention in our model
of the underlying physical determinism. At the same time however, uncertainty plays
such a large role in the climatic forcing of the water balance at its critical time-scales
that our approach must deal also with probability distributions. This is in contrast to
the approach taken in describing the radiation forcing in Chapter 3 where variability at
the time-scales of vegetation growth is dominated by deterministic planetary motions.
Analytical solutions, even approximate ones, are preferred to precise numerical
solutions whenever the objective is behavioral insight, and this model is structured ac-
cordingly. Only an outline is presented here; details are available elsewhere (Eagleson,
1978bg; Salvucci, 1992; Salvucci and Eagleson, 1992).
Plate 6.1. Giant Ficus tree exploits water source. At Ta Prom, Angkor, Cambodia,
spong tree (Ficus tetrameles nudiflora) sends spreading roots down wall to exploit soil
moisture in unvegetated courtyard. (Photograph by Peter S. Eagleson.)
172 Water balance
B Model framework
The water balance model (Eagleson, 1978a) is a one-dimensional model of soil mois-
ture dynamics forced by a stochastic climate. The model solves the equilibrium moist-
season water balance in terms of a soil moisture which is averaged temporally over
climate scale (so that storage changes may be neglected), and is averaged vertically over
the vegetation rooting depth, z r , (on the order of 1 meter). The resulting equilibrium
spacetime average soil moisture concentration, so , is the state variable determining
the average flux rates of moisture into and out of the near-surface soil. Multiplied by
a common time, season or year, these average flux rates become depths (per unit of
suface area) and must sum to zero. This sum is the water balance equation and is
commonly written on an average annual basis as
where
PA = average annual precipitation (cm),
E T A = average annual evapotranspiration (cm),
RsA = average annual surface runoff (cm), and
RgA = average annual groundwater runoff (cm).
so Simplified
spatial avg.
0
Time, t
Model framework 173
Evapotranspiration
between Capillary
Infiltration storms rise
during from
storms water table
Soil moisture
zone
zr
Wet season
percolation Continuous
to water table capillary rise
from water table
zw
Saturated Groundwater
zone runoff
Fig. 6.2. Moisture fluxes in soil column. (Eagleson, 1978a, Fig. 6, p. 708; with
permission of the American Geophysical Union.)
transpiration, E[E vj ], are found using the simplifying assumption that the spatial av-
erage soil moisture concentration, s, is at its equilibrium value, so , at the beginning
of each storm and interstorm period. A qualitative comparison between the simplified
and actual variability of s is sketched in the lower part of Fig. 6.1.
The soil moisture dynamics at the bottom boundary of the near-surface soil column
are predicated upon the restrictive assumption of a deep water table that does not
influence the transient moisture dynamics of the upper soil layer. With this assumption,
the vertical diffusive processes throughout the soil profile are assumed to attenuate the
random near-surface variations so that at the lower boundary of the soil moisture zone
the soil moisture is at steady state yielding a constant percolation, v, down to the water
table during the wet season and a constant capillary rise, w, from the water table to
the surface year around. Long-term mass fluxes into and out of the saturated zone are
balanced by lateral groundwater runoff. These fluxes are illustrated in the sketch of
the soil column shown as Fig. 6.2.
174 Water balance
C Precipitation
Point precipitation is represented by Eagleson (1978b) as Poisson arrivals of rectangu-
lar intensity pulses that have random depth and duration as is illustrated in the upper
portion of Fig. 6.1. The probability density functions (i.e., pdfs) of these variables are
incomplete gamma, (o , o ), for the storm depth, h, giving
o (o h)o 1 eo h
fh (h) = (o , o ) , (6.2)
(o )
in which o = shape parameter, or index of the distribution (dimensionless),
o = scale parameter of the distribution (cm1 ),
() = factorial gamma function (see Table 6.1a),
and the mean and variance are, repectively
m h = o /o , (6.3)
and
h2 = o /2o ; (6.4)
m tr = 1/, (6.6)
and
t2r = 1/ 2 . (6.7)
The potential bare soil evaporation rate is assumed to have the same constant
value throughout all of the interstorm periods whose duration, tb , is also exponentially
distributed,
with mean
m tb = 1/, (6.9)
and variance
t2b = 1/ 2 . (6.10)
Precipitation 175
fi (i) = ei , (6.11)
with mean
m i = 1/, (6.12)
and variance
i2 = 1/2 . (6.13)
The storm arrival rate is m , storms per year (or season), so that the mean of the
annual (or seasonal) precipitation, P, is simply
PA m P = m m h , (6.14)
and finally, the variance of the annual (or seasonal) precipitation, P, is (Eagleson,
1978b)
1
P = m P 1 +
2 2
/m . (6.15)
o
See Appendix F for the monthly climatology of these and other storm parameters at 74
first-order weather stations in the continental United States as determined by Hawk
and Eagleson (1992).
Plate 6.2. Weeping beech. Author alongside beech in Arnold Arboretum, Boston.
(Photograph by Beverly G. Eagleson.)
176 Water balance
1 1/2
f i = Si t + A0 , (6.16)
2
1
A0 = K (1) 1 + soc w, (6.17)
2
and where
ne = effective porosity of soil,
K (1) = saturated effective hydraulic conductivity of soil, cm s1 ,
(1) = saturated matrix potential of soil, cm (suction),
c = permeability index of soil,
d = diffusivity index of soil, and
m = pore size distribution index of soil.
Equation 6.19 is plotted in Fig. 6.3.
Storm inltration and surface runoff 177
m=
i(d, so)
d =2
m=1
d =3
.5
m=0
d=4
0.25
m=
d =6
1
10 1
10 1
so
Fig. 6.3. Dimensionless sorption diffusivity; see Eq. 6.19. (Eagleson, 1978c, Fig. 9,
p. 727; with permission of the American Geophysical Union.)
Remembering our restrictive assumption of a deep water table which prevents soil
saturation from below, the process physics responsible for generation of storm runoff
is as shown in Fig. 6.4 for storm durations, tr , large enough such that all three distinct
stages of infiltration are operative:
(1) in the first stage the surface retention capacity, h o , is filled at rate i. This storage
is isolated from the porous soil and so no infiltration takes place in time h o /i;
(2) in the second stage infiltration takes place at rate i because the infiltration
capacity, f i , exceeds the rainfall intensity, i, and this continues until f i = i at
time to + h o /i; and
(3) in the third stage infiltration occurs at the declining rate f i due to this rate
being less than the rainfall intensity, i, and it continues, producing runoff at the
rate i f i until the storm ends at time tr with a total storm runoff Rsj .
The average value of the storm surface runoff is obtained by taking the expected
value of Rsj over the joint probability distribution of storm intensity, i, and duration,
tr . For mathematical convenience, Eagleson (1978e) made the simplifying assumption
that i and tr are independent random variables so that with Eqs. 6.5 and 6.11, the
Appendix F shows that in the continental United States |[i, tr ]| O(101 )
178 Water balance
f i*
Flux rates
fi = i
ho to Rsj
i
f i*
0
0 ho tr
to +
i
Time, t
Fig. 6.4. Surface runoff generation during typical storm. (Eagleson, 1978e, Fig. 1,
p. 742; with permission of the American Geophysical Union.)
m h = ()1 . (6.21)
With these definitions and assumptions, the expectation of storm rainfall excess,
Rsj ,
is given by the sum of the two shaded areas of Fig. 6.4 (see Eagleson, 1978e,
p. 745)
E[Rsj ] E[E r ]
= eG2 ( + 1) + (6.22)
mh mh
and by definition of the gamma function (see Abramowitz and Stegun, 1964)
G E[RsA] 2 ( + 1)
e = = e
E[PA]
[ ]
1/3
5ne o2K(1) (1) (1so)2i (d, so)
=
6mo2
K(1)
G =
2
[1 + soc] w
[
= exp 2 3/2 ]
Dry soil
Wet soil
101 1
10 1
Fig. 6.5. Surface runoff function. (Eagleson, 1978e, Fig. 6, p. 746; with permission of
the American Geophysical Union.)
Equation 6.26 is presented graphically as the solid line in Fig. 6.5 where the
plotted points represent the convenient approximation we call the surface runoff
function,
( + 1) 2
e2 .
3/2
( ) e (6.27)
The storm infiltration, Isj , is simply the storm depth minus the surface runoff and
minus the surface retention. Averaging over all storms this becomes
E v = E pv = kv E ps , (6.29)
Potential (unstressed) transpiration 181
where
E v = long-term annual or seasonal average transpiration rate,
E pv = long-term annual or seasonal average potential transpiration rate,
kv = unstressed (i.e., maximum or potential) canopy conductance, and
E ps = long-term annual or seasonal average potential rate of evaporation from the
bare soil.
which, with the typical soil properties of Table 6.2, shows the critical soil moisture,
sc s5 , corresponding to s = sc = 5 bar to be: s5 = 0.16 for clay; s5 < 0.003
a
From Bras, R.L., Hydrology, Table 8.1, p. 352; Copyright
c 1990 Addison-Wesley
Publishing Co., Inc.; with kind permission from R.L. Bras.
b
From Salvucci (1992).
182 Water balance
for silty loam; s5 < 108 for sandy loam; and s5 < 1014 for sand (Salvucci, 1992).
Except perhaps for the clay soil, exhaustion of soil moisture from so to these small
critical values at which stress appears will take a longer time, ts , than all but the rarest of
interstorm intervals, tb . Considering the fact that as modeled, so lies at the midpoint of
the interstorm soil moisture range, ts is even larger and we conclude that transpiration
will occur primarily at the stressless, potential rate. Such reasoning is fundamental to
our basic assumption that natural selection fixes E v , the expected value of interstorm
transpiration, at the unstressed level for the given climate and soil. Using Eqs. 5.30
and 6.29, we have
E v E[E vj ] = E pv m tb = kv E ps 1 . (6.31)
The time interval, m tb , is the mean evapotranspiration window over which the maxi-
mum expected values, E[E sj ] and E[E vj ] are calculated.
Salvucci (1992) pointed out that in reality the soil moisture may reach the critical
value s5 before the end of the interstorm period, whereupon the transpiration rate would
be throttled by closure of the stomates, and E[E vj ] could be much less than is given
by Eq. 6.31. This can happen in arid and semi-arid climates for two reasons:
(1) large m tb will insure that during a fraction of the interstorm intervals the critical
soil moisture is reached, and
(2) dry climates have large m tb and hence few storms (m ) annually or seasonally,
but their storm depths have small distribution indices (o ), so that from
Eq. 6.15 the variance of annual or seasonal precipitation is high, and thus the
soil moisture at the start of an arbitrary interstorm period may be much less
than the long-term average, so .
One of the more useful features of Eaglesons (1978f) mean annual water balance
is its extension, for cases of small variance (P2 ), to the exploration of the annual water
balance (Eagleson, 1978f). However, as might be expected, when this variance is not
small the stress situation described above will occur and in a dry year Eaglesons
(1978f) annual water budget cannot be balanced. Because the condition of incipient
stress lies at the critical boundary between high productivity and disaster for the plant,
it is important that our water balance formulation at least be robust there even if not
terribly accurate. We will approach this as did Salvucci (1992) by incorporating the
first of the above two possibilities (the simpler of the two and not necessarily the
most important) in our analysis, but before doing this we must look at the bare soil
evaporation.
Evaporation Evaporation
of surface of soil
Flux rates
retention moisture
Eps
f e*
es
Er
to
Eps
te
fe fe
0
t* t * + to t * + te tb
Time, t
Fig. 6.6. Interstorm evaporation from bare soil. (Eagleson, 1978d, Fig. 2, p. 733; with
permission of the American Geophysical Union.)
(1) the first stage is evaporation of the surface retention, E r , which takes place at
the surface-referenced bare soil potential rate, E ps , until E r is exhausted at time
t = t ;
(2) the second stage is evaporation of exfiltrated soil moisture at the potential rate,
E ps , due to this rate being less than the exfiltration capacity, f e (defined
below), and it continues until E ps = f e at time t = t + to ; and
(3) the third stage is evaporation of exfiltrated soil moisture at the declining rate f e
due to this rate being less than the potential rate, E ps , and it continues until
f e = 0 and evaporation ceases at time t = t + te .
5 Union.)
4
3
2
1
2 1
10 10 1
e (d )
and that of Carslaw and Jaeger (1959, p. 80) and of Crank (1956, p. 256) to find
*d 1
e (d) = 1 + 1.85 (1)n dn , (6.35)
n=1
1.85 + n
E ps
2
E
De , (6.37)
2n 2e so2
while the dimensionless bare soil evaporation effectiveness, E, is
2n e K (1) (1)
E e sod+2 . (6.38)
m E ps
2
The average interstorm bare soil exfiltration is obtained by taking the expected value
of the integral, E sj , of the instantaneous bare soil evaporation rate, es (the smaller of
E ps and f e in Fig. 6.6), over all the possible (exponentially distributed) times between
Bare soil evaporation 185
storms. This is tedious and the somewhat unwieldy result is given in Appendix C as
Eqs. C.1C.3.
We define the bare soil evaporation efficiency, s , after Eagleson (1978d), as
E[E sj ]
s = , (6.39)
1 E ps
and its analytical expression is developed in Appendix C as Eq. C.5. For the simplified
case of zero surface retention (h o = 0), bare soil, and deep water table (w/E ps 1),
Eq. C.5 reduces to Eaglesons (1978d) important function
3
s = 1 [1 + 2E]eE + (2E)1/2 , E , (6.40)
2
which is plotted here, along with its asymptotes, in Fig. 6.8. Approximating Eq. 6.40
by its asymptotes, we see that their intersection, at E = 2/, divides the coupled
soilatmosphere behavior into two distinct zones:
10
E 1/2 3
s = 1 1 + 2 E e + (2E) 2 , E
s = 1
s 1
1/2
E
s =
2
2
Soil controlled Climate controlled
1
10
102 101 1 10
E
Fig. 6.8. The bare soil evaporation efficiency. (Eagleson, 1978d, Fig. 5, p. 737; with
permission of the American Geophysical Union.)
186 Water balance
Equation 6.40 gives a rational physical basis for the natural surface evaporation
function developed empirically by earlier investigators such as Pike (1964) and Budyko
(1974). Furthermore, the use of the asymptote intersection of such a saturating
function to separate regimes of distinctly different behavior, each governed by different
limiting processes, sets the stage for defining the limits to biological behavior in the
remainder of this monograph.
G Moisture-constrained transpiration
Transpiration can continue only as long as there is soil moisture available to the
plant roots. In Fig. 6.9 we sketch the average time variation of the root-zone spatial
average soil moisture, s, from its maximum value, s(0), at the end of the average
recharging interval of storm precipitation, m tr , to its critical minimum value, sc = s5 ,
at the end of the average drying interstorm interval, m tb . As a first approximation,
we assume the decay of s to be linear with time as it would be if soil moisture was
exhausted only by transpiration occurring at the constant unstressed (i.e., maximum)
rate, E v = kv E ps . The spacetime average (i.e., equilibrium) root-zone soil moisture
is then so = soc = [s(0) + s5 ]/2, and the volume of soil moisture, Ve , available for
exchange with the atmosphere (per unit of surface area) during the average interstorm
period is
Ve = 2n e (so s5 ) z r , (6.41)
in which n e is the effective soil porosity, and z r is again the rooting depth.
Following the development of Salvucci (1992), we continue the assumption that
after the surface retention has been evaporated at time t , evapotranspiration (only)
draws down the soil moisture storage, Ve , until that too is exhausted at the time to
stress, ts (cf. Fig. 6.10). That is
ts
Ve = [(1 M) es + M E v ] dt, (6.42)
t
in which the vegetated fraction of the surface is denoted by the canopy cover, M,
transpiring at rate E v , and the remaining fraction, (1 M), is bare soil evaporating
soc
mt s5 mt
r mt r
b
Time, t
Moisture-constrained transpiration 187
Evaporation Transpiration
of surface of soil
Flux rates
retention moisture
k *v Eps
Er
Eps t s = ts t *
t* ts tb
Time, t
Fig. 6.10. Rate of vegetal transpiration. (Salvucci and Eagleson, 1992, Fig. 3.2, p. 59.)
at rate es . Using Eqs. 6.41, 6.29, and 6.32, with the guidance of Fig 6.4, Salvucci
(1992) solved Eq. 6.42 for the time to stress, ts , with the result given in Appendix C as
Eq. C.6. The time, ts , during which the vegetation transpires, is that between the evap-
orative exhaustion of surface retention at time t , and the time, ts , when stress sets in.
That is
ts ts t . (6.43)
In the manner introduced by Salvucci (1992), we are now prepared to take the
expected value of the integral of the rate, E vj , of stressed interstorm transpiration over
all the initial surface retention conditions (resulting from the gamma-distributed storm
depths), and over all the possible (immediately following) exponentially distributed
interstorm periods, i.e.,
E[E vj ] = dtb E vj (h, tb ) fh,tb (h, tb ) dh, (6.44)
in which fh,tb (h, tb ) is the joint distribution of storm depth and the immediately fol-
lowing interstorm interval. Assuming these to be independent random variables, we
use Eqs. 6.2 and 6.8 to write
o (o h)o 1 eo htb
fh,tb (h, tb ) = fh (h) ftb (tb ) = , (6.45)
(o )
Ev Ev kv E[E vj ]
v = = = 1 , (6.46)
E pv kv E ps kv kv E ps
188 Water balance
Salvucci (1992, Table 3.2, p. 62) computed ts /m tb for a range of soil types and
plant coefficients at Beaver Creek, Arizona, and found ts /m tb 1 for all but the clay
soil (with kv = 1). However, because of the exponential distribution of tb , there will
always be some interstorm periods during which ts = ts < tb , and v will begin to
fall below unity before ts = ts = m tb . As mentioned above in Section E, the primary
cause of v < 1 is the variance of the initial soil moisture, so . Note that with a pdf
of so , we could derive the distribution of ts from Eqs. 6.43 and C.6, and thus incorporate
the variable moisture storage at the beginning of the interstorm interval as the important
origin of transient stress. We will not attempt that complicated refinement in this
work.
This completes the transpiration analysis. Using Eqs. 6.39 and 6.47, the integrand
of Eq. 6.42 can be expanded to write the expected value of the total annual evapotran-
spiration, E T A , (bare soil evaporation plus vegetal transpiration) as
m E ps
ET A = [(1 M) s + Mkv v ]. (6.49)
We recall that E ps was defined in Appendix B as an instantaneous rate keyed to the daylight hour
average insolation. Hence, in using E ps to calculate interstorm evapotranspiration volume we
must use only the daylight hour fraction of the interstorm period, 1 .
Evaporation from surface retention 189
and occurs only during the rainy season, , having mean value, m . The capillary rise
rate is
3/2 (1) mc
w = K (1) 1 + , (6.51)
mc 1 zw
and occurs throughout the entire year, TA . In the above, z w is the depth to the water
table.
In the above, the notation (i.e., subscripts) was chosen to call attention to the fact that
the long-term averages are taken for only a fraction of the year, the growing season of
mean length, m . There will normally be some difference, S (cm), in the stored soil
moisture at the two ends of a partial-year season. The major difficulty in this use of
the long-term water balance is estimating the magnitude and sign of S. The storage
change is placed on the left-hand side of the equation as in Eq. 6.53 where it will be
positive when there is carryover storage at the start of the growing season and negative
when there is carryover storage at the end of the growing season. All quantities in
Eq. 6.53, subscripted or not, are determined for the selected season.
Using Eqs. 6.26, 6.27, 6.49, 6.50, 6.51, and 6.52, and dividing through by the mean
seasonal precipitation, Eq. 6.53 gives the normalized seasonal water balance for the
common case o mLht h o < 1 to be
ho S m t E ps m K (1) c
1 eG2 [(1 M)s + Mkv v ] +
3/2
+ = b so
mh m mh mh P
mc
m K (1) 3/2 (1)
1+ , (6.54)
P mc 1 zw
which expresses the equilibrium relationship among the climate, soil, and vegetation
for a deep water table.
For the rarer case, o mLht h o 1, in which interception capacity equals or exceeds
rainfall, there will be no rainfall input to the substrate of the vegetated fraction, and the
second term on the left-hand side of Eq. 6.54 must be multiplied by the factor 1 M.
Equation 6.54 is a complex function of the equilibrium soil moisture state variable,
so , in which different physical processes (i.e., different terms in the equation) dominate
in different ranges of so . A general analytical solution for so is unavailable. However,
in keeping with our treatment in this work of other complex saturating functions
such as bare soil evaporation (Fig. 6.8) and photosynthetic capacity (Fig. 8.7), we can
solve simpler limiting cases as is detailed in Appendix C. For the simplest case of a
very deep water table, surface retention that is small with respect to the rainstorms,
and negligible carryover storage, Eq. 6.54 gives
m tb E ps m K (1) c
1 eG2 [(1 M) s + Mkv v ] +
3/2
= so . (6.55)
mh P
Salvucci and Entekhabi (1995, corrected 1997) modify Eq. 6.54 to account for water table
influence.
Vegetation statespace 191
1.0
Sum of mean annual bare soil
evaporation, transpiration,
0.8 groundwater and surface runoff,
i.e., yield + evapotranspiration
0.6
Mean annual
transpiration
0.4
Mean annual
Mean annual bare
surface runoff
0.2 soil evaporation
s0
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Effective soil saturation, s
Fig. 6.11. Components of the annual water balance. Fixed climate, soil, and
vegetation. (Salvucci and Eagleson, 1992, Fig. 2.3, p. 38.)
It is common to speak of the long-term average water yield, Y , defined as the sum
of the surface and groundwater components of runoff. From Eq. 6.55 this is
m tb E ps
Y = P [(1 M) s + Mkv v ], (6.56)
mh
while the normalized evapotranspiration is
E[E T ] E[E T ]
= = . (6.57)
E[E ps ] m m tb E ps
The separate terms of Eq. 6.55 are plotted in Fig. 6.11 for a typical humid climate
with clay soil and representative values of M and kv . Summing all the normalized
terms to unity locates the equilibrium soil moisture, so .
1.0 1.0
Mo
0.8 CL L
0.8
Mo CL
0.6 SI-L
0.6
so so SI-L
0.4 Mo 0.4 Mo
SA-L
SA-L
0.2 0.2
CL CL-L
SI-L
PA
SA-L
0.9 0.8 mP = 1 Infeasible
CL-L A SI-L for
-L clay-loam
0.8 0.6 Mo SA -L
PA
CL
mP
=1
Mo
0.7 A 0.4 Infeasible for clay
0.6 k v* = 1 0.2 k v* = 1
CL (a) CL (b)
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
M M
Fig. 6.12. The maximum soil moisture condition under average annual precipitation.
(a) Clinton, MA: climate controlled; (b) Santa Paula, CA: soil controlled. CL, clay;
CL-L, clay-loam; SA-L, sandy-loam; SI-L, silty-loam. (Eagleson, 1978f, Figs. 3 and 4,
p. 755; with permission of the American Geophysical Union.)
The two components of Eq. 6.49 are each plotted dimensionlessly in Fig. 6.13 for
two contrasting climatesoil systems under the condition of unstressed transpiration
(v = 1) and for two values of kv . Note first that the unstressed transpiration is linear
in M regardless of climate and soil. However, the bare soil evaporation will be linear
in M only when the soil properties do not influence the evaporation rate, that is in
climate-controlled situations where bare soil evaporation is limited only by the atmos-
pheric vapor transport capacity. Figure 6.13 clearly shows the Clinton climatesoil
system to be climate-controlled by this criterion. As we move to drier, soil-controlled
climates such as Santa Paula, the bare soil evaporation becomes progressively more
non-linear. This non-linearity is due to the dependence of the desorption velocity upon
the soil moisture which is in turn influenced by the M-dependent extraction of soil
moisture by the plant roots. The non-linearity produces the minimum in total evapo-
transpiration which is marked in Fig. 6.12 by the plotted circle at the canopy cover
labelled M = Mo . Looking now at the soil moisture in the upper part of Fig. 6.12, at
small but non-zero M, where bare soil desorption is the dominant mode of soil moisture
depletion, the composite evapotranspiration rate will be reduced over that for barren
soil yielding higher equilibrium soil moisture. As M increases, the vegetated frac-
tion, transpiring at the potential rate, becomes dominant and the composite evapo-
transpiration rate begins to increase, thereby reducing the equilibrium soil mois-
ture. Eagleson (1982) showed analytically that the maximum soil moisture and the
minimum evapotranspiration occur at the same M = Mo as long as M and kv are
independent.
Vegetation statespace 193
0.4
0.2
M k v* M k v*
M0
0
0.2 0.6 1.0 0.2 0.6 1.0
M M
Santa Paula, CA
1.0
k *v = 0.8 k*v = 1.0
v = 1 v = 1
0.8
0.6
Mk v*
Mk v* 0.4
(1 M ) s
(1 M ) s
0.2
M0 M0
0
0.2 0.6 1.0 0.2 0.6 1.0
M M
(a)
(b)
(c)
several forested watersheds, in which the soil properties and the plant coefficient were
estimated rather than observed, gave reasonable support to the short-term equilibrium
hypothesis. However, the work of Salvucci and Eagleson (1992) suggests a different
basis for M as we will see later in this chapter.
Summary
Equation 6.54 or its three approximate solutions given in Appendix C describe all
the possible soil moisture states in a given climate and soil for a canopy of given M
and kv . However, not all of these states are feasible, for we have seen in Section E
of this chapter that there is a limiting soil moisture potential below which the canopy
conductance, kv , declines from its open stomate value, kv , due to stomatal closure. As
we will now see, incorporating the idea of this critical moisture state (cf. Salvucci,
1992) leads to an additional, independent expression for so .
Rodrguez-Iturbe et al. (1987, 1988) infer that these two variables are essentially independent.
Evaporation from interception truncates the distribution of rainstorms incident upon the soil.
This carryover moisture storage becomes important in seasonal climates.
Table 6.3. Estimated and observed water balance parameters: Beaver Creek, Arizona
1 PinyonUtah juniper woodland 0.40 205 0.35 23.7 0.17 5.33 0.12 0.96 0.61 0.1 18.5 0.48 7.13 4.51 0.46 17.19
2 PinyonUtah juniper woodland 0.42 202 0.35 20.2 0.51 6.20 0.12 0.72 0.45 0.1 19.3 0.48 7.13 4.51 0.46 17.19
3 PinyonUtah juniper woodland 0.40 238 0.34 25.9 0.76 6.06 0.14 0.80 0.53 0.1 17.7 0.48 7.13 4.51 0.46 17.19
4 Pinyonalligator juniper woodland 0.52 181 0.33 24.6 1.02 5.69 0.12 0.62 0.49 0.1 16.4 0.48 7.13 4.51 0.46 17.19
5 Pinyonalligator juniper woodland 0.53 181 0.32 24.6 0.95 5.69 0.12 0.62 0.49 0.1 15.8 0.48 7.13 4.51 0.46 17.19
6 Pinyonalligator juniper woodland 0.52 184 0.32 22.6 0.11 5.85 0.12 1.0 0.73 0.1 15.5 0.48 7.13 4.51 0.46 17.19
718 Ponderosa pine 0.85 169 0.31 22.3 0.67 5.32 0.12 0.56 0.41 0.1 14.2 0.71 5.8 92.8 0.37 15.12
f
Watershed Years of md E psd Pd Yd S mhd ho
c ho S 3/2
number Tree type record (day) (cm day1 ) (cm) (cm) (cm) Eo m (cm) L t e G(0) (0) (cm) mh mv mh
eG(0)2 (0)
1 PinyonUtah juniper woodland 195862 160 0.14 22.0 0.78 7.13 3.02 38.5 0.62 2.27 1.63 1.05 0.33 0.53 0.30 0.04
2 PinyonUtah juniper woodland 195880 163 0.13 23.6 2.96 6.11 2.80 32.6 0.62 2.27 1.63 1.07 0.34 0.55 0.30 0.02
3 PinyonUtah juniper woodland 195867 127 0.13 18.1 1.55 5.14 2.91 39.3 0.66 2.27 1.81 1.08 0.33 0.50 0.20 0.02
4 Pinyonalligator juniper woodland 195872 184 0.14 28.4 11.3 3.16 3.10 31.8 0.77 2.58 1.38 0.91 0.44 0.57 0.13 0.04
5 Pinyonalligator juniper woodland 195872 184 0.14 28.4 11.0 3.97 3.20 31.8 0.77 2.58 1.38 0.91 0.44 0.57 0.16 0.04
6 Pinyonalligator juniper woodland 195964 181 0.14 23.8 3.84 6.96 3.15 31.5 0.72 2.58 1.46 0.96 0.44 0.61 0.31 0.04
718 Ponderosa pine 195880 196 0.08 39.1 12.8 21.0 11.9 31.8 0.70 1.77 1.29 2.12 0.48 0.69 0.94 0.00
a b c d
Williams and Anderson (1967). Appendix B, Table B.2. m = m /m tb . m h = P /m .
e
From Baker (1950) as given here in Table 3.9 using Bakers Northern white cedar for pinyonUtah juniper and Bakers lodgepole pine for pinyonalligator juniper as suggested by
similarities of needle structure as sketched in Zim and Martin (1987, pp. 25, 36, 37, and 39).
f
From Eq. 6.52 using: h o = 0.1 cm; o = 2.50 (Waring, 1983).
Source: Modified from Salvucci (1992).
Critical moisture state 197
(i.e., open-stomate) rate for as long as possible; thus we estimate the minimum (i.e.,
limiting) soil moisture volume as that necessary to maintain open-stomate transpiration
throughout the mean interstorm period. Furthermore, we will show that this limiting
soil moisture volume is insensitive to the value of sc and hence is species-dependent
only through h o and kv .
Data base
It now becomes helpful to have a data base with which to evaluate comparatively terms
of the water balance and against which to test our hypotheses. Jasinski and Eagleson
(1990) located a group of forested watersheds at Beaver Creek (near Flagstaff), Arizona
having 5 to 25 year hydrologic records (Baker, 1982, 1986) as well as a soil survey
(Williams and Anderson, 1967), and canopy cover observations (Clary et al., 1974;
Baker, 1986). The composite data (after reduction by Salvucci, 1992) are summarized
here in Table 6.3, and the soil data are also compared with the properties of typical
soils in Table 6.2b.
Denition
Using the climate and soil parameters of a typical pinyonalligator juniper watershed at
Beaver Creek, Salvucci (1992) solved the simplified water balance equation (Eq. 6.55)
repetitively by trial for the equilibrium soil moisture, so , using a different assumed pair
of the vegetation state variables, kv and M, for each solution. From this field of feasible
solutions he interpolated the isolines of so shown in Fig. 6.14a. This is an expansion into
the full, two-dimensional, statespace of the marginal, one-dimensional explorations
made earlier by Eagleson (1978f) and sampled here in the upper portion of Fig. 6.12.
Two important features of Fig. 6.14a should be noted.
(1) The locus of sMo |kv = 0, i.e., the soil moisture ridge hypothesized by
Eagleson (1978f) as the optimal operating point, is indicated by the dashed
quasi-hyperbola, so max .
(2) Somewhere between the so = 0.34 and so = 0.14 contours, the soil moisture
declines abruptly as the unstressed transpiration of this model dominates the
extraction of soil moisture. The locus of this soil moisture decline is so
topographically abrupt, particularly in arid and semi-arid climates, that
Salvucci (1992) dubbed it the cliff. This cliff is illustrated for the clay and
clay-loam soils of Fig. 6.12b, and for a typical Beaver Creek watershed in
Fig. 6.14b. In the latter illustration we have indicated the critical soil moisture
at which the abrupt decline begins as so = soc = 0.15 which will be estimated
in the next section.
The availability of this unusually comprehensive data set prompted Salvuccis (1992) test of
Eaglesons (1982) optimal vegetation hypotheses.
198 Water balance
0.6
k v* = 0.60
0.4
so
0.2 estimated soc
0
0 0.2 0.4 0.6 0.8 1.0
(b) M
1.0 1.0
0.8 0.8
v k *v
0.7 mh
1-3 4-6
0.6 0.6 mt Eps
k v*
b
0.5
0.4 0.4
(so)max
0.3
0.2 0.2
0.1
0
0 0.2 0.4 0.6 0.8 1.0
(c) M
Eagleson (1978c, p.727) assumed the rooting depth, z r , to equal the so-called pen-
etration depth, z i , of the soil wetting processes, which he estimated to be the sum
of gravitational percolation and diffusive sorption. Using the results of Appendix C,
As we will see later, this is confirmed for carbon demand-limited productivity by Eq. 10.18.
Table 6.4. Water balance summary for Beaver Creek Watersheds
a
Bedrock is at a nominal depth of 1 meter and thus the soil moisture profile does not correspond with that for a water table at z w = , the condition for which terms 6 and 7
(cf. Eq. 6.65) were derived. We assume rather arbitrarily here (after Salvucci, 1992) that term 7 is equal (and opposite) to term 6.
b
Estimated using Eqs. 7.31 and 7.14. Columns (1)(8) Averaged from Table 6.03; (9) Eq. 6.63; (10) Eq. 6.64; (11) and (12) h o = 0.1 cm; o = 2.50 (Waring, 1983); L t = 4.0
(average of all pines, Table 2.02); (13) Eq. 6.30 using sc = 5 bar; (14) Eq. 6.62; (15)(18) Eq. 6.65.
Critical moisture state 201
Section E, for the diffusive component and Eq. 6.50 for the percolation, we have
1/2 m tr K (1) soc
c
z r = z i = 1.77 m tr Di + , (6.60)
ne
Using Sc = 5 bars (cf. Section E of this chapter), the critical soil moisture, s5 , is
calculated for the Beaver Creek watersheds from Eq. 6.30 and is listed in Table 6.4
along with the resulting soc .
in which we assume the collective transpiration rate to increase linearly with time as
leaf surface is exposed, and
ho
m tb = m tb . (6.64)
E ps
This is of course a physically unrealistic artifact of replacing random variables by their time
averages for use in a non-linear equation, and demonstrates that it is the positive excursions of
this wetting from their mean that generate groundwater recharge.
Note that these are gross times in that they include the hours of darkness during which
evapotranspiration is negligible, but E ps must be the instantaneous value derived in Appendix B.
Because the leaves dry from the top of the canopy downward, we start canopy transpiration when
one-half the stored moisture has evaporated.
202 Water balance
Incorporating these times into the water balance, Eq. 6.54, we obtain
mh ho S m t E ps
(1 M) s e+G2
3/2
M kv v 1 + b ,- .
+ ,- . m tb E ps m m m m
1
+,-. + ,- . +
h h h
,- . 5
2 3 4
mc
m K (1) c m K (1) 3/2 (1)
so + 1+ . (6.65)
mc 1
.
P P zw
+ ,- . + ,-
6 7
The numbered terms of Eq. 6.65, normalized where appropriate by the mean storm
depth, m h , are:
(1) the canopy moisture flux; we assume open stomates giving v = 1, and kv for
the ponderosa pine watersheds is estimated from the internal canopy
optimization presented in Chapter 7 (cf. Figure 7.4),
(2) spacetime average surface retention; estimate h o using Eq. 6.52,
(3) seasonal climate carryover soil moisture storage; estimate from site-specific
data,
(4) bare soil evaporation; estimate s using Eq. 6.40,
(5) surface runoff; estimate and G using Eqs. 6.23 and 6.24 respectively,
(6) deep percolation; estimate from site-specific data,
(7) capillary rise; estimate from site-specific data.
Terms 4 to 6 of Eq. 6.65 contain the equilibrium soil moisture, so = soc , which we
estimate using the independent Eq. 6.62. The resulting magnitudes of these terms for
the Beaver Creek catchments are given in Table 6.4. Note that the sum of terms 4, 6,
and 7 is less than 3% of the normalized storm depth, m h /m h = 1, allowing us to write,
for v = 1,
Ve
Mkv = , (6.66)
m tb E ps
o L t h o
where for mh
< 1,
ho G(0)2 (0)3/2 S
Ve m h 1 e + , (6.67)
mh m mh
o L t h o
and (cf. Section J of this chapter) for mh
1,
ho G(0)2 (0)3/2 S
Ve m h 1 (1 M) e + . (6.68)
mh m mh
Critical moisture state 203
Note that both terms in the denominator of Eq. 6.66 are evaluated on the basis of
consistent times.
Equations 6.666.68 express the principal role of hydrology in the structuring of
vegetation communities. The right-hand side of Eq. 6.66 uses the seasonal and storm-
scale hydroclimatology along with the soil properties dimensionlessly to define the
available rate of water supply, while the left-hand side expresses the canopy water
demand. We will return to these equations in Chapter 7 to determine separately the
canopy state variables M and kv , and in Chapter 11 to discuss their implications with
respect to climate change.
Plate 6.4. Slash pine. (Photographed by Peter S. Eagleson, Jr. in dry, sandy soil near
St. Petersburg, Florida.)
204 Water balance
The quantities Pd and Yd were observed (Baker, 1982, 1986) along with the climatic
variables needed to estimate E psd and the season length, m d , and they are tabulated
here in the lower half of Table 6.3.
An attractive approximation to Eq. 6.66, strictly applicable only in very dry, non-
seasonal climates, but useful later for generalizations concerning habitat is
mh
Mkv , (6.70)
m tb E ps
which is plotted on Fig. 6.14c as the topmost hyperbola using climate and soil param-
eters approximating an average pinyonalligator juniper watershed at Beaver Creek.
Equation 6.70 is seen there to give a reasonable representation of the critical moisture
condition for such canopies at least in the semi-arid conditions for which the current
approximations apply. In the highly seasonal case of the ponderosa pine canopies,
the carryover storage is large (see Table 6.3) and plays an important role in support of
vegetation growth, but is difficult to estimate accurately.
The lowermost hyperbola of Fig. 6.14c is a reproduction of the maximum soil
moisture ridge line of Fig. 6.14a. The two plotted points are the estimated canopy
Walter (1962) states In effect the density of the vegetational cover is proportional to the
precipitation . . . .
Critical moisture state 205
d
ite
lim
M k *v
ly
pp
su
er
0.4
at
w
0
0 0.2 0.4 0.6 0.8 1.0
Ve
mt Eps
b
states for the two groups of pinyonjuniper watersheds as given by Salvucci (1992).
Note that they lie much closer to the cliff than to the ridge. The two dashed
hyperbolas on Fig. 6.14c represent the locations of the ridge and the cliff following
the above-mentioned modifications in the water balance introduced by Salvucci (1992)
to accommodate the geological peculiarities of these particular watersheds.
Finally, the critical moisture state hyperbola, Eq. 6.66 and its approximation,
Eq. 6.70, are compared, in Fig. 6.15, with observations from the three groups of Beaver
Creek watersheds (two dry pinyonjuniper groups, watershed numbers 13 and 46,
and one dry ponderosa pine group, watershed numbers 718). The relevant climate,
soil, and vegetation parameters are summarized, with their sources, in Table 6.4 along
with the estimated components of the available soil moisture Ve /m h . Two conditions
of this comparison should be noted:
(1) Observations of M (Salvucci, 1992 as listed in Table 6.3) and estimation of kv
(Eqs. 7.4, 7.14, and 7.31) allow comparison of Mkv with Eq. 6.66 (plotted
circles) and its approximation Eq. 6.70 (plotted triangles) for the two groups of
pinyonjuniper watersheds as well as for the ponderosa pine watersheds. The
product, Mkv , of the two vegetation state variables fixes the ordinate for these
observations while the abscissa is calculated from properties of the water
environment using Eqs. 6.66 or 6.70.
(2) o = 2.50 for all pines as suggested for ponderosa by Waring (1983).
As we will see in Chapter 8, there is an upper limit to Mkv for a given species at a
given temperature beyond which the demand for water rather than the supply of water
is constraining. This limit is (kv ) M=1 kv1 and occurs, in a given soil, at the available
moisture conditions, Ve , which produce a closed canopy. For larger Ve there is no
206 Water balance
change in Mkv . This is indicated in Fig. 6.15 by the horizontal lines for the different
species.
While the sample is small, the observations presented in Fig 6.15 strongly support
Eq. 6.66 (plotted circles) as defining the operating state of these water-supply-limited
canopies provided that we acknowledge an underrepresentation of the surface runoff
(and hence a slight overestimation of Ve ) in the wetter climates due to our expe-
dient choice of exclusively Hortonian runoff. The approximation, Eq. 6.70, shown
on Fig. 6.15 as the plotted triangles, is inadequate qualitatively as well as quan-
titatively in this highly seasonal climate due to its additional neglect of carryover
storage.
In evaluating Fig. 6.15 we must remember that our analysis of a highly variable
semi-arid climate has been based upon mean values producing the critical soil moisture.
This idealization can distort reality in a significant way as is easily shown: Suppose
the plant can adapt to a certain mild level of water stress. Under water-limiting cir-
cumstances the community of such plants can then develop a greater canopy cover
than if no stress was tolerable and may generate higher production. This path to higher
productivity seems certain to ensure that in the high variability of semi-arid climates
some species will develop a measure of water stress tolerance.
M Conclusions
We have seen that the concept of critical soil moisture potential, c , is essential to
understanding the limitations on stressless plant transpiration in a given climate, but
we have also seen that its magnitude affects only the negligible terms of the water
balance (i.e. terms 46 of Eq. 6.62) and hence is irrelevant to estimation of the canopy
mass flux capacity, M kv E ps . Assuming adequate nutrition, we draw the important
conclusion that species habitat is determined by parameters of the climate and not of
the soil.
Conclusions 207
It is important for the work of later chapters to note that the remarkable validation of
Eq. 6.66 presented in Fig. 6.15 may be interpreted as confirmation of our hypothesis that
natural selection seeks stresslessly to maximize the productivity (cf. Appendix E) of
biomass for a given species in a given climate. This is demonstrated by the adequacy of
the canopy mass flux capacity to deliver the available moisture to the free atmosphere
in the time available. Indeed, Fig. 6.15 shows the canopy mass flux capacity, M kv E ps ,
to be tuned to exactly match the rate, Ve /m tb , required for moisture exhaustion to the
point of incipient stress. Therefore, while water resource-limited by virtue of M < 1,
these systems are not limited by their moisture flux capacity. It seems reasonable to
conclude that this is no accident, and that natural selection in the given environment has
fixed and L t at the appropriate values to bring about this productivity maximization.
We return to this conjecture in Chapter 7.
This confirmation of the details of our water balance formulation gives reasonable
confidence in its later use in testing the important habitat hypotheses of Chapter 9.
Part II
Darwinian ecology
7
209
210 Optimal canopy conductance
clearly as follows:
The optimal density of leaves is that which just reduces the light intensity below
the tree to the compensation point; as long as there is more light, additional leaves
would add more by photosynthesis than they would subtract by respiration, but
once the light reaches the compensation point, additional leaves respire more than
they photosynthesize.
We saw in Chapter 3, that for homogeneous leafy canopies, the so-called Monsi
Saeki relationship for radiation extinction is
I ( )
= exp [ L t ], (7.1)
I0
in which I0 is the insolation at canopy top, I ( ) is the insolation transmitted to level
of the canopy, and is an empirical decay or extinction coefficient which is dependent
primarily upon the quality of the light, the extent and arrangement of the foliage
elements, and the solar elevation. At the base of the crown, = 1, Horns (1971)
argument as quoted above puts
I ( = 1) = Ik , (7.2)
where Ik is the compensation insolation, making the special case of Eq. 7.1
Ik
Iko = exp [ L t ]. (7.3)
I0
Values of Iko determined in laboratory experiments using artificial light with different
tree species are listed in Table 3.9 as taken from Baker (1950, Table 12, p. 143). Using
the condition of optical optimality from Chapter 3, i.e., that = , these tabulated
values define the desired species-specific horizontal leaf area index, L t , as
1
L t = L t = ln o , (7.4)
Ik
The species specificity of L t is demonstrated in Appendix H.
212 Optimal canopy conductance
which are plotted as the solid hyperbolae in Fig. 7.1. It should be noted that Baker
(1950, p. 143) considers these values to be inaccurate (Iko too high) due to conditions of
the experiment. Field observations of L t , , and have been presented in Tables 2.2
and 2.3 for various tree species. The average of these observed values is plotted in
Fig. 7.1 in comparison with the experimental L t = L t curves for each of the species
considered by Baker (1950) and is keyed to the Baker curves by the associated number
in Table 3.9. Agreement of the values obtained by the two methods is obtained only
for pine.
rc rci , (7.5)
1 1 1 1 1 1 L*t 1
1
= + + + + = . (7.6)
rc ri 2ri 3ri (L t 1)ri ri n=1 n
*
n
1 1 *
Ak
= h + ln (n) + , (7.7)
k=1
k 2n k=2 n(n + 1) (n + k 1)
In particular, it is not clear whether the conditions of the experiment included a different full
sunlight, I0 , representative of the habitat of each species, or used a common full sunlight such
as the 107.6 klx defined by Gates (1980, p. 94). In addition, the compensation values may be
sensitive to climate.
Given the biological requirement for compensation light intensity at the lowest leaf (cf. Section A
of this chapter), homogeneous canopy foliage area distribution is realistic only for cylindrical
crowns. For tapered crowns the vertically averaged foliage density must increase with radius from
the tree axis.
This feature distinguishes the formulation of canopy water vapor efflux from that of canopy CO2
influx. The stomatal source of water vapor is at saturation vapor pressure which is constant at all
elevations in isothermal crowns, while the CO2 sink concentration within the stomata varies with
depth in the crown in a manner that is dependent upon the foliage density and upon the
concentration needed at the lowest leaf to support compensation-level photosynthesis there. The
difficulty of estimating the gradients driving the canopy fluxes motivates and justifies the use of
the big leaf approximation of the canopy intoduced for the derivation of canopy conductance in
Chapter 5.
Resistance ratio for closed canopies 213
ra
z=h =0
ri ri ri n=1 ri
ri ri n=2 ri
ri n=3 ri
n = L t 1
ri
z = hs = h hs
(a)
ra
ri 1
ri ri ri 2
ri ri ri ri 3
ri ri 4
ri L t 1
(b)
For the homogeneous cylindrical multilayer, ri is given by Eq. 4.167, and using
Eq. 7.9, Eq. 7.11 becomes
1
[(1 m)/m] ( L t )2
ku
rc = .
1m 1 1
n s (L t 1) 1 exp Lt 0.577 + ln (L t 1) + (L t 1)
m 2
(7.12)
Dividing Eq. 7.12 by Eq. 4.143, and using Eq. 4.59, we finally have the important
resistance ratio for the homogeneous cylindrical crown
rc [(1 m)/m] ( L t )2
ns = .
ra 1m 1
(L t 1) 1 exp Lt 0.577 + ln (L t 1) + (L t 1)1
m 2
(7.13)
Equation 7.13 is presented graphically by the solid lines in Fig. 7.3 for leafy
plants (i.e., m = 1/2, n = 2) over the common range of (see Fig. 3.18). We note a
0.8
d
0.6 0.3
0.4
0.4
= 0.25
0.2 a
0
0 2 4 6 8 10
Lt
Resistance ratio for closed canopies 215
minimum in the resistance ratio of n src /ra = O(1) at 4 L t 6. This minimum n src /ra
translates, through Eq. 5.29, to a maximum in the evaporative flux of water vapor at the
given temperature. The cause of this minimum in the resistance ratio can be seen from
the behavior of Eq. 7.11. The denominator, n s N (L t ), increases monotonically with L t ,
so the interesting behavior lies within the numerator, ri , as given by Eqs. 4.165 and
4.167:
For low L t , dL
drc
t
is controlled by the change in the dimension, z L , the numerator
of ri , which declines rapidly with increasing L t giving dLdri
t
= (L t 1) 2.
For high L t , dL t is controlled by the change in the eddy viscosity, K m (z), the de-
drc
= (L
L t (L t 2)
2
dri
nominator of ri , which declines rapidly with increasing L t giving dL t t 1)
.
We will return to further discussion of the physical basis for this interesting behavior
in Chapter 10.
In conjunction with Eqs. 5.29 and 5.30, we see that this minimum in n src /ra maxi-
mizes the canopy conductance at constant temperature.
Equation 7.14 is plotted as the maximum vapor flux curve in Fig. 7.4 for the
leafy plant case of m = 1/2, n = 2.
In Chapter 10 we will see that the shape of this function also leads, for large L t , to a
growth-limiting constraint on the flux of atmospheric carbon to the leaves.
216 Optimal canopy conductance
Oak
Beech
Sugar maple
L 6
t
Observed
range
Observed
mean
r
Maximum vapor flux (minimum c )
ra
2 at constant temperature with M = 1
0
0 0.2 0.4 0.6 0.8 1.0
or
Once again, Eqs. 7.4 and 7.14 together determine and L t for a given species in
a given light environment. Their joint solution is illustrated graphically in Fig. 7.4 for
four species covering the full range of those reported by Baker (1950) and repeated
here in Table 3.9. We refer to this foliage state as optimal because of the following
accompanying flux conditions:
(1) The radiant energy absorbed is the maximum possible (i.e., the crown is
optically optimal) for larger L t the lower leaves will receive insufficient
light, while for smaller L t unused energy will be transmitted through to the
forest floor.
(2) The flux of transpirate from the soil to the atmosphere, and with it the flux of
nutrients from the soil to the plant, will both be maximum for a given
temperature because the minimization of rc /ra accompanying this joint
solution assures maximization of the normalized canopy conductance, kv
(cf. Eqs. 5.29 and 5.30).
We wish to know if nature selects for this multiply optimum state. However, we are
assuming natural selection to favor increased annual productivity which depends upon
factors other than just the average canopy conductance. We return to this question in
Chapter 10 and for now offer only the following observations relative to this theoretical
optimum.
There is evidence (Baker, 1950, pp. 141142) that these values may be climate-dependent.
This statement is strictly true only for the limiting, M = 1 case. As we will see later in this Chapter,
for M < 1 we must also specify the moisture supply and the geometry of the individual plant.
Resistance ratio for closed canopies 217
Average 0.73
a
From transpiration measurement normalized by that from short, green grass.
b
Calculated from observations of net primary productivity.
c
Assuming M = 0.84 (maximum packing of circular discs).
d
Assuming M = 0.50 (estimated from Eyre, 1968, Plate IIIA).
e
From transpiration measurement normalized by that calculated from the heat balance of a
wet surface.
Resistance ratio for closed canopies 219
We conclude from this analysis that at typical temperatures, the reported values of
kvfor trees cannot be achieved as modeled with rc /ra = O(10) but can be achieved
with rc /ra = O(1). The deciding element, missing from our formulation, seems to be
advection and hence the value of the Bowen ratio, Rb , at the time of the observations
of either kv or rc /ra . Equations 5.29 and 5.38 give the rough estimate
rc 1
= 1 1+ |Rb | 1 + , (7.15)
ra kv o o
which shows that at typical temperatures, the Bowen ratio must be greater than unity
and considerable advected energy is required for the resistance ratio to be of order ten.
Equation 7.16 (using Eq. 7.17) is plotted in Fig. 7.5 where we see that in contrast
to the multilayer cylinder there is no minimum in the resistance ratio of the multilayer
hemisphere. As was pointed out in Chapter 4, this is due to the monotonic increase in
mo , with increasing L t , brought about by the open (i.e., rough)
the eddy viscosity, K
surface of tapered-crown canopies.
where the subscript s is added to the resistance ratio to distinguish its density as
sparse in contrast to that of the conical monolayer to be considered next, and in
220 Optimal canopy conductance
0.8 = 0.5
rc
ns
ra
0.6
= 0.5
= 0.3
0.4
0.2
0
0 2 4 6 8 10
Lt
Equation 7.18 (using Eq. 7.19) is also plotted in Fig. 7.5 and shows the same monotonic
behavior and for the same reasons explained above for the hemispheres.
For each species there is a maximum L t for each , imposed by the compensation light
requirement of that species (cf. Table 3.9).
Resistance ratio for closed canopies 221
(1) Water is never truly unlimited and its vertical flux, which rises with L t , is
constrained by the presumed need to avoid water stress.
(2) As we will see in Chapter 10, biomass productivity is indifferent to crown
shape at low L t since there productivity is water-limited. However at large
L t where water is plentiful, CO2 influx becomes limiting for cylindrical
multilayers and the resistive advantages of tapered multilayer crowns do
become important. Examples of the latter case are given by the conically
crowned Japanese larch which is found in wet climates, and by the
hemispherically crowned rainforest that displays a canopy-top looking like
scoops of ice cream (cf. Odum, 1975, Fig. 7-10, p. 197).
(3) As the foliage area increases within a crown of given dimensions, the
increasing crown foliage density causes a fundamental change in the
mechanism of canopy resistance. As was argued in Chapter 4, increasing
foliage resistance to the passage of air laterally through the tapered crowns
causes a greater fraction of the air to flow around the crowns. This increases
the role of crown shape in canopy resistance while decreasing the role of the
foliage elements. A common instance of this occurs in cold climates where
maintenance of biochemically functional leaf temperature becomes important.
Structural tree forms having drooping branches that overlap closely in a
shingled fashion at their extremities may be a selective response to this heat
problem. In such structures, leaves, usually needles because of the singular
ability of needle-leaved species to survive extremely low temperatures, are
confined to the extremities where their proximity forms an insulating layer
restricting the loss of canopy heat (Gates, 1980, p. 364). The density of needles
necessary to achieve the needed insulation is such that solid body resistance is
generated. This high foliage density also causes the available light to decay to
compensation intensity over a very short distance leaving a relatively thin
skin of active vegetation at the crown surface referred to herein as a
monolayer. Of course, such limitation of heat efflux also operates to limit
water vapor efflux as well as CO2 influx, and thereby restricts the supply of
both carbon and nutrients to the plant. We examine the effect of this limitation
on plant productivity in Chapter 10, but for now we are only concerned with
the production of solid body resistance by the monolayer crown, and how this
resistance affects vertical mass flux in the canopy.
Accurate description of the transition, with increasing foliage density, from dynamic
domination by foliage element resistance to that by solid body resistance in a canopy
of tapered crowns is problematic. We base our estimate of this transition upon the
limiting cases of low-density multilayers and high-density monolayers. Having just
examined the lower limit provided by the homogeneous conical multilayer, we now
look at the upper limit given by the high-density conical monolayer.
See for example: Sakai and Weiser (1973), George et al. (1974), and Sakai (1978, 1979).
222 Optimal canopy conductance
Plate 7.3. Fir foliage texture: Douglas fir. (Photographed by William D. Rich in
Arnold Arboretum, Boston; Copyright c 2001 William D. Rich.)
As developed in Appendix A (Eq. A.106) using the definitions of Fig. 7.6a, the
limiting (i.e., solid body) reistance ratio is
rc rc 2 n L t 1
ns = ns = , (7.20)
ra ra d 3 n L t exp (n L t ) K
o
1 m d
1 exp (n L t )
in which the subscript d on the resistance ratio is to signify its dense character
in contrast to that of the homogeneous multilayer discussed previously, and in which
o )d is given by Eq. A.57
(K
Kmo = 0.454 + 0.009L t + 0.002L 2t . (7.21)
d
Equation 7.20 (using Eq. 7.21) is plotted as the upper dashed line in Fig. 7.7 and
Eq. 7.18 (using Eq. 7.19) is plotted as the lower dashed line in the same figure.
Together they represent the upper and lower limits respectively of the resistance ratio
for conical monolayers.
We now estimate a transition function between these limits in terms of the monolayer
foliage density.
z hs
b
(a)
1 a
2
3
L
t
L
t
b
t
c
(b)
224 Optimal canopy conductance
d
=1
t
10
rc
ra d
Norway spruce
d
Tamarack
= 0.033
t
rc 0.017
ns
ra 0.011
1
0.006
rc
ra s d
0 (multilayer)
t
t
0.1
0 2 4 6 8 10
L
t
Fig. 7.7. Theoretical canopy resistance for conical monolayers (M = 1, = 0.50).
d
f = L t. (7.22)
t
Landsberg and Thom (1971) give average diameters for the needles and twigs of
spruce as 1 mm and 3 mm respectively, thus we take d 2 mm. Further, we estimate
Resistance ratio for closed canopies 225
(cf. Landsberg and Thom, 1971, Fig. 1) the diameter of the spruce shoot at 3 cm.
From these dimensions we consider four arbitrary but reasonable monolayer densities,
t = two shoots, d = 0.067; t = four shoots, d = 0.017; t = six shoots,
d/t; t t
d
t
= 0.011, and t = ten shoots, dt = 0.006.
With the definition of Eq. 7.22, the crown surface becomes solid as f 1 making
rc rc
= , (7.23)
ra ra d
and we can then write the general expression for the resistance ratio
rc rc rc
= f + (1 f ) . (7.24)
ra ra d ra s
Equation 7.24 is evaluated using Eqs. A.100 and A.106, and is plotted as the solid
lines in Fig. 7.7 for the four chosen values of dt , and for the representative value
for spruce crowns, = 0.50 (cf. Fig. 3.18). Note that the resistance ratio now has a
minimum indicating a preferred L t in the range 3 L t 6 which is to be compared
with the value L t = 3.3 observed by Landsberg and Jarvis (1973) for Sitka spruce
in northeast Scotland. Note also that over the fairly wide range of the parameters
displayed, the minimum resistance ratio lies in a quite narrow range, 0.7 < rc /ra < 1.4.
The vertical dashed lines on Fig. 7.7 represent the L t for compensation light intensity
at = 0.50 for two densely foliated, conically crowned species, tamarack (i.e., larch)
Plate 7.4. Painted maple. The author alongside maple in Arnold Arboretum, Boston.
(Photograph by William D. Rich; Copyright c 2001 William D. Rich.)
226 Optimal canopy conductance
and Norway spruce as given by the values of Baker (1950, Table 12, p. 143) and
reproduced here in Table 3.9. Note that they fall in the L t range of minimum resistance
for these reasonable monolayer densities.
1
(h h s )
m
rc z/ K 2 , (7.25)
h + hs
u k dos
2
in which dos is the zero-plane displacement of the unvegetated surface. Equation 7.25
simplifies to
hs
1 1
h
rc .
dos
(7.26)
ku hs
1+ 2
h h
where u 0 is the wind speed at z = h. Equations 7.26 and 7.27 give the important
resistance ratio as
hs
1
rc u h
= .
dos
(7.28)
ra M=0 ku 0 hs
1+ 2
h h
The bulk drag parameter, ku u
0
, and the velocity distribution parameter dos /h are de-
pendent upon the substrate and the structure of both the crown and the canopy and hence
are variable with the species, the climate, and the soil. For this case of sparse roughness
elements (i.e., M 1) their estimation is problematic. For M 1, Raupach (1992,
Fig. 5, p. 393) finds empirically that smooth substrates give ku u
0
= 0.137, while forest
substrates maintain the value ku 0 1 found theoretically in Chapter 4 for M = 1. For
u
very sparse roughness elements, the displacement height, dos , of a fully developed
velocity profile must approach zero (Brutsaert, 1982, p. 116), while for homogeneous
canopies of M = 1, the average of many observations on vegetated surfaces gives
do /h = 2/3 (Brutsaert, 1982).
For our current purposes, we assume the bottom shear stress to remain ku u
0
= 1, and
the displacement height to have the value dos /h = 1/3, midway between that for the
smooth substrate and that for the closed forest. With these values Eq. 7.28 becomes
hs
1
rc h ,
hs
(7.29)
ra M=0
0.33 +
h
which is plotted as the solid line in Fig. 7.8 along with a dashed line representing the
limiting curve for d0 / h = 0 (i.e., smooth, bare substrate).
Note that our assumption of reduced effective eddy viscosity makes these estimated
resistance ratios for M 1 much larger than those for the M = 1 case presented in
rc
2.0
ra M = 0
d os
= 0.33
h
1.0
d os
=0
h
0
0 0.5 1.0
hs /h
228 Optimal canopy conductance
This relation is a function of the crown structure through the resistance ratio, r
c /r a ; of
the climate and soil through the canopy cover, M; and of the climatic growing season
temperature through o . It serves to join the internal optimization of the canopy
resistance to the external optimization of the securityproductivity tradeoff as we
will see in the next chapter.
Resistance ratio for intermediate canopy cover 229
Some typical examples of the Mkv relation are given in Fig. 7.9 for broad vege-
tation classes using representative parameter values cited herein and which are listed
in tabular form in the caption. In interpreting this figure, it cannot be overemphasized
that the characteristic shape of the function, that is rising kv with increasing M, is
dominated both in kind and in degree by our crude model of the canopy resistance for
M 1 under which the relative values of h s /h and dos /h fix the M = 0 intercept and
hence essentially determine the slope.
1.0
1
2
0.8 4
3
0.6
5
k *v
0.4
Eq. 7.31
0.2
0
0 0.2 0.4 0.6 0.8 1.0
M
Fig. 7.9. Canopy conductance vs. canopy cover (multilayers). 1, deciduous; 2, loblolly
pine; 3, ponderosa pine; 4, pinyonjuniper; 5, spruce.
M =0 M =1
a h l
See text. Table 3.03. Table 3.09.
b i m
Eq. 7.29. Assumption. Table 3.09, Northern
c j
Fig. 3.18. Christensen (1988, white cedar.
d
= from Eq. 3.52. Figs. 11.4, 11.5). n
Table 3.09, Lodgepole
e k
Fig. 7.03. Williams and Anderson pine.
f
Assumed cylindrical crown. (1967, Fig. 5).
g
Assumed conical crown.
230 Optimal canopy conductance
As we have seen in Chapter 6, the case of M < 1 is one in which available soil
moisture, Ve , determines the transpiration rate. Therefore, an alternative to Eq. 7.31
for evaluating kv in such cases is Eq. 6.66. Of course this doesnt remove the large
uncertainty in estimating kv but merely shifts it from (rc /ra ) M0 in Eq. 7.31 to Ve and
the open-canopy E ps in Eq. 6.66.
Ve
Mkv = . (7.32)
m tb E ps
Equations 7.31 and 7.32 are sketched in Fig. 7.10. If the climate and soil are such that
the soil moisture, Ve , available for stressless transpiration of a given species during the
average interstorm period, produces an intersection with the kv (M) line (i.e., Eq. 7.31)
Comparison of kv as estimated by these two very different methods should be a useful and
informative exercise. We carry it out implicitly in Fig. 7.11.
Cover and conductance 231
at M < 1, i.e., if
Ve 1 + /o
< , (7.33)
m tb E ps rc
1 + /o
ra M=1
the canopy will be water-limited and its optimum Mkv state is as shown in Fig. 7.10
by the intersection of the two functions. On the other hand, should
Ve 1 + /o
, (7.34)
m tb E ps rc
1 + /o
ra M=1
the canopy is light-limited and its optimum Mkv state is M = 1 with
Ve
kv . (7.35)
m tb E ps
Of the Beaver Creek species, only ponderosa pine is tabulated by Baker (1950). We use the
physical resemblance of needle structure (cf. Zim and Martin, 1987, pp. 36, 39) to represent the
pinyonalligator juniper as pinyon using Bakers (1950) L t for lodgepole pine, and to represent
the pinyonUtah juniper as juniper using Bakers (1950) L t for Northern white cedar.
232 Optimal canopy conductance
kv* 0.6
b
(iii) PinyonUtah
juniper
0.4
0 0.2 0.4 0.6 0.8 1.0
M
The final step in resolving the optimal canopy state is to plot, for each of the three
Beaver Creek species, the Mkv hyperbola of Eq. 7.32 representing maximum water
supply. From Chapter 6, this equation expands to
mh ho S
Mkv = (1 M) eG(0)2 (0) +
3/2
1 , (7.36)
m tb E ps mh m mh
the terms of which are quantified in Tables 6.3 and 6.4. These hyperbolae are presented
as curves (b) on Fig. 7.11 and their intersections with curves (a) define the optimal
canopy state.
We have observations only of M (cf. Table 6.3) and these are shown as the vertical
dashed lines in Fig. 7.11. Their proximity to the theoretical optimal state is remarkable
but must be viewed in relation to the sensitivity of the (a)(b) intersection to use of
other than the optimal value of . To do this we return to Fig. 7.3 and for the L t
of the given species (cf. Table 3.9) we pick off the resistance ratio corresponding to
the range of from = 0.25 to = 0.60. This allows recalculation of curve (a)
for this range of (rc /ra ) M=1 , and determines a corresponding range of non-optimal
intersections with curve (b). This last range is highlighted on Fig. 7.11 by the shaded
areas. The optimal state is that of mimimum resistance ratio (and hence highest kv ),
and thus coincides with the lower limit of this range in M. We see that for dry climates
such as the pinyonUtah juniper, the range is quite narrow and non-optimality makes
Note that Fig. 7.11 offers nothing not provided earlier by Fig. 6.15. The data are the same, but the
coordinate axes have been changed for the particular purpose of this chapter.
Cover and conductance 233
Optimal bioclimate
234
Photosynthesis 235
A Photosynthesis
The concept
Photosynthesis is the primary process of life for green plants (and for some bacteria).
It is the assimilative (i.e., anabolic) process through which these primary producers
utilize sunlight to synthesize carbohydrates from CO2 and water. In this process they
must incorporate inorganic nutrients (principally nitrogen) obtained from the soil,
thereby raising them to a higher energy level.
In biophysical and biochemical terms (Gaastra, 1963, p. 114; Gates, 1980, pp. 8,
4748), photosynthesis consists of three sub-processes: (1) a gas-diffusion process
through which the green leaf takes up CO2 from the air and transports it to the reaction
site in the chloroplasts of the palisade cells; (2) a photochemical reaction taking place
within the leaf chloroplasts which converts absorbed light energy into chemical energy
for the reduction of CO2 to carbohydrates; and (3) biochemical enzymatic processes
that both precede and follow the CO2 reduction and are highly temperature dependent.
The photochemical energy conversion can occur through one of several processes
which each release gaseous oxygen while manufacturing amino acids, proteins, and
other biologically useful compounds; they depend upon the availability of light (i.e.,
PAR) and of CO2 . The evaporation of plant water within the leaf stomatal cavities
contributes to the regulation of leaf temperature, and the evaporate diffuses out from
the leaf stomates at a rate which is closely related to that at which CO2 diffuses in
through those stomates.
Simultaneously, by both day and night, within the cells of all the plant organs,
green and non-green, the destructive (i.e., catabolic) process of respiration, sometimes
referred to as dark respiration, utilizes oxygen to break down substances in order
to provide energy for cell metabolism. Respiration releases CO2 which is diffused
outward into the atmosphere. In addition, as a component of leaf photochemistry, there
is release of CO2 in a process called photorespiration or sometimes light respiration.
Measurements of the light-induced CO2 assimilation by leaves necessarily include all
daylight respiration within their estimates of net photosynthesis.
During the day, the rate of CO2 uptake per unit of plant mass required for photosyn-
thesis is normally greater than the rate of release of CO2 in respiration. The resulting
net uptake of CO2 during daylight hours is proportional to and defines the net rate of
photosynthesis and hence plant productivity.
Plants may be distinguished in part by the predominant chemical process or path-
way they use in the fixation of carbon during photosynthesis. These processes are
several in number but the most common are (Osmond et al., 1980):
(1) Dicarboxylic acid pathway (C4 ) in which the uptake of CO2 (and hence
water loss) follows closely the changes in radiation intensity. These plants can
utilize even the most intense solar radiation, and thus their family includes
tropical grasses as well as agricultural plants such as millet, sorghum, and
maize.
236 Optimal bioclimate
30
CO2 ).
Pssm
20 C3
10
I
k
10 30 50 70 90
I(PAR) (klx)
Table 8.1. Observations of the photosynthetic reaction (natural CO2 availability and
optimal temperature)a
Ik b ISL b
(total radiation) (total radiation) Psm c
Plant group (klx) (klx) (mg CO2 dm2 h1 )
Herbaceous plants
C4 plants 13 over 80 3080
C3 plants
Crops 12 3080 2045
Heliophytes (sun) 12 5080 2040
Sciophytes (shade) 0.20.5 510 420
Xerophytes (dry) 2045
Grains and fodder grasses 1535
Woody plants
Tropical and subtropical trees
Fruit trees 1822
Forest canopy trees 1224
Understory trees 510
Broad-leaved evergreens of
warm tropics and
sub-temperate zones
Sun leaves 1018
Shade leaves 36
Seasonally deciduous trees
Sun leaves 11.5 2550 1525
Shade leaves 0.30.6 1015 510
Conifers
Winter deciduous
Evergreen 518
Sun leaves 0.51.5 2050
Shade leaves 0.10.3 510
Sclerophylls 515
Bamboos 510
Palms 610
Desert shrubs 620
Understory ferns 0.10.5 210 35
Mosses and lichens 0.42 1020 0.53
a
This table was prepared before the 1995 publication of Larchers latest edition of Physiological Plant Ecology.
The updated values presented in Larcher (1995, Table 2.4, p. 85, and Table 2.8, p. 98) are only quantitatively
different and do not affect those conclusions of the present work which are based upon the data presented in
this table.
b
From Larcher, W., Physiological Plant Ecology, Table 3.6, p. 104, for single leaves, Copyright c 1983
Springer-Verlag GmbH & Co.; with kind permission from Springer-Verlag GmbH & Co., and from
W. Larcher.
c
From Larcher, W., Physiological Plant Ecology, Table 3.4, p. 94, Copyright c 1983 Springer-Verlag GmbH
& Co.; with kind permission from Springer-Verlag GmbH & Co., and from W. Larcher.
238 Optimal bioclimate
Values for various tree species are given as a percentage of full sunlight by
Baker (1950, Table 12, p. 143) while leaving the term full sunlight
undefined. Ik has an important variation with temperature because respiration
and photosynthesis have different temperature dependencies. No fixed light
value marks the threshold of existence of any species (Baker, 1950, p. 139).
Leaves adapted to shade respire at a lower rate than do those adapted to strong
light, hence the former reach the compensation value at a lower value of the
light intensity (e.g., Larcher, 1975, Fig. 26, p. 41). However, the relative
abilities of different species to survive under shade remains constant from one
light environment to another (Baker, 1950, p. 140). This suggests that the
photosynthetic behavior of a given species is relative to the incident light rather
than being absolute and accounts for the reporting of most compensation
insolations, Ik , as a percentage of the incident insolation, I0 , as we have seen in
Table 3.9. Using results of this and later chapters, we prove in Appendix H that
Ik /I0 is a species constant.
(2) The value, ISL , of insolation at which the net photosynthesis, P, can be
considered to have reached its saturation limit, Psm . ISL is called the (CO2 )
saturation insolation. It too varies from species to species, but because of the
saturating behavior of P, estimation of ISL from observations of P is
problematic. Observations of the ISL are given in Table 8.1. In Table 8.2 we
give the average (from Table 8.1) observed ISL for C3 crops and for the sun
leaves of deciduous trees and evergreen coniferous trees. We will assume that
ISL is the largest light intensity to which the plant has adapted, and therefore in
regions with larger incident radiation we expect other species with higher ISL
to be more productive and thus to predominate.
(3) The light-saturated rate of net photosynthesis, Ps . Figure 8.2 (from Kramer
and Kozlowski, 1960, Fig. 3.11, p. 83, after Muller, 1928, Fig. 2, p. 29) shows
photosynthetic capacity curves for Arctic willow (Salix glauca) for unlimited
water at three values of the ambient temperature, T0 . There we also see that
photosynthesis saturates in weak light at low temperature and in much stronger
light at high temperature. However, for a given light condition the maximum
photosynthetic rate, Ps , does not increase monotonically with ambient
temperature. Rather, it has a maximum, Psm , at an intermediate, optimum,
value of ambient temperature, T0 = Tm , as shown (for unlimited water)
for Pinus cembra in Fig. 8.3 (from Kramer and Kozlowski, 1960, Fig. 3.10,
p. 82, after Tranquillini, 1955, Fig. 6, p.169). Species-dependence of this
optimal temperature, Tm , is illustrated by the relative curves of Fig. 8.4a (from
Larcher, 1983, Fig. 3.35, p. 114), and some reported values of Tm are collected in
Table 8.3. Adaptation of Tm by several degrees Celsius can occur over a few days
in response to changing environmental conditions (Larcher, 1983, pp. 118119).
Using the definition of Gates (1980, p. 94), full sunlight = 107.6 klux, making the values of Ik in
Table 3.9 much higher than those for trees in Table 8.1.
Photosynthesis 239
We have been assuming in this work that the leaf temperature, Tl , is that, Tm , for
which the photosynthetic capacity is the species maximum as is shown in Fig. 8.4a,
and furthermore, that the canopy sensible heat transfer is small enough that Tl T0 .
Under these assumptions,
Average 0.81
(0.78)
a
gs /g = 0.5.
b
I (total) = 2I (PAR).
c
From Table 8.1. Updated, parenthetical values from Larcher (1995) make only a 4% difference in and
are not used.
d
Exponential fit to data from source.
e
For seedlings.
f
For individual needles.
According to Larcher (1983, p. 119) natural selection adjusts Tm to the average climatic
temperature, T0 , of the habitat.
240 Optimal bioclimate
o
T 0= C
3
50 cm hr1)
o
10 C
2
Net photosynthesis
o
0 C
1
2
2
1 3 5 7 9 11
Light intensity, I (klx)
Fig. 8.2. Photosynthetic capacity of Arctic willow (water non-limiting, and ambient
CO2 ). (From Muller, D., Die Kohlensaureassimilation bei arktischen Pflanzen und die
Abhangigkeit der Assimilation von der Temperatur, Fig. 2, p. 29, Planta, 6, Copyright
c 1928 Springer-Verlag GmbH & Co. KG; as given by Kramer, P.J., and
T.T. Kozlowski, Physiology of Trees, Fig. 3.11, p. 83, Copyright c 1960 McGraw-Hill
Book Co., Inc.; with kind permission from Springer-Verlag GmbH & Co. KG and from
T.T. Kozlowski.)
2.5
Photosynthesis and respiration
2.0
is
hes
1.5
ynt
tos
pho
ss
sis
1.0
Gro
nthe
tosy
pho
0.5
Net
n
io
at
p ir
Re s
Tm
0
0 10 20 30 40
Temperature range of
net photosynthesis (C)
Optimum photosynthetic
Plant type or species temperature, Tm ( C) Reference
100 5
2
CO2 uptake (%)
1 4
50
6
3
0
20 10 0 10 20 30 40 50 60 70
(a) Tl ( C)
g (T
Tl ) Typical
0.5
broadleaved
plant
Arctic pine
Beech
0
30 0 30
(b) Tl (C)
Ps I
Pt = P + Pr = , (8.2)
I +k
in which for an isolated leaf
1
(water non-limiting, optimal
Asymptotes temperature, and ambient
CO2 ).
o
P o P
P = = o o
Pssm P I k
0.5
o o
k I SL = Relative saturation insolation
0
0 1
o
Relative insolation, I
the ratio Pr /Ps is only approximately 10% of Pt /Ps for trees (Larcher, 1983, Tables 3.4,
p. 95, Table 3.5, p. 97), we neglect it and rewrite Eq. 8.2 to estimate the net rate of
photosynthesis, P/Ps . This gives
P Pt I
= , (8.3)
Psm Psm I +k
dP k 1
= = , (8.4)
dI (I + k)2 k
I =0 I =0
from which we see that the saturation insolation ISL k. Note that with this definition,
we have effectively replaced the photosynthetic characteristic curve of Eq. 8.3 by its
asymptotes, and therefore the saturation radiances derived in this work will be smaller
than those reported in the literature and listed here in Tables 8.1 and 8.2.
Alternatively, since photosynthesis is biophysically controlled by the uptake of
CO2 through diffusive processes in the external leaf boundary layer and in the internal
leaf protoplasm, it seems equally reasonable to fit observations with an exponential
244 Optimal bioclimate
and intersection of the asymptotes again gives ISL k as is sketched in Fig. 8.6a. This
reduces Eq. 8.5 to
P
= 1 exp (I /ISL ). (8.6)
Psm
Equation 8.6 is compared in Fig. 8.6b with the same observations used by Horn
(1971, p. 69) where in the present case, the oak and pine photosynthetic rates have
been normalized using the observed asymptotic values. Here we see that Eq. 8.6 fits
all three species reasonably well using k o = ISL
o
= 0.11, a value which corresponds
1 rate of photosynthetic
reaction for leaf (water
Asymptotes
non-limiting, optimal
o o temperature, and ambient
P 0.63 o P I / k
P = =1e CO2 ).
Psm
Psm (a) Analytical form;
(b) comparison with
observations: , Eastern red
o o oak (Kramer and Decker,
k I SL = Relative saturation insolation
1944); , European beech
0 (Boysen-Jensen, 1932, after
0 1 Kramer and Kozlowski,
o
(a) Relative insolation, I 1960; , Loblolly pine
(Kramer and Clark, 1947).
1.2
1.0
o
0.8 P I o
= o o , with I = 0.05
Psm SL
I + I SL
P
0.6 o o
Psm P I / I SL o
=1e , with I SL= 0.11
Psm
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1.0
o
(b) I
Light control of productivity of an isolated leaf 245
P
ry
ist
ng
m
he
Photosynthetic capacity
ni
oc
pe
ot
ao
Ph
at
om
St
1 ISSL I
SW
0
0
I
Fig. 8.7. Definition of the leaf light characteristic (unit basal C3 leaf area, water
non-limiting, optimal temperature, ambient CO2 ).
1.2
)
22 11
Jarvis, Photosynthesis in
(s mm )
1
h
CO
rss 1971 Blackwell Science
0 I SL Ilc 0
Ltd; with kind permission
Respiration
from Blackwell Science
31.25
0 50 100 150 Ltd, Oxford, UK.)
Il 2)
light intensity (all else again being constant), and the reverse occurs as light intensity
falls. With this understanding, and with water availability non-limiting, we see that
the stomatal opening increases with increasing light intensity along the rising, photo-
chemical asymptote of Fig. 8.7. At photosynthetic saturation (the intersection of the
asymptotes in this idealization), the stomates are fully open, the stomatal resistance has
its minimum value, and the assimilation rate becomes constant at its maximum value
(consistent with the ambient concentration of CO2 ) along the asymptote controlled by
enzymatic processes (and hence temperature).
A piece of evidence supporting the use of the intersection of the asymptotes of the
photosynthetic capacity curve to define the critical point of fully open stomates is given
here in Fig. 8.8 (adapted from Ludlow and Jarvis, 1971, Fig. 3, p. 934). Observations
of net CO2 influx, P, and of stomatal resistance, rls , are presented there as a function
of irradiance, Il , for an isolated shoot of Sitka spruce (Picea sitchensis) undergoing
the usual phytotron test. The saturation value, P = Psm , is estimated at Psm = 2.10 g
m2 h1 as shown on the figure. Evaluating Eq. 8.4 at Il = ISL gives
P
(Il = ISL ) = 1 1/e = 0.632,
Psm
yielding P(Il = ISL ) = 1.33 g m2 h1 . Entering Fig. 8.8 with this value locates ISL
as is shown by the arrows entering from the P scale.
The rls data points of Fig. 8.8 were fitted by Ludlow and Jarvis (1971) with the
solid line shown, but over the range of observations could as well have been fitted
with the dashed straight line segments shown. We use the intersection of these dashed
segments to approximate the insolation, Ilc , at which the stomates first become fully
open.
Light control of productivity of an isolated leaf 247
The proximity of Ilc to ISL supports our use of the asymptote intersection of the
exponential photosynthetic capacity curve as the point of effectively full stomatal
opening. In so doing it provides the major physical justification for our use of this
intersection to define the normalizing critical insolation when characterizing the
dimensionless photosynthetic capacity curve. We recognize that values of ISL deter-
mined in this manner will be at variance with most of the existing observations, so we
have developed our own set of ISL values using observations, taken from the litera-
ture, of the photosynthetic characteristic of single leaves of eight separate species, and
fitting each of these with an exponential function as described above. Psm is normally
measured in milligrams of CO2 assimilated per unit of stomated leaf surface per unit
of time at the optimum leaf temperature, Tl = Tm , and ISL is measured in a variety of
units of either total or shortwave incident radiant energy per unit of illuminated leaf
area per unit of time. We have converted all energy fluxes to PAR (assumed to be one
half the total radiation) in units of megajoules (i.e., MJ).
The species studied are the bottom eight entries in Table 8.2 (i.e., Sitka spruce
through Arctic willow), and their experimentally determined photosynthetic capacity
curves are from the sources in the last column of Table 8.2. The value of Psm was
determined in each case by estimating the saturation asymptote of the source curve
and is listed in the first column of Table 8.2. For each species, several points were
read from the source curve over the full range of P/Psm and were used with Eq. 8.6
to obtain an average ISL for that species. These values are listed in the second column
of Table 8.2, and the resulting values of = Psm /ISL are given in the third column.
The individual points transferred from the source curves and put in common units
are displayed in Fig. 8.9. We note there that of those species studied, all but the creosote
bush (Larrea divaricata) appear to cluster about a common exponential (i.e., the upper
curve of Fig. 8.9) while creosote bush, the lower curve, reaches maximum productivity
at a much higher insolation.
Also shown in Table 8.2 are average values of these parameters, taken from
Table 8.1, for C3 crops, and for the sun leaves of deciduous trees and evergreen
coniferous trees, the three main vegetation types under consideration here.
This is another of our expedient assumptions. Larcher (1983, p. 93) points out that stomatal
opening responds to the interplay of water potential, humidity, and temperature as well as to light,
and that the stomates are completely open only rarely and briefly.
248 Optimal bioclimate
1.2
1.0
Creosote bush
0.8
P
(1 e1) Arctic willow (Kramer and Kozlowski, 1960)
Pssm 0.6
Eastern red oak (Kramer and Decker, 1944)
0.4
European beech (Boysen-Jensen, 1932)
ISSL
ISSL Sitka spruce (Jarvis et al., 1976)
0.2
Creosote bush (Ehleringer, 1985)
0
0 1 2 3 4
Fig. 8.9. Photosynthetic capacity curves for isolated leaves (water non-limiting,
optimal temperature, and ambient CO2 ).
in which the intercepted radiation used is sometimes short wave and sometimes total,
while the mass of roots is usually excluded when is reported for whole trees. We
recognize as the slope, Psm /ISL , of the photochemical asymptote of the leaf light
characteristic presented in Fig. 8.7. However, in evaluating and in using values of
from the literature, we must be careful of the units.
The estimates of Psm given in the first column of Table 8.2 have been converted to
the resulting mass of solid matter using the accepted average conversion factor
g(solid matter) gs
= 0.5, (8.7)
g(CO2 assimilated) g
as given by Penning de Vries et al. (1974), and Ledig et al. (1976, fide Russell et al.,
1989b, p. 27).
The needles of some conifers have stomates on more than one side (see Chapter 2)
in which case the the stomated and illuminated areas of the leaf are unequal. To
accommodate such cases we introduce the additional modifier,
stomated (i.e., assimilating) leaf area
o = . (8.8)
illuminated (i.e., projected) leaf area
For example, loblolly pine (Pinus taeda) has stomates on all three needle sur-
faces (Kramer and Decker, 1944, p. 352), thus in direct light this species has
o arc -chord = 2.36 (Raison et al., 1992). The observed value of Psm for this species
perimeter
was reported per unit of stomated area and thus was multiplied by o = 2.36 for listing
in Table 8.2. For broad leaves o = 1.
Light control of productivity of an isolated leaf 249
Table 8.4. Estimates of the dry matter : radiation quotient, , for whole plants (water
or nutrients not limiting; main period of growth; radiation is total intercepted)
Plant name Species Reference
(gs MJ1 )a
Agricultural C3
Rape Brassica napus 1.1 Mendham et al., 1981
Pigeon pea Cajanus cajan 1.2 Hughes et al., 1981
Pea Pisum sativum 1.5 Heath and Hebblethwaite, 1985
Potato Solanum tuberosum 1.4 Scott and Allen, 1978
Typical 1.4 Monteith, 1977
Woody C3
Monterey pine Pinus radiata 0.9b Linder, 1985
1.45 Linder, 1985, plus Forrest and Ovington
(1970), and Forrest (1973), fide
Cannell (1982, p. 15), give: (root dry
weight)/(total dry weight) = 0.38.
Scots pine Pinus sylvestris 0.85b Linder, 1985
1.03 Linder, 1985, plus Malkonen (1974);
Albrektson (1980a, 1980b); Ovington
(1957, 1959, 1961); fide Cannell
(1982, pp. 63; 225; and 244), give:
(root dry weight)/(total dry weight) =
0.24 (3 sites); 0.11 (15 sites); 0.31
(3 sites).
Black pine Pinus nigra 0.85b Linder, 1985
1.06 Linder, 1985, plus Miller et al. (1980),
fide Cannell (1982, p. 247), give: (root
dry weight)/(total dry weight) = 0.20
(7 sites).
Sitka spruce Picea sitchensis 0.85b Linder, 1985
1.05 Linder, 1985, plus Deans (1979, 1981);
fide Cannell (1982, p. 242), give: (root
dry weight)/(total dry weight) = 0.19
(1 site).
Eucalyptus Eucalyptus globulus 0.45b Linder, 1985
0.56 Linder, 1985, plus Feller (1980); Rogers
and Westman (1977, 1981), Westman
and Rogers (1977a, 1977b); fide
Cannell (1982, pp.15; 18), give: (root
dry weight)/(total dry weight) = 0.10
(1 site E. regnans, and 1 site
E. obliqua); 0.38 (1 site, E. signata
and E. umbra).
Willow Salix viminalis 1.4b Cannell et al., 1987
a
Radiation is total intercepted.
b
Above-ground biomass only (i.e., roots excluded).
Source: Russell, G., B. Marshall, and P.G. Jarvis, editors, Plant Canopies: Their Growth, Form
and Function, Cambridge University Press, 1989, Table 2.1, p. 28 and following text; reprinted
with permission of both Cambridge University Press and G. Russell.
Creosote bush
2.0
1.6
C3 crops
2 h 1)
city
pa
ca
tion
1.2
m ila g
ssi <>
s
s
la = 0.81
ica MJ(tot)
em
Psm
h 1
oc
0.8 Bi
Sitka Deciduous trees
spruce
Coniferous trees
Goethalsia
0.4 Beech Loblolly pine
Arctic Red oak
willow
White oak
0
0 0.4 0.8 1.2 1.6 2.0 2.4
I SL (MJ(tot) m2 h1)
Fig. 8.10. Biochemical assimilation capacity of C3 leaves. Data from Table 8.2. Note
that these data have been updated by Larcher (1995, Table 2.4, p. 85 and Table 2.8,
p. 98) since the original preparation of this figure. Use of the updated data here would
result in greater scatter leading to <>= 0.78.
(a) I
Optimum
Species no. 2
Pssm2
1 I0 = I SL2 I SW
(b) I
252 Optimal bioclimate
biomass production. Referring back to Fig. 8.10 now we note the direct
proportionality between Psm and ISL (i.e., constant ). Therefore (return to
Fig. 8.11a), species having larger saturation insolation at this T0 will be more
productive (i.e., larger Psm ). We hypothesize that evolutionary pressure will
move the biological system in this direction. In a single-species world we
might expect this species to evolve or adapt biochemically to a new form
having ISL2 = I0 as is indicated by operating point no. 2 in Fig. 8.11a.
However, in our multispecies world, we expect that a species no. 2, having the
same Tm but larger Psm , will take over (as indicated by the upper dashed curve)
such that ISL2 = I0 , and it will operate at point no. 2. Species no. 2 is thus
optimal for this climate; after Lindeman (1942) we will refer to it as the
climax species, that is the species that is maximally productive there, and hence
stable.
(4) Continuing our consideration of Fig. 8.11a, if T0 = Tm for species no. 3 but
I0 < ISL3 , the operating point is again no. 2. Here the photosynthetic rate, Psm2 ,
is less than Psm3 , and at operating point no. 2 species no. 3 is not producing at
capacity. Under such suboptimal operation, species no. 3 lives in an average
state of stress due to insufficient light with associated increased risk of trauma
during transient periods of lower insolation. Furthermore, plant support of such
unproductive capacity is suboptimal because the constant and non-productive
light requirements of respiration consume a greater proportion of I0 than they
do when I0 has the saturation value, ISL . Therefore, natural selection would
seem strongly to favor a new species (species no. 2), having smaller ISL at this
I0 , such that ISL2 = I0 as is shown by the lower dashed line in Fig. 8.11a. At
operating point no. 2, this species will be just as productive as species no. 3
would be at this I0 , but optimally so. Species no. 2 would again qualify as a
climax species for this environment.
(5) We conclude that with ample water and a continuous range of species-specific
Tm , the optimum relations between climate and species are T0 Tl = Tm , and
I0 = I S L , for here P = Psm which is the maximum possible photosynthetic rate
(as is shown in Fig. 8.11b). We refer to the suboptimal state I S L > I0 as
light-limited and the suboptimal state I S L < I0 as species-limited.
Note that we make this climax classification solely on the basis of full productive use of the
environmental light resource provided, of course, that water is not limiting. To the first
approximation (i.e., neglecting the stimulation of added assimilation capacity by the water-borne
flux of added nutrients) productivity is independent of excess water supply beyond that needed to
keep the stomates fully open.
These discussions of pressures for species substitution are simplistic when viewed from the
viewpoint of species succession in natural ecosystems (cf. Colinvaux, 1973, pp. 550572) where
various other strategies are believed to be operative at different successional stages.
Water control of productivity of an isolated leaf 253
of transpiration. We thus expect the photosynthetic rate of at least the uppermost leaf
initially to have the saturation value, P = Psm . For subsequent times, it is helpful to
understand the energetics of water movement in the plant from its reservoir in the root-
zone soil up to the surface of the spongy mesophyll cells lining a stomatal cavity in this
uppermost leaf where it undergoes an evaporative state change from liquid to vapor.
Water moves through the plant in the xylem under the influence of a potential gra-
dient between xylem terminals in the root and the leaf. However, many crucial details
of this extremely complex process remain controversial. For purposes of qualitative
reasoning only, we will follow here the simplified engineering explanation offered by
Hendricks and Hansen (1962).
Figure 8.12a gives a schematic representation of the energy grade line, and
Fig. 8.12b shows typical values of the associated energy potentials. The pressure
intensities, pc , of fluid held in the soil pores by capillary forces are negative (i.e., the
fluid is in tension) as are the related soil moisture (capillary) potentials, s = pc /
(f g) pc /f , where f and f are the mass density and the specific weight of the
fluid respectively, and g is the gravitational constant. The root-zone capillary potential
is related to the root-zone soil moisture concentration by the well-known relation
from soil physics (cf. Eq. 6.30)
in which s (1), the potential at saturation (i.e., s = 1), and m are hydraulic properties
of the soil.
As time progresses in our imaginary transient experiment, transpiration and down-
ward percolation in the soil reduce the average root-zone soil moisture concentration,
s, and the associated root-zone soil moisture potential, s , becomes more negative.
The withdrawal of moisture from the soil adjacent to the roots lowers the capillary
potential there by the additional amount, h s , needed to support flow to the root surfaces
against the resistance of the soil. A positive osmotic potential, Ho , exists across the
root cell membranes due to concentration differences between the soil solution and the
sap solution (Kramer, 1969, pp. 150173). This potential acts as a pump which may
(seasonally at least) develop a rise in pressure head on the order of 1 to 2 atmospheres
(Kramer, 1969, p. 158). On the plant side of the root walls there is a drop in potential
due to friction losses in the roots (h r ), xylem (h x ), and leaves (h l ), while the height of
the tree gives a positive gravitational potential, HG . Across the leafs spongy meso-
phyll cells, from the veins to the evaporation surfaces in the stomatal cavities, there is a
positive capillary potential, HL l , which can be sizeable depending upon the effec-
tive capillary diameter of the spaces between the mesophyll cells. In Fig. 8.12a these
potentials are added to their respective elevations, z, at every point along the flow path
to obtain the conventional hydraulic grade line (i.e., piezometric head line). When
the hydraulic grade line lies below the elevation of the point in question, the moisture
potential at that point is negative. As is shown in Fig. 8.12a, the fluid is under tension
in much of the tree. Conventional wisdom has rejected tension-induced cavitation by
Water control of productivity of an isolated leaf 255
o
zl
l
z h
(a)
Ho
Ho
(b)
Fig. 8.12. Energetics of transpiration. (a) Hydraulic grade line; (b) energy potentials.
(After Hendricks, D.W., and V.E. Hansen, Mechanics of evapo-transpiration, Fig. 3,
p. 73 and Fig. 4, p. 74, Proc. A.S.C.E., J. Irrigation and Drainage Div., 88, IR2,
Copyright c 1962 American Society of Civil Engineers; with kind permission from
the American Society of Civil Engineers.)
256 Optimal bioclimate
virtue of cohesion between adjacent fluid molecules (Briggs, 1950) and by adhe-
sion between the fluid molecules and the elastic vessel walls (Slatyer, 1967, p. 212).
However, more recent studies (cf. Milburn, 1993) demonstrate that xylem hydraulic
conductivity may be substantially reduced by cavitation and the resulting formation of
gas-filled passages. Holbrook and Zwieniecki (1999) hypothesize that these gas em-
boli are continuously repaired, perhaps diurnally (Zwieniecki and Holbrook, 1998),
and that measured hydraulic conductivity is a dynamic balance between the processes
of damage and repair.
Because the flow velocities are so small, the kinetic energy of the plant fluid is
negligible, and the potentials of Fig. 8.12a can be summed to give Hf , the potential
available to support friction loss, as
in which HG zl z g .
As time progresses (without additional precipitation remember!), the magnitude of
s increases sharply (Eq. 8.10), and the plant must work ever harder to pump moisture
from the soil to the stomates at the atmospheric demand rate, E ps . The values of Ho
and HG in Eq. 8.11 are fixed by the given plant, so to maintain constant Hf (and hence
E v = E ps and P = Ps ), the leaf potential, l , must increase by the same amount as
|s |.
At some (critical) value of the falling soil moisture, s = sc , the plants available
pumping capacity, Ho + l , can no longer overcome the opposing potentials, |s | +
HG + Hf , while keeping the flow at its maximum rate, E ps . At subsequent times,
without some change in the opposing potentials, the atmospheric demand would begin
to desiccate the plant structure in order to make up the deficit in the plants transpiration
capacity. To prevent this, the plant has evolved a mechanism to keep evaporative
demand more-or-less in balance with the falling transpiration capacity.
A biological sensor in the guard cells surrounding the stomatal opening senses
the moisture potential, l , in the leaf. When l reaches the critical value, lc (cor-
responding to the critical soil moisture, sc ), the guard cells respond by reducing the
stomatal opening and thereby putting the plant into a stressed (as opposed to relaxed)
physiological state. This restriction to the vapor flow raises the vapor density within
the stomatal cavity and thereby decreases the transpiration rate. The friction loss, Hf ,
then falls requiring (see Eq. 8.11) that l < s for a given negative s. This de-
cline in the transpiration rate for l > lc implies a proportional decline in the rate of
water-borne nutrient flux from the soil as well as in the rate of CO2 assimilation. Thus
there is a decrease in net biological productivity as the soil dries below the critical
value sc .
Figure 8.13 shows the typical variation of the rates of transpiration and of net
photosynthesis for several crop and tree species as a function of l under conditions
Zwieniecki et al. (2001) find that salt-sensitive gels in certain xylem membranes constantly shrink
and swell as needed to regulate the water flow.
Water control of productivity of an isolated leaf 257
lc
1.0 1.0
o
Net photosynthesis, P
3
4
(relative units)
1 2
o
kv
s = sc
0.5
0 0
0 10 20 30 40 50 60 70 80
l (bar)
of unlimited light as adapted from Larcher (1983, Fig. 3.46, p. 123). These ordinates
are presented as normalized by their maximum rates such that
kv P
kvo , and Po . (8.12)
kv Psm
Both normalized ordinates have the maximum value of unity while the stomates are
fully open (see Chapters 6 and 8). We see in Fig. 8.13 that lc has virtually a common
value of about 10 bars for the species considered.
Note in Fig. 8.13 that the stomatal control of transpiration is singularly weak in the
creosote bush. This is probably due to low soil moistures characteristic of that plants
habitat, for at low soil moisture, s /s is very large requiring a comparably large
l to accomplish a given reduction in Hf (i.e., in E v ).
Thus, in the manner just outlined for this case of transient drying, gradual stomatal
closure permits the plant to continue production, albeit at an ever-decreasing rate,
without desiccation but in a state of physiological stress until the drying soil forces
complete shut-down of the stomates. Equation 8.11 can be used to transform the
independent variable of Fig. 8.13 from l to s. With laminar flow in the soil pores and
plant veins, friction losses are proportional to the velocity in these passages and we
can write
Hf = kL E v , (8.13)
in which kL (sec) is the sum of the geometrically scaled potential-loss coefficients for
the soil, root, xylem, and leaf passages. Equations 8.11 and 8.13 then give
l |s | + Ho HG
Ev = . (8.14)
kL
258 Optimal bioclimate
o
response to falling soil
Transpiration, E v
1.0 moisture (loblolly pine;
Relative instantaneous
transpiration efficiency.
clo
ma
s
ta
s oc
Cuticular
transpiration
0
so
Stomata
(b) closed
Using Eqs. 8.10 and E.12, Eq. 8.14 provides the desired coordinate transformation to
C1
Po = l s (1)s 1/m + Ho HG . (8.15)
kL
With Eq. 8.15 each pair of values (P o , l ) for a particular species in Fig. 8.13
maps to a pair (P o , s) having the transient form illustrated for a loblolly pine canopy
in Fig. 8.14a. Note here that with falling root zone soil moisture, the photosynthetic
rate remains constant at its maximum value until the critical transient soil moisture
content, s = sc , is reached and the stomates begin to close. This maximum value of
photosynthesis corresponds to the maximum value of the canopy conductance, kv = kv
or kvo = 1. This transient case is of course the natural one where occasional additional
precipitation produces soil moisture recharge which intermittently resets the decaying
transpiration rate at a higher value 0 < E v E pv . Note particularly in Fig. 8.14a the
proportionality of P and E v .
Here the proportionality results from control of both P and E v by stomatal closure induced by the
drying soil. In Appendix E we demonstrate approximate proportionality for the condition of open
stomates.
Water control of productivity of an isolated leaf 259
Steady-state case
Time-averaging the succession of soil moisture transients characterizing a given cli-
mate, we define an imaginary steady-state which we assume to be representative of
climatic behavior to a degree that becomes exact only at the limit in which variance
of the weather vanishes. This steady-state approximation has yielded useful insights
into the behavior of the climatic water balance (Eagleson, 1978f), and was introduced
in Chapter 6 as the basis for analysis in this book. The time-averaged (i.e., climatic)
root-zone soil moisture concentration is denoted by so . If we change the independent
variable of Fig. 8.14a from the transient s to the climatic so , as in Fig. 8.14b, then
the sketched function represents the relative productivity of that species as a function
of soil moisture climate provided that there is always enough light in this climate for
photosynthesis at the rate P equivalent to Ev . The insolation, I , at which so attains its
critical value, soc , and plant desiccation begins is an important bioclimatic parameter
which, in comparison with ISL , and the actual climatic insolation, I0 , determines plant
habitat. In making this comparison through the radiation-driven transpiration rate we
will be careful to deal with daylight-hour time periods and rates. For example, we use
Eq. 6.46 to write, for a unit vegetated area
E v = v kv E ps
, (8.16)
in which E ps 2E ps is the average daylight hour potential rate of evaporation from
a wet surface, and v is approximated from Eqs. 6.43 and 6.48 for the period follow-
ing evaporative exhaustion of surface retention by retaining only the second term of
Eq. C.6. That is
4 ne zr
v 1 exp s
o
. (8.17)
M kv E ps
Equation 8.17 is sketched as the solid line labeled stomatal transpiration in Fig. 8.14b.
That is, the time-averaged depth evaporated in one-half the time, making m t E ps m t E ps .
b b
Note that in discussing the behavior of a single leaf, kv and v can be less than unity only by
virtue of stomatal closure. This contrasts with our use of kv to describe the conductance of a
multilayer canopy in which all stomates are assumed fully open.
260 Optimal bioclimate
variable transformation is important in that it allows us to present the falling leaf pro-
ductivity of Fig. 8.14b as a result of increasing environmental insolation rather than
of decreasing soil moisture, and thus facilitates comparison of the leaf water charac-
teristic with the leaf light characteristic of Fig. 8.7. Using first Eq. C.17 and then the
first-order approximation to Eq. E.1 we get, for a given moisture supply, the functional
relationships
1
so = f 1 E ps = f 2 (I 1 ). (8.18)
Note that the function P(I ) given in Appendix E is different when carbon is supply-limited.
Water control of productivity of an isolated leaf 261
Pwm = Psm
Stomata
Stomata fully open
P closing
Stomata
closed
I SL I SW
0
I
Fig. 8.15. Definition of the leaf water characteristic (unit basal leaf area of given
species, unlimited saturation insolation, optimal temperature, ambient CO2 , given soil
and precipitation).
Assuming that increased transpiration maintains optimum leaf temperature.
262 Optimal bioclimate
Their relative values determine the feasibility, stability, and optimality of a given
climatevegetation system, as well as the direction of any evolutionary pressure. Al-
though there are six permutations of the relative values of these three insolations,
the feasible set is smaller when we consider a given species and observe a series of
constraints:
as is demonstrated in the sketch of Fig. 8.16. This range of I0 defines the feasible
radiational habitat of the given species for the given soil and water supply. A few
comments about this figure are in order:
The solid line represents the actual photosynthesis, PL , of the leaf given by its actual
light characteristic as influenced by the moisture supply of the given climate, while
the dashed line represents the potential water-limited photosynthesis of the leaf, PW ,
as influenced only by the climatic moisture supply and the stomatal opening as though
light saturation did not occur. The jointly controlled leaf photosynthetic characteristic
follows the form of the light characteristic in Fig. 8.7, until photosynthetic saturation
at I = ISL . The amount of light controls PL for I ISL , and the species, through its
photosynthetic saturation, Psm , controls PL for ISL < I ISW . In this latter range, the
stomates are fully open, there is excess water since PW > PL , and thus the canopy
cover is complete at M = 1. At I = ISW , the stomates begin to close under contol of
the available water and both the actual and the water-limited potential photosynthesis
of the individual leaf fall accordingly. At ISL = I = ISW decline of the climatic water
supply will be felt through decrease of the canopy cover, M, leaving Psm = Pwm .
Additional constraints follow from optimality considerations as we now see.
To the extent that I0 > ISL , the given species is climatically suboptimal and hence
is selectively unstable in the sense that it does not make maximum productive use
of the available radiation (cf. Fig. 8.11a) and there will be selective pressure to do so.
To the extent that I0 = ISL but I0 < ISW , the species is photosynthetically optimal and
stable, even though it does not make maximum productive use of the available water,
since there is no resulting selective pressure. Such a system is light-limited because
with larger I0 a species with correspondingly larger ISL and Psm may be substituted.
Accordingly, for optimal conditions we add additional constraints.
energy balance as written in the simple form of Eq. 5.5. Incorporating the Bowen
ratio estimator of Eq. 5.38 we have, for the daylight hours
Rn = (1 Rb ) E T , (8.22)
Rn (1 T ) I qb , (8.23)
where T = spectral reflectance (i.e., albedo) of canopy, and qb is the net longwave
back radiation as is given empirically by Eq. B.12.
Together Eqs. 8.22 and 8.23 approximate the daylight hour balance of energy fluxes
for a unit vegetated area in which we replace the nominal (i.e., 24 hour) evaporation
rate, E T , by the actual daylight-hour rate, E v , giving
1
I = (1 Rb ) E ps + qb , (8.24)
1 T
where E v is linearly related to the photosynthetic rate by a climate- and species-
dependent coefficient derived in Appendix E, and where the refers to sensible heat
rejection (+) or acceptance () by the canopy.
Substituting Eqs. 6.46, 6.66, and 8.16 into Eq. 8.24 gives
1 Ve v
I = (1 Rb ) + qb . (8.25)
1 T m tb M
At the critical condition I = ISW , so = soc , and v = 1. We then rewrite Eq. 8.25
for this single leaf layer,
1 Ve
ISW = (1 Rb ) + qb , (8.26)
1 T m tb M
in which, to the first approximation Eqs. 6.67 and 6.68 become independent of the soil
properties (cf. Eq. 6.65 and Table 6.4) and define, for deep water tables
ho S
Ve m h 1 + . (8.27)
mh m mh
We see from Eq. 8.26 that the larger the moisture supply rate, Ve /m tb , the larger will
be the critical radiance, ISW .
It is helpful to remember the relationship between M and ISW as the average availa-
ble soil moisture changes.
(1) Start with the situation I0 < ISW signifying that there is excess available
moisture which insures that M = 1, and that the stomates are fully open.
For the purpose of calculation in Chapter 9, we will use the PenmanMonteith formulation,
Eq. 5.28.
266 Optimal bioclimate
(2) Now decrease the rate, Ve /m tb , at which soil moisture is available (all other
environmental variables remaining constant), until the system arrives at the
critical average state, ISW = I0 , with the stomates still open, and M = 1.
(3) Continue to decrease Ve /m tb and, since ISW < I0 , there are only two
alternatives: (a) the stomates close to reduce the rate of moisture use while
maintaining M = 1 (but placing the plant in a state of stress), or (b) the
stomates remain fully open (and the plant remains unstressed) while the
canopy cover declines just enough to keep Ve /(m tb M) constant, ISW = I0 , and
thereby maintain the critical water supply for each leaf. Adhering to our basic
premise of a stressless average state, the latter alternative is the only feasible
operating state. However, to get there it is necessary for individual plants to
die off through transient pursuit of alternative (a). As the most sensitive plant
succumbs to stress, the additional soil moisture thereby made available reduces
the stress in those plants still living.
The insignificance of percolation to groundwater under these conditions of critical
soil moisture means that whenever the average operating state is I0 = ISW , ground-
water can be generated during the growing season only due to finite variance of the
climatic properties, or during the dormant season due to E v 0. This demonstrated
insensitivity of ISW to soc may be the physical basis for the previous observation by
Denmead (1973) that transpiration is almost independent of soil conditions.
In summary, our hypothesis of an optimal bioclimatic state or climax state for
the individual leaf is founded on the concept that any average operating state which
requires stomates to be less than fully open (i.e., ISW < ISL ) is stressful to the plant and
will be selected out naturally. It proposes that in a given climatesoil, natural selection
and/or adaptation leads to Tl = Tm = T0 at I0 = I S L I SW , at which point P = Psm .
We will refine these conditions after acounting for the principal interactions among
the leaves in a real canopy.
Such benefit of the group through seemingly altruistic self-sacrifice by the individual plant is not
a negation of the selfish organism assumption of natural selection (cf. Dawkins, 1989) but is
instead a reflection of variance in the transient responses of the individuals in the group.
See Table 6.4 for estimates of the terms in the water balance leading to reducing Eq. 6.67 to the
form of Eq. 8.24.
Once again we caution that this excludes significant soil moisture supply from groundwater
which is considered as a special case in Eq. 6.72.
Bioclimatic optimality for the canopy-average leaf 267
(1) Since all leaves tested from various tree and C3 crop species appear to have a
common value of (cf. Fig. 8.10), we are led to the enabling, but gross,
approximation that all productive leaves of a given species have identical light
capacity curves at the same temperature. Under such circumstances, Psm is the
only variable light characteristic, changing with both species and operating
temperature. Of course, Table 8.1 seems to belie this assumption with its
demonstration of different photosynthetic characteristics for sun leaves and
shade leaves of the same species at optimal temperature. We acknowledge
this by further assuming that only the sun leaves are significant contributors to
the net productivity which yields seeds, and hence only these leaves are the
determinants of plant survival strategy.
268 Optimal bioclimate
Canopy-averaged insolation
Using the hat, . . . , to distinguish crown-averaged values from those of a single leaf,
the above assumptions lead to the effective canopy saturation insolation, I SL , for a
crown of a given species remaining unchanged from the individual leaf at
Psm
I
SL ISL = , (8.28)
While the comparable effective canopy environmental insolation, Il , is obtained by
averaging the actual insolation, Il , over the crown volume.
For cylindrical multilayers, the insolation, Il , incident upon a leaf decays with depth
into the canopy according to Eq. 3.53 which we average over the canopy depth to get
1
Il 1 e L t
f I ( L t ) = exp ( L t ) d = . (8.29)
I0 0 Lt
For tapered multilayers (e.g., cones or hemispherical segments), the compensation
light intensity is assumed to exist for z = h s at all radii, leading again to Eq. 8.29.
Invoking our important earlier finding that = (cf. Eq. 3.52), we then write for all
crown shapes
Il 1 e L
f I ( L t ) = . (8.30)
I0 Lt
Equation 8.30 is plotted over the important range of L t in Fig. 8.17. Over this range
Eq. 8.30 yields the approximate mean, f I ( L t ) = 0.36.
For monolayers of course, the incident insolation acts undiminished on all leaves
giving
Il
f I ( L t ) = 1. (8.31)
I0
Similarly, with ISW determined by the uppermost leaf in the canopy using Eq. 8.26,
we define the effective leaf desiccation insolation,
ISW , for the entire canopy to be
ISW f I ( L t )ISW . (8.32)
Although generally overlooked, these canopy-average reductions of the insolation incident at
canopy-top are the effective insolations for use with big leaf canopy models.
Bioclimatic optimality for the canopy-average leaf 269
0.4
Il
I0 0.3
fI (Lt )
0.2
0.1
0
0 2 3 4 5
Lt
With these averagings, we have created a fictitious canopy in which every leaf sat-
isfies Eq. 8.28 and is exposed to the canopy environmental insolation,
Il f I ( L t )I0 ,
in which I0 is the daylight-hour average environmental insolation during the growing
season.
Heat
T0 Tl = Tm ; (8.33)
Light
I
SL Il , (8.34)
in which the equality, together with Eqs. 8.28 and 8.30, defines the canopy climatic as-
similation potential, Psm = f I ( L t )I0 , while the degree of inequality signifies species
instability and measures the productivity pressure for species substitution; and finally,
to ensure no water stress,
Water
Il
ISW , (8.35)
270 Optimal bioclimate
Plate 8.5. Cedar among oaks. Left to right: scarlet oak, white oak, and Eastern red
cedar. (Photographed in Holden Arboretum, Cleveland, by William D. Rich; Copyright
c 2001 William D. Rich.)
Species stablility
I
SL = Il ISW , (8.38)
The feasible and productively optimal bioclimatic space for single-species canopies
is illustrated by the prism sketched in Fig. 8.18a using the boundaries provided by
Eq. 8.37. The stable subset of this region is shown by the cross-hatched sloping surface
according to Eq. 8.38, and the productively optimal climate for a given species is
defined from Eq. 8.39 by the heavy edge of this latter surface. Note in this figure that the
vertical coordinate is a species parameter since ISL Psm /, while the two horizontal
coordinates define the environmental controls of light and water (including the soil).
We see clearly from this that the range of feasible species increases with increasing
(a) I (water)
SW
Infeasible Climatic
productive
optimum
1
Stable, light-limited optimum
Feasible
I
SL
L = Infeasible
Il
Infeasible
0
0 1
I
SL
=
W
(b) I
SW
272 Optimal bioclimate
availability of light and water while the climate range becomes more restricted. The
inverse is true in resource-poor climates.
It is perhaps easier to examine this space by reducing its dimensionality through
the use of dimensionless variables. Dividing Eq. 8.37 by ISL and inverting each term
we have
1 L W , (8.40)
Note that in this representation of bioclimatic space, the branch for W 1 is shared by both the
stable light-limited climax and the limiting unstable sub-climax.
Bioclimatic optimality for the canopy-average leaf 273
Infeasible
I SL > Il = ISW
I
SL
= 1 Water-limited climax
W I
SW ISL = Il = ISW
P
ISL = Il = ISW
I SL
P I SW
I
Il
I
Light-limited climax
I SW
ISL = Il < I SW
P Pw
P
Il L
ISL
I
Sub-climax thus unstable
ISL < Il = ISW
0
0 1
M
given by Eqs. 6.67 and 6.68, and even disappear entirely with our
approximation of Ve in Eq. 8.27. Soil hydraulic properties are indeed important
in controlling stomatal closure. However, at the critical soil moisture state, soc ,
which determines ISW , the stomates remain fully open (but at incipient
closure). The soil is therefore only weakly influential (cf. Table 6.4) as long as
the water table is deep. Nevertheless, the literature emphasizes the importance
of plantsoil water relations (e.g., Slatyer, 1957a, 1957b; Gardner, 1960; Eyre,
Note the role that advected heat (i.e., negative Rb ) from the bare soil fraction can have in increased
transpiration and thus in reducing ISW (cf. Eq. 8.26), and the opportunity for regulation of ISW by
through its effect upon T . Note also that such increased water use will bring increased nitrogen
flux with resulting increase in ISL and in L t , the latter of which produces decreased
Il .
Restriction of the availability of carbon and of soluble nutrients in the soil will be considered
briefly in Chapter 10.
274 Optimal bioclimate
Refer to Salvucci and Entekhabi (1995, corrected 1997) in order to find Ve under water table
influence.
9
The separate heat, light, and water propositions of the bioclimatic optimality
hypothesis are tested against the characteristics of observed canopies and their
environments. On the basis of the limited evidence presented all three proposi-
tions are strongly supported. Seven of the nine vegetation communities tested
appear to be in their hypothesized natural habitat, and only two are more than
6 percent removed from the hypothesized climax state.
275
276 Natural habitats and climax communities
Plate 9.1. Creosote bush desert community. (Larrea tridentata photographed in the
Mohave Desert, Nevada, by Robert H. Webb; courtesy of Robert H. Webb and the U.S.
Department of the Interior.)
Test
In testing Eq. 9.1 we are limited to the species for which experimental determinations
of Tm are available and for which the geographic location (and hence T0 ) of their stable
habitat is also known. Values of Tm are given for several species in Fig. 8.4a taken
Proposition no. 1: Heat 277
from Larcher (1983, Fig. 3.35, p. 114), and the natural North American habitats of
many species are given by Burns and Honkala (1990, vols. 1 and 2). With these data,
the test is possible for only three species.
Loblolly pine
The only exact match of species between these two data sources is for loblolly pine.
Figure 8.4a gives for loblolly pine, 23 C Tm 30 C, while Baker and Langdon
(1990, Fig. 1, p. 497) present the natural habitat for loblolly pine in North America
indicated by the shaded area of Fig. 9.1. Average July atmospheric temperatures, T0 ,
for the period 194170 are given in the table accompanying Fig. 9.1 as taken from U.S.
Fig. 9.1. Natural North American habitat of loblolly pine. (From Baker and Langdon,
1990, Fig. 1, p. 497; courtesy of the U.S. Department of Agriculture.)
Station Location T 0 a ( C) Station Location T 0 a ( C)
a
T0 = (Normal July Daily Maximum + Normal July Daily Minimum) / 2; 194170.
278 Natural habitats and climax communities
Weather Service climatological records at nine stations within this habitat. The stations
are located in the cities shown on Fig. 9.1 and are chosen to cover the geographical
range of the habitat.
American beech
Here we must compare two different varieties of beech. Tubbs and Houston (1990,
Fig. 1, p. 326) give the natural habitat for American beech as is shown in Fig. 9.2,
and U.S. Weather Service average July atmospheric temperatures, T0 , throughout the
habitat are again listed in the caption. Figure 8.4a records Tm = 23 C for European
beech.
Fig. 9.2. Natural North American habitat of American beech. (From Tubbs and
Houston, 1990, Fig. 1, p. 326; courtesy of the U.S. Department of Agriculture.)
Climatological station Location T 0 a ( C)
a
T0 = (Normal July Daily Maximum + Normal July Daily Minimum) / 2;
194170.
Proposition no. 2: Light 279
Pacific
Ocean
3
Fig. 9.3. Natural North American habitat of Sitka spruce. (From Harris, 1990, Fig. 1,
p. 260; courtesy of the U.S. Department of Agriculture.)
Station Location T 0 a ( C)
Sitka spruce
In this last case, we are forced to make this comparison between two species of cold-
weather conifer, Arctic pine for Tm (cf. Fig. 8.4a where Tm = 15 C) and Sitka spruce
for T0 (cf. Harris, 1990, Fig. 1, p. 260). The North American habitat and average July
temperatures within the Sitka spruce habitat are given in Fig. 9.3.
The result of this limited testing of our heat proposition is summarized in Fig. 9.4
where the lines indicate the range of observation and the plotted circles the averages.
The data strongly support our proposition that T0 T l = Tm . We cannot say whether
this agreement is the result of natural selection of the species for the particular envi-
ronment or of adaptation of the species to the environmental conditions. In either case,
the result is the same.
Cold
1 Arctic willow 0.30 0.15 Finland, 60 0.56 0.45 f 6.75g 3.04 0.31 1.15 Kramer and Kozlowski
1/2 (1960, Fig. 3.11, p. 83).
2 Sitka spruce 0.55 0.26 Scotland, 55 0.60 0.51 f 5.14h 2.62 f,i 0.35 0.81 Jarvis et al. (1976, Fig. 21b,
1/2 p. 229).
3 European beech 0.33 0.16 Norway, 60 0.56 3.19 j 0.30 1.05 Kramer and Kozlowski
1/2 (1960, Fig. 3.9, p. 78).
Temperate
4 Eastern red oak 0.28 0.23 1/2 4332 0.73 2.60 j 0.35 1.11 Kramer and Decker (1944,
Fig. 1, p. 352).
5 White oak 0.26 0.23 1/2 4530 0.72 2.60 j 0.35 1.10 Kramer and Decker (1944,
Fig. 1, p. 352).
6 Loblolly pine 0.40k 0.30 1/2 3527 0.75 2.58 j 0.35 0.88 Kramer and Clark (1947,
Fig. 3, p. 55).
Hot
7 Creosote bush 2.08 1.17 Mojave desert, 3520 1.24 0.37l 1.00m 0.37 0.84 0.89 Ehleringer (1985, Fig. 7.8,
clear p. 172).
8 Goethalsia 0.38 0.22 Costa 10 0.62n 0.79o 3.30o 2.61 0.36 1.02 Allen and Lemon (1976,
Rica Fig. 6, p. 290).
Average 1.00
a b
From data, using conversion factor, gs g1 = 0.5 (see Eq. 8.7). From Fig. 7.15 at P/Ps = 1 1/e, shortwave only.
c
1/2 signifies the average of very clear sky and overcast sky radiances.
d
Single entry from map, using site of phytotron test; Temperate climate double entry from North American natural range (Burns and Honkala, 1990), Hot climate range
from map in MacMahon (1988, Fig. 8.18, p. 252).
e f g h i
Daylight Hour Average from Fig. 3.16. From Table 2.3 assuming = . From Table 2.2. L t = L t /. aw = L t ; average for all spruce, Table 2.2.
j k 2 1 2 2
From Baker (1950, Table 12, p. 143); see Table 3.8. 0.17 gs m (stomated area) h 2.36 m (stomated area) / m (illuminated area) (Raison et al., 1992).
l m
For Eucalyptus assuming = (see Table 2.2). Whittaker and Likens (1975) as given by Larcher (1983, Table 3.13, p. 151).
n o
Allen and Lemon (1976, p. 298). Allen and Lemon (1976, p. 288).
Proposition no. 2: Light 281
40
36
1
Loblolly pine
32
American beech
28
Loblolly pine
T0 ,C
24
European beech
Arctic pine
20
16
Sitka spruce
12
12 16 20 24 28 32 36 40
Tm, C
insolation as averaged both in time over the growing season and in space over the
depth of the canopy. That is
Psm
f I ( L t )I0 = ISL , (9.2)
or, in dimensionless form,
I
SL ISL
L = = 1. (9.3)
Il f I ( L t )I0
Test
Data for the test of this proposition are collected from the literature for eight species
native to a wide range of climates and are summarized here in Table 9.1.
In Fig. 8.10 we see that for C3 plants, an average value of is = 0.81 gs MJ(tot)1 .
282 Natural habitats and climax communities
(1) The values of Psm are obtained by scaling from the source photosynthetic
capacity curves for individual leaves as determined from phytotron
experiments. The values of ISL are then obtained by fitting Eq. 8.6 to pairs of
values scaled from these curves.
(2) Estimation of the effective growing season average environmental insolation
presents a more difficult problem. We use the total light curves of Fig. 3.10
to estimate the daylight-hour average PAR. For all climates save the Mojave
desert and the rainforest, we assume the annual average to be the mean of the
values for very clear sky and overcast sky and signify this with the
notation 1/2. For the desert we use the very clear sky value unmodified by
cloudiness, while for the rainforest we use the reported site observation. The
latitude at which Fig. 3.10 is entered to obtain I0 is an estimate of that at which
the phytotron test was performed in cases where the natural range of the
species was unknown (i.e., the cold climates). In the other cases the ranges
are known from the referenced maps, and their limiting latitudes are used. The
corresponding irradiances are then averaged to get I0 .
(3) Conversion of irradiance from leaf to canopy (cf. Eqs. 8.298.31) requires an
estimate of the light decay coefficient, L t . For species nos. 3 through 6 this is
found using Bakers (1950, Table 12, p. 143) estimates of light intensity at the
compensation point (cf. Table 3.8) along with Eq. 3.48. In the other cases, we
make use of our finding that = (see Chapter 3) and obtain separate estimates
of and of L t from the literature as is described in the notes beneath Table 9.1.
The two outside members of Eq. 9.2 are compared in Fig. 9.5 to test the solar radia-
tion proposition. The ratio, L = f I (ISLL t )I0 , has a mean value of 1.00 as is shown in the
last column of Table 9.1, and a standard deviation of 12%. Considering all the ap-
proximation involved in arriving at the individual values, this remarkable agreement of
the average with that hypothesized in Eq. 9.3 must be regarded as somewhat fortuitous.
This very significant result is highlighted in Fig. 9.6 by comparison of the (small
sample) distribution of f I ( L t )I0 /ISL with the functional depiction of this ratio at
leaf scale for the case of light control. In this representation of the distribution, the
circle is the mean, the line shows the range, and the box includes all values within
one standard deviation of the mean. The scatter of this small sample about its mean,
L = 1, appears biased very slightly toward L > 1 which is opposite to the di-
rection we would expect if the communities were photosynthetically underdeveloped
Il > I
(i.e., SL ). Discarding this tendency and disallowing the stressed circumstance,
Il < I
SL , we accept the hypothesis L = 1.
These results strongly support this proposition.
Note that all but one of the species tested here (i.e., Sitka spruce) appear to have homogeneous
cylindrical crowns as far as the distribution of leaf area is concerned.
For loblolly pine we use Bakers lodgepole pine and for pinyonjuniper we use either Bakers
Northern white cedar or Bakers lodgepole pine as detailed in a footnote in Chapter 7,
Section E.
Proposition no. 2: Light 283
1.2
1.0 7
1
0.8
[MJ (PAR) m2 h1]
1
fI (Lt )0
0.6
L = 1
0.4
4, 5 6
0.2 1 8 2
3
0
0 0.2 0.4 0.6 0.8 1.0 1.2
ISL [MJ (PAR) m2 h1]
Fig. 9.5. Test of light proposition: f I ( L t )I0 = ISL . 1, Arctic willow; 2, Sitka spruce;
3, European beech; 4, Eastern red oak; 5, white oak; 6, loblolly pine; 7, creosote bush;
8, Goethalsia. (Data from Table 9.1.)
1
fI ( Lt )I0
L =
ISL
0 1 2 3
Crown
tic
P
W
ris
ate
te
Po =
rc
ac
Species
ar
ha
Light
ch
limited Water
ra
cte
ht
limited limited
g
ris
Li
tic
I0 = ISW T0 = Tm
I0 = ISL
1
0
0 1 2 3
0 I0
I =
ISL
Fig. 9.6. Functional view of light proposition and the supporting evidence (fixed
species, soil, and precipitation). (Data from Table 9.1.)
284 Natural habitats and climax communities
E T = E v = kv E ps = kv E ps . (9.5)
Replacing kv and E ps through Eqs. 5.29 and B.1 respectively, Eq. 9.5 becomes
R n + Dp
E v = , (9.6)
o rc
1+ 1+
ra
in which the Dp term incorporates the Bowen ratio. We expand Rn according to Eq. B.3,
evaluate Dp empirically from Eq. B.18, and note from Eqs. 6.66 and 9.5 that the
true (i.e., daylight-hour) transpiration is controlled by the available soil moisture
to be
Ve
E v = , (9.7)
m tb M
all leading to the desiccation insolation
1 Ve o rc (1 Sr )
ISW 1+ 1+ + qb 1 . (9.8)
1 T m tb M ra 0.85
In Eq. 9.8:
(1) Ve , the moisture supply to the root zone, is approximated using Eq. 6.65 with
v = 1 (for deep water tables and with term estimates from Table 6.4) as
h0 S
Ve m h 1 + ; (9.9)
mh m mh
(2) m tb , the daylight-hour time between storms, is (Eq. 6.63)
1 o L t h 0 /2 1
m tb mt
= m tb ; (9.10)
2 b E ps 2
There are obviously many common and much more complex cases in which the root-zone
moisture is supplied totally or in part from below and a very different approximation of Eq. 6.65
is called for. Accurate estimation of Ve is at once the most difficult and the most important task in
formulating the water limit to vegetation growth, and the devil is in the details.
Proposition no. 3: Water 285
Plate 9.2. Subalpine fir and Engelmann spruce forest. (Photographed at Bow Lake in
Banff National Park, Alberta, by William D. Rich; Copyright c 2001 William D.
Rich. Tentative species identification by Arthur L. Fredeen from the photograph.)
rc
(3) ra
, the crown average resistance ratio, is (Eq. 7.25)
rc rc rc
= (1 M) +M ; (9.11)
ra ra M 0 ra M = 1
and
(5) Sr , the atmospheric saturation ratio (i.e., fractional relative humidity) at screen
height is obtained from station meteorological records.
ISL
W = 1, M 1, (9.13)
f I ( L t )ISW
ISL
W < 1, M = 1. (9.14)
f I ( L t )ISW
286 Natural habitats and climax communities
5, ponderosa pine;
ISL
8, pinyonUtah juniper;
3 9, creosote bush (desert).
0.4
(Data from Table 9.2; solid
Light-limited points are Beaver Creek
0.2 climax watersheds.)
0
0 0.2 0.4 0.6 0.8 1.0
M
Test
We have already supported this proposition by the test in Fig. 6.15 using the extraordi-
narily complete data set for the three pinyonjuniper and ponderosa pine communities
at Beaver Creek, Arizona. Here we expand the test to a wider range of habitats using
the less direct formulation in terms of desiccation insolation as is called for in Eq. 9.13
and as prescribed in Eq. 9.8.
We test the proposition further in Fig. 9.7 (as was suggested in Fig. 8.19) by exam-
ining the observed variation of the ratio, W f I (ILSLt )ISW , with the canopy density, M,
for stands of nine different species covering the range from the desert shrub (Larrea
divaricata) to the rainforest (Goethalsia meiantha). The supporting data are presented
and their sources are identified in Table 9.2. In interpreting this figure remember that
ISW is the value of the growing season average insolation, I0 , at which the available
soil moisture has that critical value which is just sufficient to keep the stomates at
incipient closure at the given temperature. Therefore, given verification of the heat
and light propositions as just presented, and given the strong verification of the critical
water balance (cf. Eq. 6.15), the degree to which ISL = f I ( L t )ISW for M < 1 must
be viewed as a test of the energy balance.
As noted in Chapter 6, evaluation of Ve in ISW is problematic. This is particularly true
for the interception loss and the carryover storage for precipitation-fed soil moisture,
and is even more uncertain where soil moisture is recharged from the water table.
Among the data presented, the carryover storage is best known for the ponderosa
pine and pinyonjuniper watersheds of Beaver Creek (Flagstaff, Arizona) which were
studied in Chapter 6 and reported in Tables 6.3 and 6.4. However, the value of ISL
was not available for ponderosa pine or for either of the pinyonjunipers. On the other
Proposition no. 3: Water 287
hand, ISL is known for the other six species (see Table 8.2), but all components of Ve
are unknown at the sites where ISL was measured. Thus in all nine cases, completion
of the data set required either outright estimation of one or more parameters or the use
of observations of the given species at other sites. This must be kept in mind when
evaluating Fig. 9.7.
The parenthetical superscripts used in Table 9.2 and defined in the listing accompa-
nying that table, identify the source and estimators of the parameter values used in this
test. Most should be self-explanatory; however a few deserve expanded commentary
due to the uncertainty of their estimates:
(1) Canopy cover (M). This variable is seldom reported by the foresters and plant
physiologists responsible for the bulk of useful data. The studies by Whittaker
(1966) of natural beech and oak stands in the Great Smoky Mountains of
Tennessee which we utilize heavily here, report a small percentage, less than
10%, of light transmitted to the forest floor. We assume this to reflect the
structure of the canopy rather than that of the crown, and take M for these
canopies to be 1 minus the percentage of light transmitted.
(2) Interception (h 0 ). An interception depth, h 0 = 0.1 cm, was assumed retained
on the horizontal projection (i.e., L t ) of all flat leaves and on all sides
(i.e., o L t ) of needle leaves. Note from Table 9.2 and Eq. 6.52 that with this
assumption the Sitka spruce canopy gets soil moisture recharge from only
2% of the rainfall. This small Ve leads to f I (ILSLt )ISW = 5 for the spruce which is
off the scale of Fig. 9.7. There must be another moisture source for this canopy.
(3) Carryover moisture storage (S). The presence of soil moisture in storage at
either the beginning (+S) or end (S) of the growing season is known for
only the Beaver Creek watersheds (Salvucci, 1992) and is reported here in
Table 6.3. However, the seasonal asynchronism of heat (i.e., growing season)
and water (i.e., rainy season) present at Beaver Creek is similar to that found
for the Sitka spruce at Spokane, Washington. We see this in Fig. 9.8 where the
monthly precipitation is much higher just before than during the growing
season at Flagstaff and at Spokane. In contrast we note in Fig. 9.8 that for
loblolly pine at Savannah, Georgia the monthly rainfall builds to a peak at the
conclusion of the growing season raising the likelihood of a negative carryover
moisture.
In addition to carryover soil moisture storage, Douglas fir (Waring and
Running, 1978), and Scots pine (Waring et al., 1979) are known to store up to
1.5 mm of water per day in the sapwood of the tree. This mechanism may be an
evolutionary response to the asynchronous seasonality just discussed and also
enables available soil moisture to be carried over either to or from the growing
season.
In moist habitats such as those of the Sitka spruce, drip from fog
condensation and direct absorption of condensation by the foliage (cf. Stone,
1957; Goodell, 1963) may provide the missing moisture supplement.
Table 9.2. Test of water propositiona
Latitude E ps(24) S
Number Species Site ( N) T0 ( C) m (d y1 : h d1 ) (cm d1 ) m h (cm) m m tb (d) M T Sr(20) (cm)
1 Beech Knoxville, TN 36 31.0(20) 202(28) : 12(4) 0.36(34) 0.98(7) 63.0(7) 3.21(8) 0.93(9) 0.25(3) 0.62(36) unknown
2 Oak Allentown, PA 41 23.4(20) 183(28) : 12(4) 0.37(34) 0.97(7) 61.0(7) 3.00(2) 0.95(9) 0.25(3) 0.52(36) unknown
3 Goethalsia Turrialba, Costa 10 23.0(14) 365(4) : 12(4) 0.34(27) 0.68(23) 365.0(4) 0.90(4) 1.00(4) 0.25(3) 0.71(37) unknown
Rica
4 Sitka spruce Scotlandb 47 27.0(20) 106(28) : 12(4) 0.41(34) 0.44(7) 22.0(7) 4.82(8) 1.00(9) 0.25(3) 0.80(38) unknownc
5 Ponderosa pine Flagstaff, AZ 34 14.2(2) 169(2) : 12(4) 0.31(2) 0.70(33) 31.8(8) 5.32(33) 0.83(2) 0.25(3) 0.34(36) +21.0(2)
6 Loblolly pine Savannah, GA 32 27.0(20) 78(28) : 12(4) 0.48(34) 1.46(7) 25.0(7) 3.12(8) 0.80(25) 0.25(3) 0.60(36) unknownd
7 Pinyonalligator Flagstaff, AZ 34 15.9(2) 182(2) : 12(4) 0.32(2) 0.75(33) 31.7(2) 5.74(33) 0.51(2) 0.25(3) 0.34(36) +4.70(2)
juniper
8 PinyonUtah Flagstaff, AZ 34 18.5(2) 215(2) : 12(4) 0.35(2) 0.63(33) 36.8(8) 5.86(33) 0.41(2) 0.25(3) 0.34(36) +6.13(2)
juniper
9 Creosote bush Las Vegas, NVe 37 11.0(20) 60(15) : 12(4) 0.38(35) 0.45(7) 8.0(8) 7.50(8) 0.09(16) 0.35(26) 0.39(38) unknown
ISL
(46)
(MJ(tot) E v (40) qb (39) ISW
rc (13) qb ISL
No. m2 h1 ) N (41) o = Lt Lt f I ( L t ) ra
ho cm (MJ m2 h1 ) (MJ m2 h1 ) E v
(MJ m2 h1 ) R(19)
b f I ( L t )ISW
1 0.32(1) 0.41 1.00 0.73(42) 4.4(29) 3.19(5) 0.30(11) 1.15(44) 0.42 0.70 0.13 0.19 1.10 0.19 0.97
2 0.46(1) 0.37 1.00 0.53(42) 4.9(29) 2.60(5) 0.36(11) 0.89(44) 0.36 0.85 0.10 0.12 1.26 0.15 1.01
3 0.44(1) 0.50(4) 1.00 0.79(12) 3.3(12) 2.61(10) 0.35(11) 1.95(44) 0.36 2.73 0.20 0.07 3.07 0.73 0.41
4 0.52(1) 0.73 2.00(9) 0.51(30) 5.1(30) 2.62(30) 0.35(11) 0.76(45) 0.43 0.88 0.14 0.16 1.36 0.30 1.09
5 0.60(6) 0.24 2.50(21) 0.28(42) 6.3(29) 1.77(5) 0.47(11) 0.51(43) 0.47 0.93 0.07 0.07 1.33 0.17 0.96
6 0.60(1) 0.36 2.36(22) 0.52(42) 5.0(29) 2.58(5) 0.36(11) 0.81(31) 0.59 1.28 0.12 0.09 1.87 0.12 0.89
7 0.60(6) 0.24 2.50(21) 0.52(42) 5.0(29) 2.58(5) 0.36(11) 1.31(43) 0.43 1.14 0.07 0.06 1.60 0.07 1.04
8 0.60(6) 0.24 2.50(21) 0.42(42) 5.4(29) 2.27(5) 0.40(11) 1.32(43) 0.33 1.00 0.07 0.07 1.42 +0.13 1.06
9 2.34(1) 0.22 1.00 0.37(17) 1.0(18) 0.37(10) 0.84(11) 0.75(43) 0.10 1.88 0.09 0.05 3.03 0.22 0.92
a b
Parenthetical superscript numbers refer to the source listing given in the notes below. Aberdeen, Scotland for ISL ; Spokane, WA for climate.
c d
We assume S = +0.5Pd as was observed for ponderosa pine at Flagstaff, AZ. We suspect a positive S due to the seasonal rainfall (cf. Fig. 9.8).
e
Las Vegas for T0(20) , Ely for precipitation (Appendix F).
Sources:
1. Table 8.2. 2. Williams and Anderson (1967).
3. Ross (1981, Fig. 112, p. 332). Equation 9.8 is formulated for unit vegetated area, thus we use T = 0.25 for dense plant stands rather than T = 0.16 at tree top (Iqbal, 1983,
Table 9.4.1, p. 289).
4. Estimated.
5. L t from Table 3.8: using lodgepole pine for loblolly, Northern white cedar for pinyonUtah juniper, and lodgepole pine for pinyonalligator juniper as suggested by sketches
of needle structure (cf. Zim and Martin, 1987, pp. 25, 36, 37, and 39).
6. Assumed identical to loblolly pine. 7. Appendix F. 8. m tb = m /m .
9. Whittaker (1966, Table II, p. 107); average of [1 % light penetration].
10. Product of separate and L t estimations. 11. Eq. 8.27 or Fig. 8.20. 12. Allen and Lemon (1976, p. 288).
13. Eq. 6.52 with h o = 0.1 cm in all cases. 14. Allen and Lemon (1976, p. 289).
15. MacMahon (1988, Fig. 8.5, p. 236), shows most favorable season to be JanFeb.
16. MacMahon (1988, Table 8.2, p. 240). 17. For Eucalyptus with = , (Whitford et al., 1995).
18. Desert shrub (Larcher, 1983, Table 3.13, p. 151). 19. This apparent Bowen ratio is calculated from Eq. 5.5. See text for discussion.
20. U.S. Weather Service (1974). 21. Waring (1983). 22. Raison et al. (1992a). 23. Whitford et al. (1995).
24. Gross seasonal rate (i.e., total for 24-h day during growing season).
25. Rogerson (1967), average of ten stands in Oxford, MS.
26. Ehleringer (1985, pp. 168169). 27. Budyko (1977, Map 34). 28. Fig. 3.11. 29. Fig. 7.4.
30. Table 2.2 for = and L t = aw , average for all spruce; L t = L t /.
31. As in Note 43 but divided by 2.36 to account for parallel flux from other stomated needle surfaces (cf. McNaughton and Jarvis, 1983; Jarvis and McNaughton, 1986).
32. Average for maple; Table 2.4. 33. Salvucci (1992); Table 6.4. 34. Kohler et al. (1959, Plate 2).
35. Farnsworth et al. (1982, Map 2) gives MayOct. E ps = 1.14 m; Kohler et al. (1959, Plate 2) gives full-year E ps = 1.83 m; thus, winter E ps = 0.69 m/180 days.
36. July @ 1 pm. 37. Allen and Lemon (1976, Table III, p. 287). 38. January @ 1 pm.
r
39. qb qb [1 ( 1S
0.85
)] with qb from Eq. B.12 using daily minimum (i.e., night-time) radiation temperature.
2Ve
40. E v Mm with Ve and m tb from Eqs. 6.67 and 6.63 respectively. The factor 2 reflects evaporation during daylight hours only.
tb
41. U.S. Weather Service (1974) using N = 1 % possible sunshine. 42. L t /L t . 43. Eq. 7.25 and the data in the caption of Fig. 7.9.
44. Eq. 7.13. 45. Fig. 7.6 for d/t = 0.006. 46. Eq. 9.8
290 Natural habitats and climax communities
Growing
season
1.2 Growing
season
0.8
0.8
Oak
0.4 at
Sitka spruce Allentown, Pennsylvania
0.4
10
at
10
Spokane, Washington 0
Monthly precipitation
0
Annual precipitation
Growing
Monthly precipitation
Annual precipitation
Growing season
season
1.6 1.2
1.2 0.8
Ponderosa pine
0.8 0.4 at
Flagstaff, Arizona
Loblolly pine
0.4 at 0
2 4 6 8 10 12
Month of year
0
2 4 6 8 10 12
Month of year
We resolve this uncertainty for the Sitka spruce rather arbitrarily by relying
on the similarity (cf. Fig. 9.8) of seasonal phases at the beginning of the
growing season and making the assumption that for the spruce S is the same
+50% of the dormant season precipitation as is observed for the ponderosa pine
(cf. Table 6.3). This assumption gives S = +17 cm for the spruce, making
ISL
f I ( L t )ISW
= 1.09 as is plotted in Fig. 9.7.
(4) Light decay coefficient ( L t ). Although admittedly inaccurate, Bakers (1950)
values of L t were used for matching species. Otherwise, separate estimates
of = and of L t were made using Tables 2.2 and 2.3. The Beaver Creek
communities have a continuous variation with topographic elevation
(Baker, 1982) from ponderosa at the top, through pinyonalligator juniper
down to pinyonUtah juniper and finally to desert shrub at the bottom. Not
knowing L t for either pinyon or juniper, we use the physical resemblance of
the needle structure to represent the pinyonalligator juniper as pinyon using
Bakers (1950) L t for lodgepole pine, and to represent the pinyonUtah
juniper as juniper using Bakers (1950) L t for Northern white cedar (see Zim
and Martin, 1987, pp. 36 and 39 for drawings of the respective needle
structures).
(5) Wet surface potential evaporation (E ps ). The gross annual water surface
evaporation was estimated for the given location from climatological maps and
then reduced to the appropriate season using the ogive distribution of Salvucci
(1992) presented here as Eq. B.13.
Proposition no. 3: Water 291
(6) Sensible heat flux (H ). Because of the large uncertainty in Ve and hence in
E T , we are uncertain of the actual values of Rb , and the values of Rb
estimated in Table 9.2 using Eq. 5.5 are termed apparent Bowen ratios. The
water balance components are best known for the pinyonjuniper canopies
where we have no evidence of sapwood storage capability and S was
observed. Indeed, the apparent Bowen ratios for these canopies are closely
zero and are assumed to be actual. All other apparent Rb are negative
indicating an apparent transfer of sensible heat from the atmosphere to the
vegetation in augmentation of the net radiation. This is likely for the moist
canopies such as oak, beech, Goethalsia, and Sitka spruce, and even for the
loblolly and ponderosa although the actual magnitudes may be different due to
the water balance uncertainties discussed above. The driest, creosote bush
canopy, has a winter growing season and may well be accepting heat as the
negative Rb indicates. In any event, with the exception of the rainforest (i.e.,
Goethalsia), and beech, the apparent Rb satisfy the criterion Rb < O(101 )
which is consistent with our finding in Fig. 9.4 that T0 T l = Tm .
The results are presented in Fig. 9.7 where we see that according to this proposition
and the assumptions discussed above, seven of the nine communities are in their
natural habitat. Note in particular that only the loblolly pine fails to satisfy closely the
conditions for climax communities specified by Eqs. 9.13 and 9.14. It seems possible
that buffered stress tolerance (see Baker, 1950, pp.140143; Whittaker, 1975, p. 179)
may be responsible for this deviation. Our reasoning follows.
For simplicity and convenience, we have developed the habitat constraints using
the approximation that random variables are replaced by their mean values. However,
this formulation does not account for the fact that while in the average a species may
satisfy the climax conditions, the distributed nature of the climatic variables guarantees
that in some years the species will be under stress in this habitat. The ability of certain
species to survive the runs of stressful seasons characteristic of climatic variability
is often called the tolerance of that species (cf. Baker, 1950, pp. 140143) and is a
function of both species and habitat. To survive, the species must be in a mean state
that falls short of optimum productivity by an amount providing a margin of safety
to accommodate the resource shortages of extreme seasons. Such a stable but sub-
climax state appears to be one of so-called buffered tolerance (Whittaker, 1975, p. 45)
in which the departure (i.e., the buffering) from climax conditions depends upon the
tolerance that the particular species has for stress. It is interesting to note that Baker
(1950, Fig. 12, p. 143) lists lodgepole (similar to loblolly) pine among the intolerant
species. This classification is consistent with the behavior demonstrated by this species
in the habitat plot of Fig. 9.7. The degree of buffering provided is probably controlled
by the population density, M. That is, for a given value of the ordinate, f I (ILSLt )ISW , the
amount of water available, on the average, to each leaf can be increased by decreasing
the canopy cover, M.
292 Natural habitats and climax communities
Plate 9.3. Nurselog. Fallen fir tree nourishes new growth in the Olympic National Park,
Washington. (Photograph by William D. Rich; Copyright c 2001 William D. Rich.)
Table 9.2 shows that the second square-bracketed term on the right-hand side of
Eq. 9.15 may be neglected with respect to the first, leaving
(1 T )Psm / o rc
E v 1 + 1+ , (9.16)
f I ( L t ) ra
which is helpful to our reasoning. Note that, except for minor T variation with , the
left-hand side of Eq. 9.16 is a species constant as is the minimum resistance ratio on the
right-hand side. Climate is contained within E v and o . Two significant observations
flow from this:
(1) Figure 7.3 shows that the resistance ratio for a given species varies strongly
with without significant sacrifice of optimality. Thus, a given species can
maintain maximum productivity (i.e., maintain W = 1) over a range of
climates by varying while simultaneously varying L t so as to keep L t at the
species-constant value. This ability may account for some of the scatter of
around the theoretical value, = , observed in Fig. 3.18.
Conclusion 293
(2) Nutritional deficiency decreases the leaf area, L t . If increases to maintain the
species-constant L t , the resistance ratio increases (cf. Fig. 7.3) making
W < 1. With desert and rainforest soils notoriously short of nitrogen, can this
be the reason for the sub-climax position of the creosote bush and the
Goethalsia in Fig. 9.7?
E Conclusion
On the basis of the evidence presented, all three propositions are strongly supported,
and their combination, as presented in Fig. 9.7, defines both the feasible natural habitat
of species with given ISL and L t , and the climax limits thereto.
10
For cylindrical multilayer crowns of given leaf angle and without local carbon
recycling, there is a critical L t = L t separating species productively limited
by carbon demand from those limited by atmospheric carbon supply, and at
which the productive gain has its global maximum. For tapered multilayer
crowns the atmospheric carbon supply exceeds demand at all values of L t .
Collected productivity observations from the literature support the theory
over a wide range of plant communities from desert shrub to rainforest with
the wet communities clustering around L t .
Normalizing the productivity by the local climatic productive potential
places all species dependence in a new function of L t that displays a global
maximum over the range of L t corresponding to that observed in nature and
thereby confirming our fundamental hypothesis that selection favors increasing
productivity.
Equating the productivity of unit plant areas for three competing species is
shown to predict well the latitude of the mixed forest ecotone of eastern North
America.
294
Canopy biomass production 295
Plate 10.1. Collection of pines. Left to right: Himalayan pine, Cevennes pine, and
Japanese red pine. (Photographed in Arnold Arboretum, Boston, by William D. Rich;
Copyright c 2001 William D. Rich.)
and nitrogen in order for the plant to fulfill its productive potential. Here we assume a
nutrient supply that is non-limiting and climatic climax conditions (cf. Chapter 8) that
guarantee light saturation and sufficient water supply to keep the stomates fully open.
We now formulate the production of canopy biomass in terms of carbon supply and
demand.
Carbon balance
The accumulated green plant biomass (plus any net transfer of organic carbon from
the green plant compartment to other compartments within the ecosystem) is termed
the net primary productivity or NPP of the community (Bormann and Likens, 1979,
p. 16), and is normally quoted on an annual basis. Larcher (1983, p. 146) expresses
this carbon budget as
NPP = B + C + D, (10.1)
With NPP measured in grams of dry solid matter (i.e., gs ) and GPP measured in
grams of CO2 assimilated, we write the ratio
NPP 1
= , (gs g(CO2 )1 , or just gs g1 ), (10.3)
GPP R
1+
NPP
which was introduced here in Chapter 8. It is a necessary empirical quantity for
our work and is variable among plant types. Larcher (1983, p. 149) gives =
0.400.67 gs g1 for temperate forests, and 0.25 gs g1 for moistwarm tropical
forests. Furthermore, data presented by Larcher (1983, Table 3.12, p. 149) show that
for these two types of deciduous trees at least, the value of is less than 10% different
if the contributions of the roots are omitted from both R and NPP in Eq. 10.3. Here
we will not attempt to include the variability of with plant type and will instead use
the single value
= 0.50 gs g1 , (10.4)
reported by Penning de Vries et al. (1974) and by Ledig et al. (1976, fide Russell et al.,
1989b, p. 27). Consistent with this approximation then, we may also take the chosen
value of to represent gs (above-ground) g1 . We will find this alternate interpretation
convenient because the usual field observations of forest productivity omit the root
biomass.
To formulate canopy NPP we begin at the leaf level following the developments
of Chapter 8. Maximization of plant NPP is taken to be the driving force behind the
natural selection processes establishing the monocultures with which we deal on the
basis of the assumption that NPP is proportional to seed productivity and hence is a
surrogate for survival probability, all else being constant. This proportionality may
vary with species.
The data compiled by Cannell (1982) indicate that between 65% and 87% of total net primary
production is above-ground production.
Canopy biomass production 297
Psm will change with changing ambient CO2 concentration.
Here, unlimited implies that the natural, ambient CO2 concentration is available at every leaf
surface.
298 Net primary productivity and ecotones
1
P 1
exp e L t d.
D
= 1 (10.10)
g(T0 ) Psm 0 fI
To perform this integration, we make the substitution
e L t = u, (10.11)
in which
and for 1 z < , Abramowitz and Stegun (1964, p. 231, no. 5.1.54) give the
approximation
1 z 2 + a1 z + a2
E 1 (z) = z 2 + (z) , (10.17)
z e z + b1 z + b2
Professor David Benney, personal communication.
Canopy biomass production 299
where
a1 = 2.334733, b1 = 3.330657,
a2 = 0.250621, b2 = 1.681534,
|(z)| < 5 105 .
We now compare this above-ground annual canopy carbon demand with that of a
groundcover monolayer of this species having = 1, L t = 1, which has the above-
ground annual carbon demand
PT PT
= = fD L t, , (10.20)
pD g(T0 ) Psm Mm
NPP NPP
= = f D ( L t ). (10.22)
pD g(T0 ) Psm Mm
Equation 10.22 is plotted as curve (a) in Fig. 10.1 where f D ( L t ) may be viewed al-
ternatively as the above-ground dimensionless carbon demand function for the given
light environment. Note that f D ( L t ) satisfies all the conditions for optical optimality
for a given species as given in Chapter 3 and is independent of all other aspects of
climate as well as of crown shape. Note also that for the diffuse monolayer, L t = 1,
f D ( L t ) < 1, as compared with the groundcover monolayer (i.e., = 1, L t = 1)
where NPP = pD as is shown by curve (f) in Fig. 10.1.
Note that the ability of the plant to use light depends upon the supply of nitrogen (Sharkey, 1985),
while its ability to use CO2 depends upon the adequacy of the water supply to maintain fully open
stomates as specified by ISL f I ( L t )ISW . This latter condition is expressed in Eq. 10.20 through
the separable variable, M.
Actually of course, there is an upper limit to L t in the neighborhood of L t = 4 to 5 imposed by
compensation light requirements (cf. Table 3.9).
Remember that here Psm has the units of g(CO2 ) m2 h1 .
With the exception of PAR which is insensitive to for these h , and is thus fixed by the leaf
texture of the species.
Canopy biomass production 301
(e) (d)
(a)
0.5
2.0 0.7
=
=
=0 6 8
.3
4
7
3
CO2 and 5
opy
p 1.0
(f ) p = pD .5 (b)
12
1 2
=
(c)
e r, y
ov op
nd
c can
ou
gr
1 11
10
Fig. 10.1. Productive gain of C3 multilayer canopies. (a) CO2 demand (Eqs. 10.21 and
10.22); (b) atmospheric CO2 supply, cylindrical crown with = 0.3 (Eqs. 10.30 and
10.31); (c) atmospheric CO2 supply, cylindrical crown with = 0.5 (Eqs. 10.30 and
10.31); (d) atmospheric CO2 supply, hemispherical crown with = 0.7 (Eq. 10.33);
(e) atmospheric CO2 supply, conical crown with = 0.5 (Eq. 10.34); (f) CO2 demand,
groundcover with = 1.0, NPP/ pd = L t . Data points from Table 10.1: 1, Creosote
bush, Nye Co., Nevada; 2, ponderosa pine, Tucson, Arizona; 3, loblolly pine, Tucson,
Arizona; 4, oak, Great Smoky Mountains, Tennessee; 5, rainforest, Kade, Ghana; 6,
rainforest, Yangambi, Congo; 7, red spruce, Great Smoky Mountains, Tennessee; 8,
beech, Eastern Europe; 9, beech, Great Smoky Mountains, Tennessee; 10,
pinyonalligator juniper, Flagstaff, Arizona; 11, pinyonUtah juniper, Flagstaff,
Arizona; 12, sugar maple, Great Smoky Mountains, Tennessee.
We are assuming neutral atmospheric stability in the long-term average because it is unwise to
calculate mass fluxes using eddy momentum diffusivities under conditions of atmospheric
instability (Jarvis et al., 1976, pp. 224225).
302 Net primary productivity and ecotones
in which K m (0) is the diffusivity at the top of the canopy, = 0. Precise evaluation of
the CO2 concentration gradient is problematic. We will approximate it by assuming
that its average over the crown depth increases with increasing foliage density in the
manner
dc c0 cs
= (1 e L t ), (10.25)
dz h hs
in which c0 and cs represent the concentrations at the uppermost and lowermost leaves
respectively. For homogeneous canopies it follows that
h h s = kc L t , (10.26)
It is the elimination of unknown proportionality constants such as (c0 cs )/kc which motivates
presenting these supplydemand comparisons in terms of dimensionless ratios.
Canopy biomass production 303
trees respectively (cf. Fig. 3.18). Alternatively, we may call f S (, L t ) the above-ground
dimensionless carbon supply function. Note that the maximizing behavior of Eq. 10.31
is characteristic of turbulent vertical mass flux in the closed canopy of homogeneous
cylindrical multilayer crowns.
With vanishing canopy, the concentration gradient at the monolayer surface is max-
imized and it seems reasonable to assume that the monolayer carbon supply, ps , adjusts
to meet the monolayer carbon demand, pD . With ps = pD , we may view Fig. 10.1 as
a comparison of carbon supply and demand under the same conditions.
We see from Eq. 10.24 that K m 0 as L t gets large due to internal resistance of
the foliage elements, while from Eq. 10.25, dc dz
0 as L t 0 due to the decline of
sink strength. Thus the concentration gradient controls the turbulent flux at low L t
while the diffusivity controls that flux at large L t with a maximum in between. In
Chapter 7 we found a minimum in crown resistance which, together with Eqs. 5.29 and
5.30, dictated a maximum in plant transpiration rate at intermediate values of L t .
Note that for = 0.3, Fig. 10.1 shows the dimensionless CO2 supply rate to max-
imize at L t 1.2, while we found in Fig. 7.3 that at this same , the dimensionless
transpiration rate, kv , maximizes (i.e., rc
/ra minimizes) at L t 1.95. Equations 10.23
and 10.24 apply equally to water vapor and to CO2 , so the difference in the peak of
the two flux curves must lie with differences in the vertically averaged concentration
gradients of the two quantities. Indeed, the source concentration of water vapor is as-
sumed at saturation in all stomates throughout the crown, while the sink concentration
of CO2 in the air is presumed to decline with depth into the crown.
Lt
(1 e )2
=
L t 2
fD ( L t ). (10.32)
1e
Equation 10.32 is plotted as the solid line in Fig. 10.2 in comparison with the
optically optimal foliage state for several species (dashed hyperbolas) presented earlier
Note that maximization of the plant moisture flux rate, E v kv E ps , through maximization of kv ,
does not affect the canopy moisture flux rate, M E v = Mkv E ps , in a water-limiting climate,
because there the product, Mkv , is a climatic constant (cf. Eq. 6.66). However, because this
product is invariant in a given water-limiting situation, maximizing kv means minimizing M, and
the canopy biomass will decline (cf. Eq. 10.22). Thus evidence such as Fig. 7.11 in support of
maximum kv supports the concept of the selfish gene (Dawkins, 1976).
304 Net primary productivity and ecotones
rwa
8
ma
ys
pru
Carbon Carbon
k
Lt 6
demand
control
supply
control
average of and from
Tables 2.22.4.
, Deciduous;
4 Lt , evergreen;
, crops and grasses.
0
0 0.2 0.4 0.6 0.8 1.0
Canopy biomass production 305
Equations 10.33 and 10.34 are plotted in Fig. 10.1 as dotted curves (d) and (e) re-
spectively using values of representative of the hemispherical crowns common to
rainforests (i.e., = 0.70), and conical evergreens (i.e., = 0.50).
Before discussing the very revealing Fig. 10.1, a few comments concerning the
other major ingredients of production are in order.
M E v = Mkv E ps . (10.36)
in which Ve is the average available interstorm soil moisture, and m tb is the net in-
terstorm duration, the canopy adjusts its water demand to equal the climatic supply
by setting M < 1. Otherwise, M = 1, and there is an excess of water supply which
runs off through the natural surface water and/or groundwater drainage systems. In
this way, through self-regulation, the canopy water supply always equals or exceeds
its water demand.
Actually of course, the ecological pressure for this adjustment arises from the individual plants
through the demise of the weakest ones due to stress induced by water shortage.
Canopy biomass production 307
Plate 10.5. Riverside community. This community along the Ayeyarwady (Irrawaddy)
River in Myanmar (Burma) contains a giant hemispherical tree of unknown species
within a grove of Royal palms. (Photograph by Peter S. Eagleson.)
concentration over a wide range of the latter. Here then is a selection pressure for
maximizing the flux rate of the solvent water through the plant.
We begin our analysis of nitrogen supply by assuming the reservoir of soluble
nitrates in the soil to be large enough to impose no limit on their solution in the
soil water. The concentration of nitrate in the soil water will then be fixed by the
kinetics of the water uptake which is under the control of the climate, soil, and
vegetation.
The plant water supply equals the plant water demand which is given by Eq. 10.35
as
E v = kv E ps . (10.38)
Ns = CN kv E ps , (10.39)
in which CN is a constant for given climate, soil, and vegetation. A diffuse monolayer
under these same conditions will have the the nitrogen supply
n s = CN kvm E ps , (10.40)
in which kvm is the monolayer conductance. With Eq. 5.29, the dimensionless supply
of nitrogen and all other minerals is then
rc
1+ +
NS kv 0 ra m
= = . (10.41)
nS kvm rc
1+ +
0 ra
Unfortunately, our big-leaf model for the resistance ratio, rc /ra , does not hold at
the monolayer limit, L t = 1. However:
(1) For cylindrical crowns we infer from Fig. 7.3 that irrespective of , (r c /r a )m
must become large as L t 1. As noted above, there is a selection pressure to
maximize NS /n S and hence to minimize r c /r a . Thus, for the common observed
range of (i.e., 0.3 0.5), (r c /r a )min = 0.34 at L t = 1.8, and
(r
c /r a )min = 0.81 at L t = 2.5 (cf. Fig. 7.3), and for normal temperatures,
1 + /0 3 (cf. Table 5.1). The denominator of Eq. 10.41 therefore
increases modestly with L t while the numerator stays constant. It follows
then that NS /n S > 1 for all L t .
(2) For tapered crowns we infer from Fig. 7.5 that the denominator of Eq. 10.41
decreases with L t , and that therefore NS /n S > 1 for all L t and increases
with L t .
(1) Potential carbon supply equals or exceeds carbon demand for full local carbon
recycling and atmospheric carbon replaces that exported.
(2) The CO2 (and nutrient) demand is independent of crown shape.
(3) The CO2 supply is dependent upon crown shape.
(4) For crown shapes producing a rough canopy top (e.g., tapered shapes such as
cones and hemispherical segments) CO2 supply exceeds CO2 demand for all
values of L t for both multilayer and monolayer foliage. Therefore, the
demand function governs for these crowns at all L t .
(5) For cylindrical crowns only, the CO2 supply rate maximizes at low L t making
supply equal demand at a critical intermediate value of L t denoted
L t,
310 Net primary productivity and ecotones
which varies inversely with . At low L t for such crowns the relative canopy
resistance, rc /r a , gets very large (cf. Fig. 7.3) making the canopy conductance
and hence the transpiration rate correspondingly small. These are conditions
for low water use and regions of low water supply should favor such plants.
We have seen (cf. Chapter 8) that in water-limited situations, water demand
(i.e., use) is maintained equal to water supply (at given L t ) through
adjustment of the canopy density, M.
(6) For L t <
L t , CO2 supply exceeds demand for all crown shapes and the
demand function governs the productivity which increases with increasing L t .
All else being constant, all crown shapes should coexist in this range of L t .
(7) For L t > L t , CO2 demand exceeds supply for cylindrical crowns, and the
supply function governs the productivity which decreases with increasing
L t . For tapered crowns the demand function continues to govern. All else
being constant, tapered crowns will bring maximum dimensionless
productivity in this range of L t . Without supplemental CO2 supply such as by
regular fog or rain (cf. Wilson, 1948), or by the augmented mixing provided by
wind-induced leaf motion, only tapered crowns will follow the CO2 demand
curve in this region.
(8) For L t =
L t , the dimensionless productivity of cylindrical crowns has its
global maximum.
(9) For L t 1, a monolayer groundcover, curve (f), has greater productivity than
a monolayer canopy.
(10) For 1 < L t < 1.7, a canopy has no productive advantage over the monolayer
groundcover, L t = 1. Can it be that this window of L t is vacant of natural
canopies?
(11) For all L t , r
c /r a will be minimized in order to maximize the supply of
nutrients to each plant.
(12) For L t <
L t , where CO2 supply exceeds demand, selection pressure is on the
photochemistry to develop more efficient assimilation mechanisms.
(13) For L t
L t , where CO2 demand exceeds supply, selection pressure is on
crown shape to develop the tapered form that facilitates vertical fluxes.
(14) Water-limited communities have a more efficient CO2 flux path and thus a
higher CO2 supply curve than those for the closed canopies plotted on
Fig. 10.1. Therefore, we expect water-limited communities to be CO2
demand-limited. Under such an excess of supply over demand, natural
selection has produced more efficient photochemical processes: the C4 process
NPP limitation through foliage-induced CO2 flux constraints seems to be a new finding. Larcher
(1983, p. 85) states that CO2 supply is often a yield-limiting factor for both land and water plants
but he does not identify the constraint. Later in the same work (Larcher, 1983, p. 154) he writes,
On the continents it is almost always the water supply that limits yield, though other factors are
nutrient deficiency and . . . shortening of the production period due to cold.
Canopy biomass production 311
(found in sugar cane, maize, millet, and many savanna grasses; cf. Larcher,
1983, p. 79), that operates in a biophysically similar manner to the C3 plants
but generates much higher productivities, and the CAM process (found in
succulents such as cactus) that is fundamentally different by virtue of its
nocturnal assimilation of CO2 . Note that the demand-NPP as given by
Eq. 10.18 contains the product, Psm M. For C3 and C4 plants whose stomates
are open during the day, Eqs. E.9 and 6.46 give Psm E v E ps , while for
1
water-limited systems, Eq. 6.70 gives M E ps . Thus under water limitation
Psm M = constant. This removes the selection pressure for increasing Psm in
either C3 or C4 plants in the presence of excess carbon supply when water is
limiting. We can then expect M < 1 climax communities to have species in the
range L t <
L t . Stated in another way, L t < L t is a necessary condition for
water limitation. It is not a sufficient condition since it specifies no
climate.
(15) Following the above, species for which L t L t will be light limited. For
light-limited communities, M 1, and the NPP as given by Eq. 10.18 remains
sensitive to changes in Psm . This preserves the selection pressure to close the
potential supplydemand productivity gap. We expect that climax communities
will have closed this gap resulting in M = 1 climax communities having
species in the range L t L t.
(16) Items 14 and 15 above lead heuristically to
L t as a necessary condition
separating the species of water-limited (M < 1) and light-limited (M = 1)
communities.
Productivity observations
In Table 10.1 we have collected productivity observations for a small set of vegetation
communities from a variety of sources as described in the tables footnotes. As was
the case in Chapter 9, our emphasis in the data collection has been to test the theory
over the range of communities from desert to rainforest rather than to assemble masses
of observations for the commonly studied systems. Table 10.1 with its accompanying
notes is self-explanatory except that in plotting the points on Fig. 10.1 we have used
symbols evoking the crown shape according to our best estimate. That is, squares
are used for assumed cylindrical crowns, triangles for assumed conical crowns, and
semicircles for assumed hemispherical crowns. Tubbs and Houston (1990, p. 329)
describe American beech as having a round-topped head; photographs of red spruce
(Blum, 1990, Fig. 3, p. 254) show a conical form; Fechner (1990, p. 243) describes
the blue spruce as pyramidal in form; Odum (1975, Fig. 7-10, p. 197) pictures the
rainforest canopy-top as a set of spherical segments; and Godman et al. (1990, Fig. 2,
p. 80) show the sugar maple to have a cylindrical crown. Other than the beech, spruce,
and rainforest, we assume all crowns to be cylindrical. In making the comparison
Table 10.1. Observations of canopy productivitya
NPP
Number Species Site NPPb (gs m2 y1 ) Psm (gs m2 h1 ) M m (d y1 : h d1 ) Lt Lt Psm m M
1 Creosote bush Nye Co., NV 39.8(1) 2.08(2) 0.09(3) 60:12(4) 0.37(5) 1.0(6) 0.37 0.30
2 Ponderosa pine Tucson, AZ 61,200(7) 0.40(8) 0.83(9) 169:12(10) 1.77(11) 0.91
3 Loblolly pine Tucson, AZ 435(12) 0.40(8) 0.80(4) 80:12(13) 0.52(39) 5.0(38) 2.58(14) 1.41
4 Oak Gt. Smoky Mts., TN 963(15) 0.28(2) 0.95(16) 202:12(13) 0.53(39) 4.9(38) 2.60(11) 1.49
5 Rainforest Kade, Ghana 2,188(17) 0.38(18) 1.00(19) 365:12(20) 0.79(21) 3.3(21) 2.60 1.31
6 Rainforest Yangambi, Congo 2,884(22) 0.38(18) 1.00(19) 365:12(20) 0.79(21) 3.3(21) 2.60 1.73
7 Red spruce Gt. Smoky Mts., TN 950(23) 0.55(2) 0.92(24) 85:14.5(25) 0.51(26) 5.10(26) 2.62(26) 1.52
8 Beech Eastern Europe 1,300(27) 0.33(2) 0.93(28) 202:12(13) 3.19(11) 1.75
9 Beech Gt. Smoky Mts., TN 883(29) 0.33(2) 0.93(28) 202:12(13) 3.19(11) 1.19
10 Pinyonalligator juniper Flagstaff, AZ 0.19 to 0.25(30) 0.40(8) 0.52(31) 182:12(31) 1.68(32) 0.25 to 0.19
11 PinyonUtah juniper Flagstaff, AZ 0.20 to 0.26(30) 0.40(8) 0.41(31) 215:12(31) 2.27(32) 0.26 to 0.20
12 Sugar maple Gt. Smoky Mts., TN 1,340(33) 0.69(34) 0.96(35) 202:10(36) 0.35(37) 3.91(11) 1.00
a
Parenthetical superscript numbers refer to the source listing given in the notes below.
b
These values are above-ground and do not include roots. From Table 8.4 we see that if roots were included the total NPP would be 10% to 40% higher.
Sources:
1. Rosenzweig (1968, Table 1, p. 69, Code A). 2. Table 8.2. 3. Average of five sites in southern California (McMahon, 1988, Table 8.2, p. 240). 4. Table 9.2.
5. Eucalyptus (Table 2.3). 6. Desert shrub (Larcher, 1983, Table 3.13, p. 151). 7. Cannell (1982, p. 311). 8. Loblolly pine (Table 8.2).
9. Williams and Anderson (1967). 10. Salvucci (1992, Table 4.5). 11. Table 3.9. 12. Whittaker and Marks (1975, Table 46, pp. 8081). 13. Fig. 3.11.
14. Lodgepole pine (Table 3.9). 15. Whittaker (1966, Table VI, p. 111, average of Sample Numbers 20, 21, 27, and 28).
16. Whittaker (1966, Table II, p. 107, average of Sample Numbers 20, 21, 27, and 28). 17. Nye (1961). 18. Goethalsia (Table 8.2). 19. Assumed.
20. Costa Rica; Allen and Lemon (1978, pp. 298, 300). 21. Allen and Lemon (1978, p. 288). 22. Bartholomew et al. (1953).
23. Whittaker (1966, Table II, p. 107, average of Sample Numbers 29 and 30). 24. Whittaker (1966, Table I, p. 106), from average light penetration.
25. Fig. 3.11 for cool multilayers; white pine and red spruce cohabit (Moore, 1917); Table 3.5 with full daylight length due to high elevation.
26. Table 2.3, = , and L t = aw , average for all spruce; L t = L t /. 27. Drozdov (1971, Table 1, p. 55).
28. Whittaker (1966, Table II, p. 107, average of Sample Numbers 25 and 26). 29. Average of Bray (1964) and Whittaker (1966, Table II, p. 107, sites 15, 25, 26).
30. Clary et al. (1974). 31. Table 6.3.
32. From physical resemblance of needle structure (cf. Zim and Martin, 1987, pp. 36 and 39) we represent the pinyonalligator juniper as pinyon using Bakers (1950) L t for
lodgepole pine, and we represent the pinyonUtah juniper as juniper using Bakers (1950) L t for Northern white cedar.
33. Cannell (1982, p. 248). 34. Table 8.2: Average for deciduous trees. 35. Whittaker (1966, Table II, p. 107, Sample Number 24).
36. Fig. 3.11 with estimated shortening of day length due to low elevation and northern exposure. 37. Table 3.7 for maple. 38. Fig. 7.4. 39. L t /L t .
Canopy biomass production 313
between theory and observation, we keep in mind that the former assumes atmospheric
carbon supply and contains other major approximations:
(1) The form of the L t -dependence of the CO2 concentration gradient has been
reasoned rather than derived (and only for M = 1), and this form is crucial to
the shape of the supply curve. While analogy with the derived canopy
conductance, kv , demonstrates that the supply curve clearly must peak, our
location,
L t , of the important supplydemand equality must be approximate.
(2) We have in effect assumed that the annual net increase of above-ground
biomass is a constant percentage of the annual assimilated CO2 for all forest
types. This expedient permits us to use observed annual above-ground
biomass production to evaluate the theory but it limits us to qualitative
conclusions.
Note in Fig. 10.1 that except for the two pinyonjuniper canopies, all observations
are in reasonable agreement with the theory. In particular, we see a cluster of data points
in the vicinity of the global optimum where for cylindrical crowns with = 0.3, supply
equals demand. Furthermore, beyond this point the only canopies following the demand
curve are two apparently tapered crowns as predicted. Productivity observations in the
range L t < L t are few but as is shown in Fig. 10.1 and Table 10.1, they support L t <
L t being water-limited with M < 1 and L t L t being light-limited with M = 1.
It is particularly interesting to note that the literature finds C3 trees to be carbon
supply-limited at the current atmospheric CO2 concentration (Larcher, 1983, p. 85;
Sharkey, 1985; Gunderson and Wullschleger, 1994; Drake et al., 1997) which would
imply a direct response to anthropogenic CO2 increases. Here we find, by both theory
and limited observation, that carbon is limiting only beyond a critical basal leaf area,
L t , and then only for non-tapered crowns.
The general agreement of observation with theory supports our earlier tentative
conclusion that with an adequate nutrient reservoir, the supply of nutrients will not
limit productivity. However, the soil at the Beaver Creek site of the pinyonjuniper
canopies is described (Williams and Anderson, 1967) as a very stony clay derived
from basalt and volcanic cinders, rich in exchangeable calcium and magnesium, but
low in potassium. The nitrogen content ranges from 0.17 to 0.02 percent. It seems
possible that the productivity of these soils may be limited by the available reservoir
of either potassium or nitrogen.
In evaluating the comparisons of observations with theory we must remember that
we have not developed a theoretical predictor for m , the length of the growing sea-
son, and have estimated this important factor using a somewhat uncertain empirical
correlation (cf. Fig. 3.11). Season length is limited by complicated biochemical pro-
cesses beginning with bud-break in the spring which is commonly forecast through
some measure of accumulated heat (Valentine, 1983; Lechowitz, 1984). The season
ends in the autumn with leaf senesence. Transpiration (Gee and Federer, 1972) and
photosynthesis (Mitchell, 1936) cease as chlorophyll and other nutrients are removed
from the leaves before they fall. Our estimates of m may introduce considerable error
into the observational ordinates of Fig. 10.1.
314 Net primary productivity and ecotones
introduce the climate supportive of this carbon demand we use Eq. 8.31 (confirmed
in Fig. 8.11), and the bioclimatic light optimum of Eq. 9.2 (confirmed in Fig. 9.5),
along with the optical optimum Eq. 3.52 (confirmed in Fig. 3.19) to replace Psm by
the climatic assimilation potential
canopy productivity. For large L t , the high foliage density causes heavy light decay
and the falling average light intensity, f I ( L t ), dominates the increasing leaf surface
so that canopy productivity falls. Importantly, there is an intermediate range of L t in
which the dimensionless productivity has the almost constant global maximum with
the approximate value
NPP
0.51. (10.45)
I0 m max
Note that at least for = 0.3 as plotted on Fig. 10.1, L t = 2.65 lies in the center of
the productivity peak of Fig. 10.3.
In Appendix H we show that L t = L t is a species constant, and in Table 3.9
we relate L t to species through Eq. 3.53 and the compensation point observations of
Baker (1950). This allows us to associate L t with specific species and to draw the
vertical dashed lines on Fig. 10.3 representing the full range of species observed by
Baker (1950). Note that 10 of the 13 species lie in the range of the global maximum
productivity. We take this as observational confirmation of natural selection pressure
to maximize survival probability as represented by its surrogate, NPP.
Extreme climates may dictate extreme L t which are locally but not globally opti-
mum productively. For example, hot, dry growing seasons demand low L t to reduce
water use (e.g. L t = 0.37 for creosote bush; not plotted in Fig. 10.3), and cold but wet
growing seasons demand high L t to conserve heat (e.g. L t = 3.9 for sugar maple).
B Ecotones
Denition
An ecotone is a transition zone between different plant communities (Colinvaux, 1973).
Because they represent marginal conditions, we expect ecotones to be sensitive to
changing climate (Arris and Eagleson, 1994). Therefore, a necessary condition for the
utility of the present work in anticipating the effects of future climate change is the
ability of our analyses to locate principal existing ecotones.
The proposition
Let us assume that in the battle for survival, the plant with the highest productivity
wins due to its superior seed production. Let us also assume, for purposes of this
illustration, that the gradient of climatic factors controlling productivity of a given
species is expressed primarily through latitude. Extremes of latitude should depress
productivity due to the temperature control of photosynthesis leaving a preferred in-
termediate latitude at which the productivity is maximum. In the competition between
two monocultures, the geographical productivity variation is as sketched in Fig. 10.4
(Arris and Eagleson, 1994, Fig. 3, p. 4). Under these assumptions the local productivity
maximum shifts species where the two curves of Fig. 10.4 cross thereby marking the
ecotone location.
Ecotones 317
A test
As is shown in Fig. 10.5, the forest formations of eastern North America display
ecotones which are approximately latitudinal. The shaded region embracing the Great
Lakes is the broad, mixed forest transition (i.e., ecotone) between the evergreen boreal
(primarily spruce) forest and the deciduous (primarily oak) forest. The thinner shaded
band represents a second mixed forest ecotone between the deciduous and southern
pine (primarily loblolly).
Latitude
Tundra
Evergreen
48
Boreal forest
42
38
Mixed
Deciduous forest
forest
Approximate northern
Southern
limit of Floridan 30
pine forest
aquifer system
27
Kilometers
Broadleaved 25
evergreen 0 300 600
forest
Fig. 10.5. Forest formations of eastern North America. Map from Little (1971);
vegetation boundaries from Eyre (1968); aquifer boundary from Sun and Weeks
(1991).
318 Net primary productivity and ecotones
To compare our ecotone postulate with the observations of Fig. 10.5, we return to
Eq. 10.20 rewritten to represent an individual plant as
PT
= f D ( L t ) g(T0 ) Psm m , (gs m2 y1 ). (10.46)
M
The terms of Eq. 10.46 are evaluated as a function of latitude in eastern North
America for spruce, oak and loblolly pine as explained and listed in Table 10.2. The
reference growing season atmospheric temperature at each latitude is taken as the
average daily maximum July temperature which is presented in Fig. 10.6 for locations
in eastern North America from climatic records of the U.S. Weather Service (1974). A
common value of is assumed from Table 10.2, and Eq. 10.46 is plotted vs. latitude
in Fig. 10.7.
Examining Fig. 10.7 we find Eq. 10.46 to predict spruce as being the most productive
of the three species in the observed boreal forest zone and oak as the most productive in
the deciduous forest zone. The interchange of productive order among the three species
takes place clearly within the observed broad mixed forest ecotone. At variance with
the theory is the predicted strong superior productivity of oak in the observed southern
pine forest zone. Clearly some influence is present here in nature that the theory omits.
Arris and Eagleson (1989a) found this same behavior and have suggested nutrient
supply as is indicated in the following discussion paraphrased from their work.
m
Tl = T0 Psm d y1 : h d1 f D ( L t ) NPP/M
( C) g(T0 ) gs m2 h1 (Fig. 3.11) (Table 3.9) 102 gs m2 y1
( N) (Fig. 10.5) (Fig. 8.4b) (Table 8.2) (Table 3.5) (Eq. 10.21) (Eq. 10.22)
Spruce ( L t = 2.91) ( = 0.51) (Table 10.1)
50 17 0.98(1) 0.55 100 : 16 1.37 11.81
45 25 0.88(1) 0.55 85 : 15.5 1.37 8.74
40 29 0.77(1) 0.55 15 : 15 1.37 1.30
32
Weather Service.)
30
28
26
24
22
20
18
16
50 40 30 20
Latitude of observation in Eastern North America
(Calculations summarized in
Eq. 10.46
Shaded areas Table 10.2.)
are observed
ecotones Oak
Loblolly pine
PT
M
Spruce
0
55 50 45 40 35 30 25
, N latitude
Note in Fig. 10.5 how this narrow ecotone separating the deciduous and the southern
pine forests takes a NESW direction along the approximate northern boundary of
the Floridan aquifer system. This marks the beginning of the sandy alluvial formation
terminating in the Florida peninsula. The coarse texture of sandy soils leads to rapid
percolation of soil moisture and to better soil aeration which causes rapid decomposi-
tion of organic matter. The small fraction of clay-sized soil particles coupled with the
low organic content dramatically reduces the surface area available to hold nutrients
in the soil against the leaching effects of percolation (Brady, 1974; Spurr and Barnes,
1980). This is particularly true for deciduous trees which, in contrast to evergreens,
contribute a concentrated pulse of nutrients to the soil in the spring when the autumn
320 Net primary productivity and ecotones
leaf fall decomposes. Evergreen leaves are shed more-or-less continuously and have
both chemical and structural differences which ensure a more gradual return of nutri-
ents to the soil with less likelihood of loss through leaching from the root zone (Monk,
1966; Spurr and Barnes, 1980; Chabot and Hicks, 1982).
Monk (1966) and Chabot and Hicks (1982) suggest that evergreen species may
dominate in areas of low nutrient availability for biological reasons also. Evergreen
leaves tend to contain smaller quantities of key nutrients which reduces the amount of
replacement nutrient required in the soil. Evergreens also have more efficient internal
nutrient cycling which results in a higher carbon gain per unit of nutrient turned over
(Chabot and Hicks, 1982).
As a result of the above, it seems likely that differential nutrient requirements may
make the loblolly pine dominate the oak in the sandy soils characterizing the southern
forest zone.
11
A Summary
Problem simplication
Interaction between the atmosphere and vegetated land surface involves the exchange
of mass, momentum, and energy between the two climate components in processes
which have characteristic time-scales varying from seconds (turbulent wind gusts) to
decades or centuries (life span of individual trees). In seeking first-order understanding
of the natural selection of vegetation form and function we have chosen a priori to
focus our attention at the time-scale of the growing season and to treat the canopy as
a steady-state passive responder to climatic forcing without feedback from canopy to
atmosphere. Additional major assumptions are:
(1) Coefficients of variation of climatic variables in both time and horizontal space
are small enough to allow writing dynamic equations in terms of the climatic
321
322 Summary, speculations, and opportunities
averages over the growing season for the local vegetation community
(Chapter 4).
(2) Atmosphere is neutrally stable in this climatic average (Chapter 4).
(3) Flux divergences of heat and water vapor above the canopy are negligible in
this climatic average (Chapter 5).
(4) The vegetal state in this climatic average is stressless in the interests of plant
health and productivity, and thus the stomates are in an effectively open
condition (Chapter 8).
Bioclimatic optimality
Our principal thesis here has been the existence of a relentless evolutionary pressure
toward a bioclimatically optimal vegetal state assumed to be that of maximum prob-
ability of reproductive success which we equate to maximum seed productivity and
hence, by proportion, to maximum biomass productivity (Chapter 10). From the bio-
physics we have identified the following major optimization opportunities for exercise
of this selective pressure.
Optical optimality
We propose that the optimum foliage arrangement is such as to extract the maximum
photosynthetically useful energy from the radiation beam.
This requires that the reflection of PAR is minimized, the upper leaf surfaces are
fully illuminated, the lowest leaf in the canopy has the minimum radiation intensity
needed for metabolic sufficiency (i.e., the compensation intensity), and that the basal
area of the average leaf be in full shadow. We show (cf. Figs. 3.193.20) that these
conditions result in equality of the canopy extinction coefficients for momentum and
insolation (i.e., = ) and the selection of optimum leaf angle by the solar altitude
(cf. Eq. 3.51 and Fig. 3.15, curve (b)).
We further propose that the climatic insolation, Il , on the average leaf in the crown
is just equal to the minimum insolation, I SL , for fully open stomates to insure stable
systems in which crown carbon assimilation is maximized (cf. Fig. 8.11).
This requires ISL Psm = Il f I ( L t )I0 , where f I ( L t ) averages I0 over the
canopy (cf. Eq. 8.29). It is supported over a wide range of canopies in Figs. 8.10 and
9.5, and defines the crown climatic assimilation potential, Psm = f I ( L t )I0 , deter-
mining the maximum crown CO2 assimilation rate for a given C3 species and climate.
Mechanical optimality
We propose that with optical optimality in homogeneous cylindrical multilayer crowns,
there is a particular leaf angle for each species at which the canopy conductance of
water vapor and CO2 is maximum thereby defining an optimum foliage state.
We believe this to be a new finding.
Summary 323
With stomates fully open, the resistive network for turbulent transfer in the canopy
from stomatal cavities to the free atmosphere is shown adequately to approximate
canopy total resistance, rc , by a function solely of the interleaf layer atmospheric
resistance, ri , which is a function of L t (shown in Appendix H to be a species
constant ). The relative canopy resistance, rc /ra , is thus derived as a function of both
and L t . For homogeneous cylindrical crowns rc /ra has a minimum, and thus the
vertical turbulent fluxes are maximum at a particular leaf angle for each species. For
tapered crowns rc /ra declines monotonically with L t for a given . The dependence
of canopy conductance, kv , on rc /ra and T0 provides selective pressure toward ta-
pered crowns in climates where the fluxes of carbon, nitrogen, or heat may limit
productivity.
Thermal optimality
We propose that during the growing season, the average leaf temperature, Tl , equals the
photosynthetically optimum temperature, Tm , and approximates the average screen-
height atmospheric temperature, T0 , thereby making the canopyatmosphere system
isothermal on the average.
We confirm the finding of earlier investigators by our limited demonstration in
Fig. 9.4 that T0 = Tm . Minor environmental temperature regulation is provided in the
short term by temporary modulations of leaf angle, while major genetic regulation
occurs through both foliage arrangement and crown shape. We are concerned here
only with the genetic mechanisms. Some speculation is in order:
(1) Very limited evidence is presented (cf. Fig. 3.15) that in quite hot climates the
average leaf angle tends to produce upward specular reflection probably as a
means of most efficiently reflecting heat-producing longwave radiation back to
space. Hot climates appear to favor tapered multilayer crowns perhaps because
for these the canopy conductance increases monotonically with increasing leaf
area (up until the onset of significant solid body resistance) and thus maximizes
the flux of sensible heat from plant to atmosphere.
(2) Similarly limited evidence is presented (cf. Fig. 3.15) that in cold climates the
average leaf angle tends to produce horizontal specular reflection which may
be most efficient in retaining heat-producing longwave radiation within the
canopy. Cold climates appear to favor needle leaves in monolayer arrangement
probably because these best inhibit heat transfer to the atmosphere from within
the crown. Crown shape in these climates is typically conical with the
monolayer surface nearly normal to the radiation beam in order to minimize
reflection of NIR.
We believe this to be a new finding.
Among the genetic mechanisms is a tapered multilayer crown which, as we have seen in
Chapter 7, insures a maximum canopy conductance for given and L t , and hence maximum
transfer of latent and sensible heat, CO2 , and dissolved nutrients.
324 Summary, speculations, and opportunities
Hydrological optimality
We propose that at the end of the average growing season, water-limited communities
(i.e., M < 1) achieve a critical moisture state at which the individual plant is at incipient
stress and the crown-average insolation is I SW , the maximum insolation for open
stomates as given by the leaf water characteristic (cf. Fig. 8.15, Eqs. 6.67 and 8.26).
This insures that the stomates remain open for carbon assimilation the maximum
possible time and is borne out by the water balance data presented in Fig. 6.15. Joint
achievement of the optimum foliage state and the critical moisture state defines the
optimum canopy state which compares well with limited observations (cf. Fig. 7.11).
Nutritional optimality
We propose that the optimum foliage state will be that giving maximum canopy con-
ductance, because that will provide maximum water-borne nutrient flux from soil to
plant.
This requires that rc /ra = (
rc /ra )min for cylindrical multilayers as is illustrated in
Fig. 7.3 which, along with the conditions of optical optimality, selects the species and
fixes the leaf area index in this crown shape for given leaf angle (cf. Fig. 7.4). For
given species having the tapered crowns which are favored at high leaf area, the solar-
determined leaf angle along with optical optimality fixes the leaf area index and thus
the resistance ratio as given in Fig. 7.5. Strong support of this proposition is obtained
from the detailed study of three semi-arid canopies summarized in Fig. 7.11.
We further propose that tapered multilayer crowns are indicated at high basal leaf
area in order to insure that through high canopy conductance, carbon supply satisfies
the light-driven carbon demand.
This is borne out by the productivity data presented in Fig. 10.1.
Climax conditions
We propose that the maximum habitat insolation at which stomates can remain open
without water stress, I SW , is always equal to or greater than the minimum insolation
for open stomates, I SL , as given by the leaf light characteristic.
Along with optical optimality this proposition insures that the stomates are always
fully open and that for M < 1, all available moisture is used. As given in Chapter 8,
this defines the water-limited climax condition, Psm = f I ( L t )ISW for M < 1, and
the light-limited climax condition Psm f I ( L t )ISW for M = 1, which receive strong
support for a variety of canopies and habitats in Fig. 9.7.
Productivity
The net primary productivity of a multilayer canopy is derived in Chapter 10 and
normalized by that of a groundcover monolayer of the same species. Derived using
We believe this to be a new finding.
Speculations 325
the optimality conditions cited above, this ratio is called the productivity gain of the
canopy. It is shown to agree quite well with observations from the literature over
a wide range of the basal leaf area, L t , which is shown in Appendix H to be a
species constant. Limited by carbon demand, productivity gain rises with increasing
L t until, for non-tapered crowns, a critical value,
L t , is reached beyond which the
foliage density limits atmospheric carbon supply to the leaves. The zone of productive
limitation by carbon demand appears to contradict implications in the literature that C3
plants are always carbon supply-limited. It is reasoned that L t <
L t , and L t Lt
are necessary conditions for water-limited and light-limited communities respectively.
For tapered crowns we show that carbon supply is non-limiting for all species.
B Speculations
Synthesized selection
With the above mechanisms in mind, how might we synthesize the selection of the
most productive species in a given habitat?
We begin with temperate climates and the cylindrical multilayer crowns that seem
to predominate there:
(1) At the given latitude, , the mean solar altitude, h , determines the mean leaf
angle, cos1 , through optical optimality by Eq. 3.51. This equates the canopy
decay coefficients for light and momentum giving = .
(2) Mechanical optimality gives L t using Eq. 7.14 and bioclimatic optimality fixes
Psm through Eq. 8.41.
(3) The optimum canopy state (kv , M ) for this , L t , and habitat is determined
from the critical moisture state through Eqs. 6.66 and 7.31 as is illustrated in
Figs. 7.10 and 7.11.
(4) The season length, m , is estimated empirically from Fig. 3.11 and Table 3.6.
We believe this to be a new finding.
326 Summary, speculations, and opportunities
The developments of Chapter 10 offer the following possible scenario for accli-
mation by trees to increased concentration of atmospheric CO2 provided the trees are
under CO2 supply limitation, and provided the nitrogen supply is adequate to support
additional foliage:
(1) Increase in ambient CO2 concentration initially brings added productivity since
it removes the CO2 supply constraint.
(2) Over a growing season (or longer), the increased productivity causes an
increase in the crown absorption index, L t , which restores the CO2 supply
limitation (cf. Fig. 10.1) and productivity falls.
C Opportunities
The first-order relationships presented here provide a simple, albeit approximate, ana-
lytical basis for sensitivity analyses and probabilistic forecasting in the manner ex-
ploited by the author in studies of the water balance (cf. Eagleson, 1978g). Accordingly,
the following extensions seem particularly promising:
(1) Both the productivity formulation and its comparison with field observations of
NPP suffer from the expedient assumption that nutrient availability is
non-limiting (cf. Chapter 10). It would be helpful to make the NPP forecasts
probabilistic by incorporating into Eq. 10.18 the natural variability in nutrient
supply as well as that of the important independent climate variables such as
season length, temperature, and available soil moisture. It would then be
possible to evaluate the accuracy of Eq. 10.18 better when compared with
reported observations, to gain a theoretical understanding of observed
interanual variability of primary production (cf. Knapp and Smith, 2001), and
to make sensitivity studies of climate change.
(2) The photosynthetic capacity curve used in generating the productivity function
(Eq. 10.18) as well as the habitat conditions (Eqs. 9.139.14) should be
generalized to incorporate variable ambient CO2 concentration in order to
investigate the effects of this climate change (cf. LaDeau and Clark, 2001).
Certain of the fundamental components of productivity remain empirically defined
and are treated here in an intellectually unsatisfactory manner which might benefit
from a fresh look. Particularly troublesome are the length of the growing season with
its multitude of constraints and influences, and the leaf angle for tapered crowns which
we have concluded to respond to the thermal need to retain or reject longwave radiation.
Finally, far more work is needed in careful testing of our results against observation
before we can claim understanding of the true basis for natural selection in trees.
Drake et al. (1997, Table 1, p. 612) collected 12 long-term studies of 8 field-grown species (crops
such as wheat and cotton, and non-forest ecosystems such as prarie, grassland and wetland) and
found only a 3% increase in leaf area following a doubling of the CO2 .
Appendices
A
329
330 Appendix A
0 h
A
s ho
r
b/2
hs
ro
z
1
b
(a)
0 h
ho
d = 2r
1 hs
b = 2ro
(b)
Fig. A.1. Definition sketches for multilayers. (a) Hemispherical; (b) conical.
spherical segment as
2
s = h o 3r + h 2o , (A.1)
6
in which geometry allows the radius, r , to be written
where
one-sided leaf area
acr = crown one-sided leaf area density = .
crown volume
Effect of crown shape on ow in canopy 331
s( ) = B. (A.9)
As we did for the cone in Eqs. 4.874.90, we can now find the velocity distribution
3
3
2
u( )
= exp L t
B0
, (A.10)
u0 1
3
where again B0 BL t . Equating velocity gradients at = 0 gives B0 = 1. Then,
with m = 1/2 and " = 1 (see Eq. 4.111 and Fig. 4.26b), Eq. 4.37 gives the eddy
332 Appendix A
viscosity variation for the arbitrary hemispherical segment having a single base
3 2 3
exp L t
K m ( ) 3 1
K mo = . (A.11)
K m (0) 6 3 2
1+
3 1
The limiting cases of Eq. A.11 are of particular interest. For = 1 we have the
hemisphere and Eq. A.11 becomes
2
3 3
exp L t
2 2
K mo = , (A.12)
3 2
1 + 3
2
and for 1 we have a palm tree or an umbrella-like savanna tree and Eq. A.11
gives
exp [ L t ( 2 )]
K mo = , (A.13)
1 + 2
both of which are plotted in Fig. 4.24a.
at ( ) = L t (3 2 ) = L t a( ), (A.18)
a( ) = 3 2 . (A.19)
Effect of crown shape on ow in canopy 333
Using Eq. A.19 in Eq. 4.17 gives the areal average momentum flux variation
f ( )
= exp n L t 3 . (A.20)
f (0)
Choosing the interference function again as s( ) = B , the velocity variation is
u( )
= exp L t 3 B0 , (A.21)
u0
in which B0 BL t . Equating the velocity gradients at z = 0 again gives B0 = 1.
From Eq. 4.111 and Fig. 4.24b, " = 1, and with m = 1/2, Eq. 4.37 gives for the cone
exp L t ( 3 )
K mo = , (A.22)
1 + 3 2
which is also plotted in Fig. 4.24a.
K
mo K m ( )
1
K m ( )
d. (A.23)
K m (0) 0 K m (0)
and
limiting hemispherical segment ( 1):
mo = 0.54 + 0.086( L t ) + 0.012( L t )2 ,
K (A.26)
which are displayed in Fig. 4.25a where the plotted points represent the
numerically integrated values.
334 Appendix A
which is also shown in Fig. 4.25a. The values of K mo at the common value
L t = 1.5 are added to Fig. 4.24a as vertical dashed lines. Once again we call
attention to the quite profound qualitative difference in these functions for
differing crown shape.
(a)
(b)
Effect of crown shape on ow in canopy 335
2r t(1)acr t
at ( ) = (2 2 )1/2 acr . (A.28)
ro (1)
2 ro
From the definition of foliage area density, and using Eq. A.28
1
t
at ( ) d L t = acr , (A.29)
0 ro 4
Using Eq. A.31 in Eq. 4.17 gives the momentum flux variation
f ( ) 4
= exp n L t (2y y ) dy ,
2 1/2
(A.32)
f (0) 0
which can be integrated (Gradshteyn and Ryzhik, 1980, p. 82, no. 2.262-1 and p. 81,
no. 2.261) to obtain
f ( ) 2
= exp n L t arc sin(1 ) (1 )(2 )
2 1/2
.
f (0) 2
(A.33)
The interference function will again be taken as in Eq. A.9, which gives the velocity
distribution
u( ) 2
= exp L t arc sin(1 ) (1 )(2 ) B0 ,
2 1/2
u0 2
(A.34)
in which B0 = 2mBL . Equating velocity gradients at the canopy top and setting " = 1
t
(Eq. 4.111 and Fig. 4.26b) establishes that B0 = 2 . Taking m = 1/2 then allows us
to write
2
exp L t (1 ) arc sin(1 ) (1 )(2 )
2 1/2
2
K mo = .
4
1 + (2 2 )1/2
(A.35)
336 Appendix A
at ( ) = 2 L t , (A.38)
a( ) = 2. (A.39)
Using Eq. A.39 in Eq. 4.17 gives the momentum flux variation
f ( )
= exp n L t 2 . (A.40)
f (0)
Once again the interference function is taken as s( ) = B , giving the velocity distri-
bution
u( )
= exp L t 2 B0 , (A.41)
u0
where B0 = mLBt . Equating velocity gradients at = 0, and setting " = 1 gives B0 = 1.
With m = 1/2 the eddy viscosity variation is
exp [ L t ( 2 )]
K mo = , (A.42)
1 + 2
which is identical to that found for the multilayer hemispherical segment with
ro
hh s
= hroo 1. Once again these monolayer eddy viscosities are compared for the
typical value L t = 1.5 in Fig. 4.24b.
the crown shape itself and our flow field becomes the canopy rather than the interior
of the crown. In this case, the interference drag results from the separation between
adjacent crowns at a given elevation rather than from varying separation of neighboring
needles and/or shoots within the crown. This peculiar foliage area distribution is found
primarily in the conifers of cold climates where the high density serves to maintain
sufficient needle warmth for photosynthesis. We will confine our treatment of high
densities to conical monolayers.
Using the interference function, Eq. 4.87, the shear stress is again as in Eq. 4.88,
and we can write the velocity variation
u f ( )
= exp 2m L t 2 + m B . (A.50)
u0
Equating velocity gradients inside and outside the canopy at = 0, and using Eq. 4.111,
one obtains
u h hs
mB = L t . (A.51)
ku 0 h d0
Assuming Fig. 4.26b applies, " = 1, and Eq. A.51 becomes
m B = L t = mn L t . (A.52)
With the single-sided foliage area equal to half the lateral crown surface area in this
limiting case, n = 2, and 2 , reducing Eq. A.52 to
2
B = L t. (A.53)
At the high crown shape Reynolds numbers associated with even moderate wind
speeds, m = 1/2 (Table 7.1). Thus from Eqs. 4.37, A.49, and A.50
2
exp L t +
2
K m ( )
K mo = = , (A.54)
K m (0) 1 +
which is compared with the low density monolayers for the value L t = 1.5 in
Fig. 4.24b. Note for the conical monolayer that the change in drag behavior induced
by the high foliage volume density causes much less vigorous vertical mixing than
was present in the low density case.
Cylindrical monolayer
The monolayer cylindrical crown is just a shallow multilayer cylinder and its K mo has
been given as Eq. A.24.
mo are presented in Fig. 4.25b as a function of L t where
All monolayer values of K
they may be compared with the multilayer values given in Fig. 4.25a.
5 resistance ratio
Homogeneous hemispherical multilayer (M = 1)
Referring to Fig. A.1a, Eqs. A.1 and A.2, and with hroo , the volume of the spherical
segment of height h o is
3
s = h o 6ro h o 2h 2o = ro3 2 , (A.58)
6 3
s
0.7 1.85 . (A.59)
ro3
s 1.85
= 0.7
ro3
s
3 101
ro
1.85
exact
s 3
= 2
r o3 3
102
101 1
340 Appendix A
and
1/1.85 1/1.85 0.54
1 1 0.95
1 1 = = (A.64)
1.05 Lt 1 Lt 1
is the thickness of the uppermost layer. If we now let s (2 ) = 2s , we get
2so
s (2 ) = 1.05so 21.85 = , (A.65)
Lt 1
or
0.54
(2)(0.95)
2 = , (A.66)
Lt 1
and the second layer down has the thickness
0.95 0.54 0.54
2 = 2 1 = [2 10.54 ]. (A.67)
Lt 1
Continuing in this fashion, the bottom layer has thickness
0.95 0.54
L t 1 = L t 1 L t 2 = (L t 1)0.54 (L t 2)0.54 . (A.68)
Lt 1
Using Eqs. 4.38 and 4.40, Eq. 4.165 gives the interleaf layer resistance as
h hs mo ,
ri,n = n / K (A.69)
ku (h d0 )
and using Eq. 4.111, Eq. A.69 is
L t u 0 mo .
ri,n = n / K (A.70)
u 2
Effect of crown shape on ow in canopy 341
As before for the cylindrical crown, the ri,n are arrayed in parallel series to give the
reciprocal of the canopy resistance as
1 1 1 1
= + + + , (A.71)
rc ri,1 ri,1 + ri,2 ri,1 + ri,2 + ri,3 + + ri,L t 1
but in this case the ri,n are unequal due to the varying thickness of the atmospheric
layers containing equal foliage area. Using Eq. A.67, Eq. A.71 becomes
1 mo L t 1 0.54
u 2 K
=
rc L t u 0 0.95
1 1 1
0.54 + 0.54 + + , (A.72)
1 1 + (20.54 10.54 ) (L t 1)0.54
or, with Eq. 4.143
o
ra u 2 K L t 1 0.54 L*t 1
1
= m
. (A.73)
rc ku 0 L t 0.95 n=1
n 0.54
in which K o is given by Eq. A.25. Equation A.74 is plotted in Fig. 7.5 for the common
range of , and we notice that there is an important difference between the resistance
ratio for the hemispherical and cylindrical crowns.
b2
cr ( ) = (h z) 2 = co 3 . (A.79)
12
As we did earlier for the hemisphere, the thickness, 1 , of the topmost atmospheric
layer follows from Eqs. A.76 and A.79 as
1/3
c (1 ) 1/3 1
1 = 1 = = , (A.80)
co Lt 1
while that of the second atmospheric layer is
1/3
2c (1 ) 1/3 c (1 ) 1/3 1
2 = 2 1 = = [21/3 11/3 ],
co co Lt 1
(A.81)
or,
1/3
1
L t 1 = (L t 1)1/3 (L t 2)1/3 . (A.83)
Lt 1
Using Eqs. 4.38 and 4.40, Eq. 4.165 gives the interleaf atmospheric resistance, ri,n ,
as shown for the hemisphere in Eq. A.70. The ri,n are again arrayed in parallel series
to arrive at the reciprocal of the canopy resistance as given by Eq. A.71. The ri,n are
unequal due to the varying thickness of the atmospheric layers containing equal foliage
area. Using Eqs. A.80A.83, Eq. A.71 becomes
1 o (L t 1)1/3
u2 K
= m
rc L t u 0
1 1 1
1/3 + 1/3 + + , (A.84)
1 1 + (21/3 11/3 ) (L t 1)1/3
or, with Eq. 4.143
2 o
ra u K m (L t 1)1/3 L*t 1
1
= . (A.85)
rc ku 0 L t n=1
n 1/3
in which Kmo is given by Eq. A.27. Equation A.86 is plotted in Fig. 7.5 for the common
range of , and we notice there the same important difference between the resistance
ratio for the conical and cylindrical crowns that we previously found between the
hemispherical and cylindrical crowns.
or
1/2
2
2 = , (A.95)
Lt 1
so that the second layer down has the thickness
1/2
1
(2 ) = 2 1 = (21/2 11/2 ). (A.96)
Lt 1
Continuing in this way, the bottom layer has thickness
1/2
1
L t 1 = L t 1 L t 2 = (L t 1)1/2 (L t 2)1/2 . (A.97)
Lt 1
Using Eqs. 4.38 and 4.40, Eq. 4.165 gives the interleaf atmospheric resistance, ri,n ,
as is shown for the hemispherical multilayer in Eq. A.70. The unequal ri,n are again
arrayed in parallel series to arrive at the reciprocal of the canopy resistance as given
by Eq. A.71. Using Eqs. A.93 to A.97, Eq. A.71 becomes
1 o (L t 1)1/2
u2 K
= m
rc L t u 0
1 1 1
1/2 + 1/2 + + , (A.98)
1 1 + (21/2 11/2 ) (L t 1)1/2
or, with Eq. 4.143
2 o
ra u K m (L t 1)1/2 L*t 1
1
= . (A.99)
rc ku 0 L t n=1
n 1/2
in which
5 6
*
L t 1
1
N (L t ) . (A.101)
n=1
n 1/2
In Eq. A.100, the subscript s on the resistance ratio and on the eddy viscosity signifies
they are for the case of sparse monolayer in which the wind blows freely among the
leaves or needles. Evaluating the diffusivity through Eq. A.56, Eq. A.100 is plotted as
the lower dashed line in Fig. 7.7 for the common value, = 0.5. We notice there that
rc
ra s
declines monotonically with L t just as it does for the conical and hemispherical
multilayers (see Fig. 7.5).
Effect of crown shape on ow in canopy 345
where the subscript d on the resistance ratio and on the eddy viscosity signifies they
are for the limiting case of dense monolayer in which the wind blows around the
solid crown. In the absence of observations of the bulk drag parameter for the case of
solid cones, we assume kuu
o
= 1 (see Fig. 4.26b), and use Eq. 4.51 to write Eq. A.105
rc 2 n L t 1
= .
n L t exp (n L t ) K
(A.106)
ra d 3 mo
1 d
1 exp (n L t )
For the solid cone, n = 2, = 2 , and the diffusivity is given by Eq. A.57.
Equation A.106 is plotted as the upper dashed line in Fig. 7.7 for the common value,
= 0.5. We see there that the sparse and dense conical monolayers have opposing
variations with L t . All else being constant, Eq. A.100 must apply for very small L t ,
while Eq. A.106 applies for very large L t . A weighting function to define rc /ra for the
practical, intermediate case is given in Chapter 7 as Eq. 7.24.
B
Net radiation
The net radiation, Rn , is defined as the difference
R n qi qb , (B.3)
qi (1 T )I0 , (B.4)
The material of this Appendix is adapted from the authors unpublished lecture notes for the
Tenth Annual John R. Freeman Memorial Lecture entitled Climate, Soil, and the Water Balance:
A Framework for their Analytical Coupling, MIT, Cambridge, MA, 1977.
346
Potential evaporation from wet simple surfaces 347
Ilw = o TsK
4
, (B.8)
in which E s is the surface emissivity (o = 0.97 for water and o = 1 for other
surfaces), and TsK is the surface temperature in degrees Kelvin.
The net longwave back radiation under clear skies, Rc , is then
Rc = Ilw Ilw . (B.9)
Finally, the effect of cloud cover on the longwave radiation exchange between
atmosphere and surface is approximated by writing the net longwave back
radiation
qb = (1 kl N )Rc , (B.10)
which becomes, using Eqs. B.9, B.8, B.7, and B.6, in cal cm2 min1
qb = (1 kl N ) o TsK
4
(1 l ) 0.971 1010 T0K
4
0.245 , (B.11)
in which
N = fraction of the sky covered by clouds,
kl = reduction factor < 1,
l = longwave albedo of the surface,
o = longwave emissivity of the surface, and
= StefanBoltzmann constant = 0.826 1010 cal cm2 min1 K4 .
Here, we apply Eq. B.1 using time-averaged variables to estimate the annual aver-
age evaporation rate. However, because of the non-linearity of Eq. B.1 with respect to
This development of net longwave back radiation follows that of Snow Hydrology (U.S. Army
Corps of Engineers, 1956, pp. 156161).
348 Appendix B
Introducing the above and remembering that all quantities are now time-averaged,
Eq. B.11 reduces to
qb = (1 0.80 N ) 0.245 0.145 1010 T0K
4
, (B.12)
es
= 6989 51.87 TsK + 0.0966 TsK
2
, (dyne cm2 K1 ) (B.15)
Ts
10
Assuming Ts = T0 , the second factor, F2 = 0.245 0.145 0 4
T0K
es /
, of Eq. B.14 will vary by
less than 25% over the nominal range of T0 . However, the dimensionless third factor,
Potential evaporation from wet simple surfaces 349
280
Observations
240
(Tables of Physical
Properties of Fluids)
200
Fitted approximation
Dynes
cm2 K
160
es
120 = 6989 51.87 T + 0.0966 T 2
T
T
es
80
40
0
270 280 290 300 310
T (K)
F3 1S 1
r
, will vary by 150% over the accompanying range of relative humidity. The
first factor of Eq. B.14 will be randomly variable according to cloud cover, wind speed,
and surface roughness. We thus expect, for given ra and N , that
qb
f (Sr ). (B.17)
Dp
We now explore Eq. B.17 empirically.
Van Bavel (1966, p. 456) has pointed out that
Potential evaporation can be defined for any situation in terms of the appropriate
meteorological variables and the radiative and aerodynamic properties of the
surface. When the surface is wet and imposes no restriction upon the flow of water
vapor, the potential value is reached.
Accordingly, we use the readily available observations of free water surface evapo-
ration (cf. Farnsworth et al., 1982) in developing our estimation procedure for simple
surfaces such as bare soil and leaves.
Jensen and Haise (1963, Table 3, pp. 3031) tabulated the observed mean monthly
values of insolation, I0 , in many different climates as taken from weekly U.S. Weather
Bureau records. The resulting annual averages are plotted in Fig B.2 where the least-
squares linear fit to the observations is given by the dashed line. For comparison, the
annual average clear sky insolation is also shown as calculated by Budyko (1958).
From Jensen and Haise (1963) we select those sites for which the presence of ice
for a significant fraction of the year may be neglected, and we estimate annual water
Free water surface evaporation is that which occurs from a thin film of water having
insignificant heat storage and thus it closely represents the potential evaporation from adequately
watered simple natural surfaces such as individual leaves and bare soil (Farnsworth et al.,
1982, p. 4).
350 Appendix B
0.50
0.40
Io (cal cm2 min1)
0.35 Average of
observations
0.30
0.25
Observations
(Jensen and Haise, 1963)
0.20
0 10 20 30 40 50 60
Latitude (N)
surface evaporation at these sites from the contour map of annual lake evapora-
tion (presented by Kohler et al., 1959, Plate 2). For these water surfaces we assume
an annual average reflectance, T = 0.05 (U.S. Army Corps of Engineers, 1956).
The associated meteorological variables, N , Sr , and To at the chosen sites are averaged
for the year from the values given by the U.S. National Weather Service in their Annual
Climatological Summary using the station closest to the site. Finally, the parameter
o / is given as a function of T0 in Table 5.1.
With these assembled data (listed in Table B.1), Eq. B.1 is solved for Dp using the
observed annual evaporation at each of the nine sites, and the relationship of qb /Dp
The term lake evaporation has been used by Kohler et al. (1959) with the same meaning as
free water surface evaporation (cf. Farnsworth et al., 1982, p. 4).
Table B.1. Estimation of potential evaporation for water surfaces (negligible ice)
E psd E psk
f g
Site I0 a T0 c u2c (observed) qi e qb Rn Dp i (calculated)
o h
location ( N) (ly min1 ) T b ( C) Nc (m s1 ) Sr c (m y1 ) (ly min1 ) (ly min1 ) (ly min1 ) 1+
(ly min1 ) (m y1 ) Reference
1 Phoenix, AZ 33.4 0.371 0.05 21.3 0.20 2.7 0.43 1.82 0.352 0.114 0.238 1.44 0.061 1.92 Jensen and Haise
(1963)
2 Davis, CA 38.4 0.310 0.05 15.7 0.30 3.9 0.52 1.27 0.295 0.109 0.186 1.59 0.045 1.36 Jensen and Haise
(1963)
3 Fresno, CA 36.8 0.317 0.05 16.8 0.25 2.8 0.43 1.52 0.301 0.114 0.187 1.56 0.084 1.48 Jensen and Haise
(1963)
4 Stillwater, OK 36.1 0.286 0.05 15.6 0.40 5.5 0.50 1.45 0.272 0.104 0.168 1.59 0.095 1.26 Jensen and Haise
(1963)
5 Brownsville, TX 25.9 0.307 0.05 23.2 0.40 5.2 0.77 1.47 0.292 0.090 0.202 1.40 0.033 1.42 Jensen and Haise
(1963)
6 Fort Worth, TX 32.8 0.320 0.05 18.6 0.40 4.8 0.63 1.45 0.304 0.095 0.209 1.50 0.039 1.46 Jensen and Haise
(1963)
7 Midland, TX 31.9 0.334 0.05 17.7 0.25 4.5 0.50 1.86 0.317 0.113 0.204 1.53 0.121 1.55 Jensen and Haise
(1963)
8 Spokane, WA 47.6 0.257 0.05 8.5 0.40 3.9 0.42 0.98 0.244 0.105 0.139 1.89 0.072 0.98 Jensen and Haise
(1963)
9 Lake Cochituate, MA 42.5 0.286 0.05 10.7 0.33 4.9 0.70 0.59 0.200 0.111 0.089 1.78 0.031 0.63 Hamon (1961)
a
Jensen and Haise (1963).
b
Eagleson (1970, Fig. 3-10, p. 37).
c
U.S. Weather Service Annual Climatological Summary, 194170 (at nearest station).
d
Kohler et al. (1959, Plate 2).
e f g h i k
qi = (1 T )I0 . Eq. B.6. R n = qi qb . Table 5.1. Eq. B.1. Eqs. B.1 and B.12.
352 Potential evaporation from wet simple surfaces
3.0
qb
6 5
2
Dp
2.0 1
8
3
4
1.0
7
0
0 0.5 1.0
Sr
to Sr is plotted in Fig. B.3. In agreement with our prior reasoning (refer to Eq. B.17),
we see that over the included range of variables their functional relationship may be
approximated empirically by
qb 0.85
= . (B.18)
Dp 1 Sr
Using Eqs. B.3, B.4, B.12, and B.18 in Eq. B.1 gives finally
1 (1 Sr )
E ps = (1 T )I0 qb 1 . (B.19)
1 + o / 0.85
Goodness of t
Because of the secondary importance of Dp in Eq. B.1, the errors inherent in Eq. B.18
(due to our omission of dependencies on wind speed and surface roughness) are reduced
when it is used in Eq. B.19 to estimate E ps . This is shown in Fig. B.4 where E ps as
given by Eqs. B.1 and B.18 is compared with the observed value for each of the nine
water surfaces of Table B.1. Equations B.19 and B.12 provide the estimates of E ps in
this work.
This material is for use in Chapters 6, 7, and 9.
Potential evaporation from wet simple surfaces 353
200
7 1
3 1
150 5
4 6
Observed Epps (cm y1)
100 8
9
0
0 100 150 200
Calculated Epps (cm y1)
Fig. B.4. Free water surface evaporation: goodness of fit. Site numbers refer to
Table B.1.
are assumed to be only weakly variable. In this work we use potential evaporation as a
base for estimating annual transpiration, so we are concerned with climatic variables
which are time averaged over only the growing season.
In Table B.2 we present the data for 18 forested watersheds within the Beaver
Creek basin of Arizona studied by Baker (1982, 1986), Salvucci (1992), and others.
The atmospheric temperatures listed are the average for the growing season, and along
with the other average variables are used with Eq. B.19 to calculate the potential free
water surface evaporation, E ps , during the growing season, .
Cooley (1970) presents the cumulative percentage of annual water surface evapora-
tion as a function of Julian day as observed in Arizona, and Salvucci (1992) fitted this
with an ogive curve which may be approximated well (except for its cold-weather tails)
Estimation of the growing season is considered briefly in Chapter 3.
Table B.2. Estimation of free water surface potential evaporation for Beaver Creek watersheds, Arizona
1 0.51 0.43 18.5 130 284 0.315 0.18 1.51 0.093 0.165 0.054 0.35 65 0.14
2 0.50 0.43 19.3 126 287 0.315 0.18 1.48 0.092 0.166 0.054 0.35 68 0.13
3 0.50 0.43 17.7 125 287 0.315 0.18 1.53 0.093 0.165 0.055 0.34 68 0.13
4 0.55 0.43 16.4 141 278 0.315 0.18 1.57 0.095 0.163 0.050 0.33 59 0.14
5 0.55 0.43 15.8 144 276 0.315 0.18 1.58 0.095 0.163 0.050 0.32 56 0.14
6 0.56 0.43 15.5 144 276 0.315 0.18 1.60 0.095 0.163 0.049 0.32 56 0.14
718 0.58 0.43 14.2 151 271 0.315 0.18 1.64 0.097 0.161 0.048 0.31 65 0.08
a
Baker (1982).
b
NOAA data for Flagstaff, AZ, as reported by Salvucci (1992, Table B.2).
c
Baker (1986).
d
Interpolated from Williams and Anderson (1967).
e
estimated from topographic maps.
f
Fig. B.2, average of observations at = 35 N.
g
Kondratyev (1969, Table 7.5, p. 425): assumed 50% grass @ T = 0.22, and 50% forest @ T = 0.14.
h
Table 5.1. i Eq. B.6. k Eq. B.3. l Eq. B.18. m ( /365) Eq. B.19. n Eq. B.20.
o E ps 1
E psd = 365 % tot
1 .
Potential evaporation from wet simple surfaces 355
* 0, 0 tJ 65,
tJ 65
%E ps = 100 300 65
, 65 tJ 300, (B.20)
100, 300 tJ 365,
in which tJ = Julian day. Using Eq. B.20 and the tabulated season length, we find
E ps as a percentage of the total annual E ps , and from this we calculate the free water
surface potential evaporation, E psd , during the dormant season. Listed in Table B.2,
these values are useful in the water balance calculations of Chapters 6, 7, and 9.
C
1M M 2 kv + (1 M)w/E p
B + , (C.2)
1 + Mkv w/E p 2(1 + Mkv w/E p )2
356
Water balance equations 357
and
1
C (Mkv w/E p )2 . (C.3)
2
In the above:
= reciprocal of mean time, m tb , between storms (s1 ),
[ ] = incomplete gamma function (dimensionless),
[ ] = gamma function (dimensionless),
o = shape parameter of gamma distribution of storm depth (dimensionless),
o = scale parameter of gamma distribution of storm depth (cm1 ),
[1] = saturated soil matrix potential (cm suction),
e = dimensionless exfiltration diffusivity,
d = soil diffusivity index (dimensionless),
h o = surface retention capacity (cm),
m = soil pore size distribution index (dimensionless),
n e = soil effective porosity (dimensionless),
so = effective soil moisture concentration (dimensionless),
E p E ps = bare soil potential evaporation rate (cm s1 ),
kv = unstressed canopy conductance E pv /E ps (dimensionless),
M = vegetated surface fraction (dimensionless),
w = rate of capillary rise (cm s1 ),
E = bare soil evaporation effectiveness (dimensionless), and
K (1) = saturated hydraulic conductivity (cm s1 ).
Defining the bare soil evaporation efficiency, s , in the manner of Eagleson (1978d),
E[E s j ]
s = . (C.4)
1 E ps
Again letting E p E ps and using Eq. C.1, Eq. C.4 becomes
(o , o h o ) h o /E p o [o , (o h o + h o /E p )] B E
s 1+ e
(o ) o h o (o )
(o , o h o ) B E(h o /E p ) w
+ 1 1e 1 + Mkv + E(2B) 1/2
(o ) Ep
w
+ eC E(h o /Ep ) Mkv + E(2C)1/2 + (2E)1/2 eh o /Ep
Ep
3 3 h o /E p o [o , (o h o + h o /E p )]
,CE , BE + 1+
2 2 o h o (o )
3 3 C E w
(2E) 1/2
,CE , BE +e Mkv + E(2C) 1/2
2 2 Ep
w
eB E Mkv + E(2B)1/2 . (C.5)
Ep
358 Appendix C
2 moisture-constrained transpiration
The time to stress is
1 n e so z r wE 1
ts /2 = t + + 1 (B C) + C E(1 M)
Mkv E ps kv E ps M
E 1 1 1
+ 1 (2B)1/2 (2C)1/2 + B E M 1 + ,
kv M Mkv kv
(C.6)
which integrates to
kv E p [o , h o ] ts o [o , ( + /E p )h o ]
E[E vj ] = e 1+
(o ) Ep (o )
(o , h o )
+ 1 e(ts +h o /Ep ) 1 . (C.8)
(o )
in which
1/2
n e K (1)(1)e
Eo 2
. (C.10)
m E ps
Because this asymptote of s has the value s = 1 at E = 2 , we can use Eq. C.9 to
define the upper limit of soil-controlled equilibrium soil moisture as
so = Eo 2/(d+2) . (C.11)
We now use Eqs. 6.52 and C.9 in Eq. 6.54 to get an approximation to the water
balance under soil-controlled conditions:
ho S m K (1) 3/2 (1) mc
1 eG2
3/2
+ + 1+
mh m mh m P mc 1 zw
m tb E ps m K (1) c
= (1 M)E o so1+d/2 + Mkv v + s (C.12)
mh m P o
in which the left-hand side represents moisture supply, SW , to the root zone, and the
right-hand side represents the exhaustion of that moisture by bare soil evaporation,
transpiration, and percolation.
As so 0, G G(0), and (0). The data of Table 6.3 show that for so less
than about so = 0.25, the last (percolation) term on the right-hand side of Eq. C.12
vanishes. Under these dry condition and with a deep, homogeneous soil, we expect
that climatic depression of the water table, z w , will make the capillary rise term vanish
as well. Finally, since m tb m tr in dry climates we have
m P /m h m / m tb + m tr m /m tb , (C.13)
ho S m t E ps
1 eG(0)2 (0) [(1 M)E o so1+d/2 + Mkv v ],
3/2
+ = b
mh m mh mh
(C.14)
in which
and
1/3
5n e 2 K (1)(1)i (d, 0)
(0) . (C.16)
6mo2
Equation C.14 gives a dry climate approximation for the soil moisture state variable:
1 mh G(0)2 (0)3/2 ho S
1+d/2
so = 1e + Mkv v .
(1 M)E o m tb E ps mh m mh
(C.17)
360 Appendix C
Equations. C.14 and C.18 are identical when the bare soil evaporation asymptotes
meet as specified by Eq. C.11. For the catchments of Table 6.3 at least, this will occur
at so 0.75, at and above which the final bracketed term on the right-hand side
of Eq. C.18 (i.e., the percolation) dominates that side of the water balance. Further
assuming negligible capillary rise and so 1, we have G G(1), and M 1. For
m tr /m tb 1, and since G(1) 2G(0), Eq. C.18 yields a wet climate approximation
for the soil moisture state variable:
mh 2G(0) ho S 3/2 (1) mc
so
c
1e + + 1+ kv .
m tb K (1) mh m mh mc 1 zw
(C.19)
while through soil physics, Brooks and Corey (1966) relate these parameters according
to c = 2d 1 (cf. Eq. 6.34).
Equations 6.34 and C.20 are equal at c = 5, and are plotted in Fig. C.1 where
dashed horizontal lines represent the nominal values of c for the classic soil types
as given in Table 6.2. We note from this plot that the water balance equation does
indeed approximate a quadratic in s so1+d/2 for silty loam soils provided the above
constraint on so is met and provided the surface runoff parameters have the fixed values,
G(so ) = G, and (so ) = . With those approximations, the only positive solution of
Eq. C.12 is
B + [B2 4A C ]1/2
s so1+d/2 = , (C.21)
2A
where
K (1)
A , (C.22)
E ps
B (1 M)E o , (C.23)
Water balance equations 361
66
properties for analytical
, 19
8 solution of water balance
rey
Clay equations.
Co
and
oks
7
Bro
ce
an
c = 2d 1
bal
c 6
er
at
w
]
c=2 d+2
]
5 2
Silty loam
ic
at
dr
ua
Q
4
Sandy loam
Sand
3
0 1 2 3 4 5
d
0.6
d
[erf (Z)]
dZ
Z=0
0.4
0.2
0
0 0.4 0.8 1.2 1.6 2.0
z
Z=
2 (Di t)1/2
considered here. We will solve Eq. C.26 under the initial and boundary conditions
s = 1; z = 0, t > 0
. (C.27)
s = s5 ; z > 0, t = 0
The solution is (Crank, 1956, p. 30, Eq. 3.13)
s1 z
= erf erf[Z ], (C.28)
s5 1 2(Dt)1/2
which is plotted in Fig. C.2 using tabulated values from Abramowitz and Stegun (1964,
Table 7.1, p. 310). The asymptote of this function evaluated at Z = 0 is also given in the
above-referenced table and is plotted as the dashed line on Fig. C.2. The intersection of
the asymptotes of erf [Z ] locates the (dimensionless) characteristic penetration depth
of this diffusion as Z = 0.885.
D
A = A0 ekt/ , (D.1)
(t/ ) 1
= . (D.5)
(A/A0 ) k
The tangent to A/A0 at the origin thus intersects the horizontal axis at the average
of the independent variable as we see in the sketch of Fig. D.1. This is the intersection
363
364 Appendix D
1.0
0.8
= ekt/
A
A0
0.6 A
( A )
A 0
A0
( t )
0.4 t
e1
0.2
Asymptote
intersection
0
0 0.4 0.8 1.2 1.6 2.0
t =k t
t
Fig. D.1. Characterization of exponential decay.
of the asymptotes of the exponential function and provides the physical justification
for using this point to characterize such a function.
E
The canopy
In Appendix B we used the energy equation to derive the potential evaporation from
a wet simple surface in Eq. B.19. In terms of average rates over the growing season,
this is
1 (1 Sr )
E ps = (1 T )I0 qb 1 . (E.1)
1 + o / 0.85
and therefore, to the first approximation, the back radiation from the wet surface may
be neglected and Eq. E.1 reduces to
(1 + o /)
I0 = E ps . (E.3)
1 T
Ev
E ps = . (E.4)
kv
365
366 Appendix E
Equations E.3 and E.4 give, for a horizontal unit of vegetated surface,
1 + o / E v
I0 = . (E.5)
1 T kv
in which
Written in terms of the canopy-average insolation, and with g(Tl ) = 1, Psm is given
by Eqs. 8.9 and 9.3 as
where
The functions f I ( L t ) and f D ( L t ) are defined in Eqs. 8.30 and 10.21 respectively.
Finally, using Eqs. E.5 and E.7, Eq. E.6 is written in dimensionless form:
NPP 1 1 + o /
= F = f I ( L t ) f D ( L t ) . (E.8)
m M E v kv 1 T
in which
In Eq. E.8 the product, m M F, is a constant for a given climate and species.
Transpiration as a productivity surrogate 367
An isolated leaf
Again using Eqs. E.5 and E.7, the productivity, Ps , of an isolated horizontal leaf
can be written
(1 + o /)
Ps = g(Tl )Psm = g(Tl ) E v , (E.9)
1 T
where for the single leaf, f I ( L t ) 1, rc /ra 0, and therefore kv 1. This demon-
strates the linear dependence of single-leaf productivity on the transpiration rate in
which the proportionality constant is a function primarily of the leaf temperature.
1.2
5
1.0
2
3
0.8
v
E
M
NPP
0.6
m
ET
=
1
T
E
1
0.4
4
0.2
1
0
0 0.2 0.4 0.6 0.8 1.0 1.2
1 . f (L ) . f (L ) 1 + /
F=
kv*
I t D t (1
)
Fig. E.1. Test of NPP vs. E v linearity. 1, Creosote bush; 2, tropical forest;
3, chestnutoak; 4, loblolly pine; 5, beech.
Table E.1. Annual NPP as a linear function of transpiration rate
NPP
NPP E T
m References for
o c 1+ o
Number Vegetation type (gs m2 y1 ) E T (mm y1 ) dimensionless L t a fI a f Db T0 a
Ta 1T
kv d Ma dimensionless F e NPP and E T
1 Creosote bush 40(7) 126(7) 0.16 0.37 0.84 0.22 21 0.44 0.35 2.22 0.84 0.09 0.08 0.49 Odum (1959)
2 Tropical forest 2188(8) 1230(8) 0.89 2.61 0.35 1.44 32 0.25 0.25 1.67 0.88 1.00 0.50 0.84 Nye (1961)
3 Chestnutoak 1400(9) 832(9) 0.85 2.60 0.36 1.44 31.5 0.26 0.25 1.68 0.88 0.95 0.25 0.99 Whittaker (1966)
4 Loblolly pine 380(10) 525(10) 0.36 2.58 0.36 1.42 27 0.32 0.25 1.76 0.83 0.80 0.11 0.73 Whittaker (1963)
5 Beech 977(11) 5(11) 1.00 3.19 0.30 1.68 27 0.32 0.25 1.76 0.88 0.93 0.28 1.00 Whittaker (1966)
a
Table 9.2.
b
Eq. 10.21 with = (cf. Fig. 10.1).
c
Table 5.1.
d
Fig. 7.9.
e
Eq. E.8.
Transpiration as a productivity surrogate 369
Note that as the climate gets more moist, the annual transpiration from a unit of
horizontal area, m M E v , grows in magnitude approaching the annual evapotranspira-
NPP
tion, E T , in the limit. We plot E T
versus the right-hand side of Eq. E.8 in Fig. E.1
where Eq. E.1 is represented by the 45 limit-line along which E T = m M E v . Since
E T > m M E v always, we expect the observations to lie below the line (which they
do), and to be farther below the line the drier the climate (which they may be). Not
much more can be said about this very limited test of the NPP vs. E v linearity proposed
by Eq. E.8 except to remind the reader of the many approximations involved.
It is interesting to speculate that the degree of agreement of these evapotranspira-
tion data with the proposed linear transpiration relation may be due to compensating
approximations. On the one hand, we neglected (as being small) the back radiation
term in Eq. E.1 which makes E ps smaller than it should be and results in the left-hand
side of Eq. E.8 being too large. This error will be largest in the hot, dry climates. On
the other hand, the left-hand side of Eq. E.8 has been evaluated from observations of
E T rather than Ev , and since E T > E v , makes the left-hand side of Eq. E.8 too small.
This error will also be largest in the hot, dry climates.
Note that the dimensions of the constituent variables of this parameter are adjusted from those
given in Table E.1 such that the resulting plotted parameter is dimensionless. With NPP in
gs m2 y1 ; E T in mm y1 ; = 0.81 gs MJ(tot)1 ; = 585 cal g1
w ; and there being
23.8 104 MJ cal1 this parameter must be divided by 1.99 to be dimensionless.
Further evidence of the NPP vs. E v proportionality is presented in Fig. 8.14a in the form of a
comparison of measured photosynthesis and transpiration during a drying transient for loblolly
pine.
F
Table F.1 is an index of 75 first-order stations of the U.S. Weather Service in the
continental United States containing the station identification number, name, period of
record, percentage of record useable, location, elevation, and mean annual precipitation
(Hawk and Eagleson, 1992, Table 7-5, pp. 277290).
In Table F.2 are tabulated eight parameters of the Poisson rectangular pulses model
of station rainfall averaged over each month of the year from the historical records
of hourly precipitation for each of the 75 stations identified in Table F.1. The column
headings of the table translate into the notations used in this work as follows:
Column Definition
1 Station identification number
2 Latitude of station (northern hemisphere)
3 Longitude of station
4 Month of year (January = 1)
5 Number of independent storms per month (m )
6 Storm duration, hours (m tr )
7 Storm depth, mm (m h )
8 Storm intensity, mm h1 (m i )
9 Time between storms, hours (m tb )
10 Intensityduration correlation ([i, tr ])
11 Shape index of storm depth distribution (o )
12 Scale parameter of storm depth distribution, mm1 (o )
370
Storm rainfall in the United States 371
Percentage
Station
of records Latitude Longitude Elevation PA
State ID Name Years usable ( N) ( W) (m) (m)
Alabama
33 Huntsville 195888 96 34:39 86:46 189 1.41
34 Montgomery 194888 98 32:18 86:24 67 1.29
67 Mobile 195888 94 30:41 88:15 64 1.65
Arizona
1 Flagstaff 195088 99 35:08 111:40 2137 0.56
2 Tucson 194888 98 32:08 110:57 786 0.29
55 Phoenix 194888 95 33:26 112:01 338 0.19
Arkansas
40 Fort Smith 194888 98 35:20 94:22 137 1.03
California
3 San Diego 194888 98 32:44 117:10 3 0.25
4 San Francisco 194888 98 37:37 122:23 3 0.51
75 Los Angeles 194888 98 33:56 118:24 34 0.31
Colorado
5 Denver 194888 98 39:46 104:52 1609 0.39
Connecticut
51 Bridgeport 194888 98 41:10 73:08 3 1.04
Florida
6 Fort Myers 196088 95 26:35 81:52 6 1.34
7 Jacksonville 194888 98 30:30 81:42 9 1.32
68 Key West 195788 96 24:33 81:45 12 1.02
69 Daytona Beach 194288 98 29:44 81:02 6 1.39
Georgia
8 Athens 195788 95 33:57 83:19 244 1.26
9 Atlanta 194888 98 33:39 84:26 308 1.23
56 Savannah 194888 98 32:08 81:12 15 1.25
Idaho
35 Boise 194888 98 43:34 116:13 866 0.30
Illinois
10 Chicago 194888 96 41:44 87:46 189 0.91
11 Peoria 194888 98 40:40 89:41 198 0.90
Indiana
12 Indianapolis 194888 98 39:44 86:16 241 1.01
Iowa
13 Des Moines 194888 98 41:32 93:39 293 0.80
Kansas
42 Dodge City 194888 98 37:46 99:58 786 0.53
Kentucky
52 Paducah 194888 94 37:06 88:36 101 1.13
Louisiana
45 Baton Rouge 194788 97 30:32 91:09 18 1.43
Maine
49 Portland 194888 98 43:39 70:19 3 1.11
Massachusetts
14 Boston 194888 98 42:22 71:02 6 1.09
(continued )
372 Appendix F
Percentage
Station
of records Latitude Longitude Elevation PA
State ID Name Years usable ( N) ( W) (m) (m)
Michigan
66 Sault Ste Marie 194888 98 46:28 84:22 219 0.86
Minnesota
15 Duluth 194888 98 46:50 92:11 436 0.77
16 Internat. Falls 194888 98 48:34 93:23 360 0.62
Mississippi
48 Meridian 194888 98 32:20 88:45 88 1.37
Missouri
44 Springfield 194888 98 37:14 93:23 387 1.05
Montana
36 Billings 194888 98 45:48 108:32 1088 0.37
64 Glasgow 195788 97 48:13 106:37 695 0.27
Nebraska
41 Norfolk 194888 98 41:59 97:26 472 0.63
54 Valentine 194888 98 42:52 100:33 789 0.46
Nevada
17 Ely 194888 98 39:17 114:51 1908 0.24
New Hampshire
18 Mt. Washington 194888 98 43:12 71:30 107 0.94
New Mexico
19 Albuquerque 194788 97 35:03 106:37 1618 0.21
New York
53 Buffalo 194888 98 42:56 78:44 216 0.96
70 Rochester 194888 98 43:07 77:40 168 0.80
North Carolina
20 Raleigh 194888 98 35:52 78:47 116 1.05
71 Cape Hatteras 195788 95 35:16 75:33 3 1.39
North Dakota
37 Bismarck 194888 98 46:46 100:46 503 0.39
65 Williston 194888 98 48:11 103:38 579 0.35
Ohio
31 Cincinnati 194883 92 39:09 84:31 232 1.03
32 Cleveland 194888 98 41:25 81:52 235 0.92
Oklahoma
43 Oklahoma City 194788 97 35:24 97:36 390 0.82
Oregon
46 Salem 194888 98 44:55 123:01 61 1.03
61 Eugene 194888 98 44:07 123:13 110 1.20
62 Astoria 195388 99 46:09 123:53 3 1.73
Pennsylvania
21 Allentown 194888 98 40:39 75:26 119 1.23
22 Philadelphia 194888 100 39:53 75:14 3 1.06
South Carolina
23 Columbia 194888 95 33:57 81:07 64 1.23
73 Charleston 194888 98 32:47 79:56 3 1.17
Storm rainfall in the United States 373
Percentage
Station
of records Latitude Longitude Elevation PA
State ID Name Years usable ( N) ( W) (m) (m)
South Dakota
38 Huron 194888 98 44:23 98:13 390 0.49
Tennessee
24 Knoxville 194888 98 35:48 84:00 290 1.17
Texas
25 Abilene 194088 95 32:26 99:41 544 0.59
59 El Paso 194288 96 31:48 106:24 1195 0.21
60 Corpus Christi 194788 97 27:46 97:30 12 0.75
74 Brownsville 194288 96 25:54 97:26 6 0.67
Utah
57 Salt Lake City 194888 98 40:47 111:57 1286 0.40
58 Milford 194888 97 38:26 113:01 1533 0.23
Vermont
50 Burlington 194888 98 44:28 73:09 101 0.86
Virginia
26 Richmond 194888 98 37:30 77:20 49 1.10
27 Roanoke 194888 97 37:19 79:58 351 1.01
72 Norfolk 195288 96 36:54 76:12 6 1.13
Washington
39 Spokane 194888 98 47:38 117:32 719 0.43
63 Seattle 196488 96 47:27 122:18 137 0.94
West Virginia
28 Charleston 194888 98 38:22 81:36 311 1.07
Wisconsin
29 Greenbay 194888 98 44:29 88:08 207 0.72
30 Lacrosse 194869 95 43:52 91:15 198 0.74
Wyoming
47 Sheridan 194888 98 44:46 106:58 1201 0.37
In deriving the G-function neglecting scattering we follow the work of Ross and Nilson
(1965) (see Ross, 1975, 1981). We refer to Fig. 3.13 for notation.
For the special case rs = rz = z, h = 2 and the angle being averaged is that
between the vertical and the leaf normal giving
If g(L , L ) is the probability density function of foliage area orientation, its integral
over the celestial hemisphere must satisfy
2
1
g(L , L ) d# 1, (G.2)
2 0
393
394 Appendix G
With r in the Sun direction (i.e., r = rs ), the G-function can then be written
2
1 2
G( ) = dL g (L ) | cos r
S rL | dL . (G.5)
2 0 0
cos r
S rL = cos cos L + sin sin L cos( L ). (G.6)
With azimuthal symmetry, the azimuth, , of the solar vector, rs , is arbitrary, so for
convenience we will choose = 0, and Eq. G.6 becomes
cos r
S rL = cos cos L + sin sin L cos(L ). (G.7)
Since for r = rs
h, (G.8)
2
we can write Eq. G.7 in the alternative form
cos r
S rL = sin(h ) cos(L ) + cos(h ) sin(L ) cos(L ). (G.9)
| cos r
S rL | = sin(h ) cos(L )|[1 + cot(h ) tan(L ) cos(L )]|. (G.10)
However, by definition
2
g (L ) cos(L ) dL , (G.12)
0
(2) For L > h , cot(h ) tan(L ) > 1, and the bracketed term of Eq. G.10 will be
negative whenever
The value, o , of the azimuthal angle at which the sign change takes place is
then
(G.18)
Integrating on L gives
G(h ) 2 2
= g (L ) cos(L ) o 1 dL
sin(h ) h
2 2
+ g (L ) cos(L ) cot(h ) tan(L ) sin(o ) dL . (G.19)
h
Replacing o by os , noting that sin(o ) = sin(os ), and using Eq. G.17 we have
G(h ) 2 2
= g (L ) 1 + [tan(os ) os ] cos(L ) dL , L > h . (G.20)
sin(h ) h
Finally, Eqs. G.14 and G.20 can be combined to write the general relationship
G(h ) 2 2
b = + g (L ) [tan(os ) os ] cos(L ) dL , 0 L .
sin(h ) h 2
(G.21)
in which for L h , the second term is identically zero. Remember that Eq. G.21
neglects scattering! We note that for L > h , the second term is always positive and
that it is dominated by the magnitude of [tan(os ) os ] as given from Eq. G.17.
H
With I = Ik , the so-called compensation light intensity, the net rate of photosynthesis
is zero by definition and Eq. H.1 becomes, for an isolated leaf
Ik
Ps
I0
Pr = . (H.2)
Ik k
+
I0 I0
396
Canopy absorption index and compensation ratio 397
From the approximation in Fig. 8.5 we see that k is approximately equal to ISL and
Ps = Psm by definition, so that from Eq. H.2 for an isolated leaf
Ik
Pr I0
= . (H.3)
Psm Ik ISL
+
I0 I0
Expanding to the whole canopy, we first assume that the compensation light intensity
is the same for all leaves, i.e., Ik = Ik . Next, Eqs. 8.28 and 9.02 give I
SL = ISL =
f I ( L t )I0 , making Eq. H.3 for a climax canopy
Ik
Pr I0
= , (H.4)
Psm Ik
+ f I ( L t )
I0
and using Eq. 7.4 gives finally for the climax canopy
Pr 1
= . (H.5)
Psm 1 + f I ( L t )e L t
Both Pr and Psm are species constants which are separately temperature dependent
(cf. Larcher, 1983, pp. 94, 97), but the temperature is fixed here by our optimization
at the species constant, Tm . The ratio Pr /Psm is thus a species constant at optimum
temperature as is the canopy absorption index, L t , through Eq. H.5.
Finally, since Ik /I0 = exp ( L t ) from Eq. 7.4, Ik /I0 is itself a species constant.
2
Ponderosa
Pine
Larch
1
1 2 3 4
I0
In
Ik
398 Appendix H
Observational conrmation
For a few of Bakers (1950) genotypes listed in Table 3.9, observations of either
the canopy absorption index, L t , or both (or ) and L t , were found in the literature
and are given in Tables 2.2 and 2.3. We use these to test the equivalence of L t and,
using Eq. 3.53, Bakers Ik /I0 as is shown in Fig. H.1. Essential agreement over the full
range of Bakers species, from ponderosa pine to maple, lends observational support
to our interpretation of the compensation intensities as a constant percentage of the
ambient insolation, and therefore that L t is a constant for a given species. This result
plays an important role in our work.
Glossary
Adiabat = state space path of constant heat released in the condensation from 1 cm3 of
Advection = horizontal transport of a fluid saturated air at one temperature in contact
quantity by virtue of the fluid motion with 1 cm2 of liquid water at a lower
temperature
Albedo = ratio of light reflected from to that
received by a surface Azimuth = angular position about the
stem
Amphistomatous = broadleaved or
needle-leaved plant species having stomates Bare soil evaporation effectiveness E =
on both leaf surfaces independent climatesoil variable
Anabolic = pertaining to the synthesis, in determining the bare soil evaporation
living systems, of more complex substances efficiency
from simpler ones Bare soil evaporation efficiency s =
Arid climate = one in which the bare soil interstorm bare soil evaporation as a fraction
evaporation is under soil control, i.e., of the potential value (Eq. 6.40)
E 2/ Basal area = projected area on a horizontal
Assimilation = the total process of plant plane
nutrition including the uptake of external Basal leaf area index L t = horizontal
foods and the process of photosynthesis leaf area per unit of crown basal area
Assimilation ratio = grams of organic Beam radiation = radiation unmodified by
matter produced per hour by a gram of the effects of intervening clouds (also called
chlorophyll under the light to which the direct radiation)
system is adapted (Odum, 1975, p. 76) Beam ratio = direct radiance fraction of
Atmospheric emissivity a = efficiency total radiance
of black body radiation from atmosphere Big leaf model = representation of canopy
Atmospheric resistance ra = the time by a single leaf having the equivalent
required for exchange of the sensible heat properties of the full canopy
399
400 Glossary
Biochemical assimilation capacity = the assimilation per unit of crown basal area
linear empirical function, Psm = ISL , assuming unlimited carbon supply to all
relating maximum C3 assimilation rate to leaves
the carbon-critical insolation Carryover storage = the difference, S,
Biomass = weight of living plant material between the stored soil moisture at the two
Bowen ratio rate of sensible heat ends of a partial-year growing season
flux/rate of latent heat flux Catabolic = pertaining to the degradation,
Buffered tolerance = increased tolerance in living organisms, of complex substances
of unfavorable circumstances due to special into their simpler constituents
mechanisms Characteristic penetration depth =
Bulk drag parameter kuu
0
, and equal to the intersection of the asymptotes of the error
crown density (Eq. 4.35) function of the dimensionless diffusion
Canopy absorption index (or just canopy depth (Appendix C-5)
index) horizontal leaf area index = L t Climate = long-term (at least 30-year)
(shown to be equal to L t ) dimensionless average of the local weather
index of exponential decay of either Climate-controlled bare soil
horizontal momentum (using ) or evaporation = evaporation under the
insolation (using ) saturating asymptote of s where E 2/
Canopy conductance kv = transpiration
Climatic assimilation potential = the linear
rate/potential rate of evaporation from a wet
empirical relation, Psm = f I ( L t )I0 ,
simple surface (Eq. 5.29)
relating the maximum crown CO2
Canopy cover M = fraction of substrate assimilation rate to the climate and the
covered by vegetation species
Canopy mass flux capacity Mkv E ps Climatic climax = the globally optimal
Canopy resistance rc = lumped resistance Il = I
bioclimatic state
SL = ISW for M 1
of all the leaves in the crown along with all Climatic insolation I0 = daylight-hour
the labyrinthine interleaf pathways to the average of the insolation during the growing
canopy top season
Canopy-scale non-homogeneity = condition Climax = the asymptotic growth state of a
in which crowns do not fill the canopy space species; a stable stage in the evolution of a
Canopy transpiration efficiency v = plant community supplantable only by
transpiration during the interstorm period as change of the environment by some outside
a fraction of the potential value (Eq. 6.47), agency (Colinvaux, 1973); we use it in
dimensionless reference to the species that while fully
Capillary rise w = rate at which water exploiting the environments productive
rises to a bare soil surface under the forces potential is itself maximally productive, and
of capillarity (Eq. 6.51) hence stable. With this definition, the climax
state is I
SL = I l = I0 .
Carbon-critical insolation ISL =
insolation at which a leaf of a given species Closed canopy = canopy in which the
is photosynthetically saturated when water horizontally projected leaf area equals the
is non-limiting total substrate area thereby giving M = 1
Carbon demand function fD ( L t ) = the Clustered needle varieties = trees with
light-stimulated rate of crown carbon short, stiff needles clustered closely together
Glossary 401
on long twigs forming shoots; high needle Desiccation insolation ISW = insolation
drag interference causes the characteristic at which soil moisture reaches the critical
drag dimension to be that of the cluster value causing stomates to begin closing
Compensation light intensity Ik = when light is non-limiting
insolation at which gross photosynthesis is Desorption = process of soil moisture
just balanced by total respiration, making reduction through capillary rise to the
net photosynthesis equal to zero (Eq. 7.3) surface
Competing species = multiple species all Diffuse radiation = radiation arriving from
meeting the feasibility conditions in the every point of the sky
same environment Diffusely distributed foliage = assumed
Critical light intensity = intensity at which infinite number of small, unconnected
water-equivalent of absorbed solar energy resistive elements behaving as a
flux equals potential transpiration rate continuum-like sink for momentum
Critical moisture state = soil moisture Diffusive penetration depth = depth to
concentration at which stomatal closure which moisture will penetrate a soil by
begins simple diffusion during the average
Crown = that portion of the tree above the interstorm period
lowest leaves Direct solar radiation = parallel beams of
Crown density ( hd 0
hh s
) L t (cf. Eq. 4.35) the Suns rays unmodified by intervening
governs the partition of drag between the clouds (also called beam radiation)
substrate and the foliage elements for Drag area density aD (z) = foliage
closed, sparse canopies (cf. Fig. 4.26) element drag area in a unit vertical
Crown leaf area density = one-sided leaf element of the crown-circumscribing
area/crown volume circular cylinder
Crown-scale non-homogeneity = condition Drying power = the net capacity of the
in which the leaf density is variable within ambient atmosphere to augment
the crown regardless of whether the crowns (or diminish) the radiational forcing of
fill the canopy space evaporation through adiabatic
Cuticular resistance rcu = resistance transformation of excess sensible heat into
impeding diffusion through the leaf cuticle additional latent heat (Eqs. 5.22, B.2,
Fig. B.3)
Dark respiration = the use of oxygen to
break down substances in order to provide Dry matter : radiation quotient =
energy for plant metabolism grams of dry matter produced per megajoule
of intercepted radiant energy
Deep percolation = downward movement of
infiltrated water below the depth at which Dynamically active = contributing
gravity becomes the dominant motive force significant fluid drag
Dense canopy = canopy in which the Ecotone = transition zone between
momentum reaching the substrate is different plant communities (Colinvaux,
negligible and therefore the boundary shear 1973)
stress may be neglected Ecotype = same species raised in different
Dense monolayer = monolayer with foliage habitats and retaining differences when
density high enough that the canopy drag cultivated under common conditions
results primarily from the shape of the crown (Walter and Breckle, 1985, p. 198)
402 Glossary
Gross time = includes the hours of darkness Individual needle varieties = trees with long
needles which are separated to the point of
Growing season = period during which relative drag independence
light, temperature and moisture conditions
are conducive to plant growth Infiltration capacity = infiltration rate as
limited by soil properties and moisture
Habitat = the native environment of a plant content (Eq. 6.16)
species; we use it in reference to the
Infiltration rate = rate at which moisture
environment in which a given species is
enters the soil under the forces of gravity
maximally productive
and capillarity
Habitat heat proposition = Eq. 9.1, Fig. 9.4
Insolation = quantity of radiant solar energy
Habitat light proposition: L = 1 (Eq. 9.3, arriving at a unit of horizontal surface
Figs. 9.5, 9.6) during a given time
Habitat water proposition; water-limited Insolation absorption index L t
branch: W = 1, M 1 (Eq. 9.13,
Insolation extinction coefficient
Fig. 9.7); light-limited branch: W < 1,
M = 1 (Eq. 9.14, Fig 9.7) Intercellular resistance ric = resistance
controlling flow within the stomatal cavity
Heliotropism = the tendency of a plant to
move toward or away from the light Interception = precipitation trapped on
foliage surfaces
Hemispherical distribution = uniform
distribution = probability density function Interception capacity = depth of water
which is constant for all values of the which surface tension and gravity forces are
variable (e.g., distribution of the normal capable of holding on the leaf surface
direction to a unit element of the surface Interference function s( ) = function
area of a hemisphere) describing the drag-modulating effects of
Homogeneity function a(z) at ( )/L t , neighboring foliage elements and/or crowns
equal unity in homogeneous canopy Interleaf layer resistance ri = average
Homogeneous the same at all points resistance of atmospheric path between
adjacent leaf layers
Homogeneous canopy = canopy of
homogeneous cylindrical crowns Interstorm period = time between cessation
with M = 1 of one rainstorm and the beginning of
the next
Homogeneous crown = same foliage
area density everywhere in crown Irradiance = the rate at which incident
radiant energy is received per unit of
Horizontal leaf area index canopy
surface area
absorption index = L t (shown to be
equal to L t and to be a species constant) Isothermal = constant temperature
(Eq. 7.4) Isotropic = insensitive to direction at a point
404 Glossary
Abramowitz, M., and I.A. Stegun (editors), Swedish Natural Science Research Council
Handbook of Mathematical Functions with (NFR), Stockholm, 1980b.
Formulas, Graphs, and Mathematical
Allen, L.H., Turbulence and wind spectra
Tables, National Bureau of Standards
within a Japanese Larch plantation, J. Appl.
Applied Mathematics Series 55, 9th
Meteorol., 7, 7378, 1968.
Printing, U.S. Government Printing Office,
Washington, D.C., 1964. Allen, L.H., Jr., and E.R. Lemon, Carbon
dioxide exchange and turbulence in a Costa
Addicott, F.T., and J.L. Lyon, Physiological
Rican tropical rain forest, in Vegetation and
ecology of abscission, in Shedding of Plant
the Atmosphere, vol. 2, Case Studies, edited
Parts, edited by T.T. Koslowski, pp. 85124,
by J.L. Monteith, pp. 265308, Academic
Academic Press, New York, 1973.
Press, New York, 1976.
Ahlgren, C.E., Phenological observations of
nineteen native tree species in Northeastern Allen, L.H., C.S. Yocum, and E.R. Lemon,
Minnesota, Ecology, 38, 622628, 1957. Photosynthesis under field conditions, VII.
Radiant energy exchanges within a corn
Albrektson, A., Biomass of Scots pine
crop canopy and implications in water
(Pinus sylvestris L.). Amount, Development,
use efficiency, Agron. J., 56(2), 253259,
Methods of Mensuration, Report No. 2,
1964.
University of Agricultural Sciences,
Department of Silviculture, Umea, Sweden, Amiro, B.D., Comparison of turbulence
1980a. statistics within three boreal forest canopies,
Albrektson, A., Relations between tree Boundary-Layer Meteorol., 51, 99121,
biomass fractions and conventional 1990a.
silvicultural measurements, in Structure and Amiro, B.D., Drag coefficients and
Function of Northern Coniferous Forests: turbulence spectra within three boreal forest
An Ecosystem Study, edited by T. Persson, canopies, Boundary-Layer Meteorol., 52,
pp. 315327, Ecological Bulletin 32, 227246, 1990b.
409
410 References
Anderson, M.C., Stand structure and light Range Experiment Station, Fort Collins,
penetration, II. A theoretical analysis, CO, 1982.
J. Appl. Ecol., 3, 4154, 1966. Baker, M.B., Jr., Effects of ponderosa pine
Aronsson, A., and S. Elowson, Effects of treatments on water yield in Arizona, Water
irrigation and fertilization of mineral Resources Res., 22(1), 6773, 1986.
nutrients in Scots pine needles, in Baldwin, H.I., The period of height growth
Structure and Function of Northern in some northeastern conifers, Ecology,
Coniferous Forests: An Ecosystem Study, 12(4), 665689, 1931.
edited by T. Persson, pp. 219228, Bartholomew, W.V., J. Meyer, and
Ecological Bulletin 32, Swedish Natural H. Laudelout, Mineral nutrient
Science Research Council (NFR), immobilization under forest and grass
Stockholm, 1980. fallow in the Yangambi region (Belgian
Arris, L.L., and P.S. Eagleson, A Congo), Pub. Inst. Etude Agron. Congo
Physiological Explanation for Vegetation Belge, 57, 1953.
Ecotones in Eastern North America, Ralph Billings, W.D., The historical development
M. Parsons Laboratory Report 323, of physiological plant ecology, in
Department of Civil Engineering, Physiological Ecology of North American
Massachusetts Institute of Technology, Plant Communities, edited by B.F. Chabot
Cambridge, MA, 1989a. and H.A. Mooney, pp. 115, Chapman and
Arris, L.L., and P.S. Eagleson, Evidence of a Hall, New York, 1985.
physiological basis for the borealdeciduous Birkebak, R., and R. Birkebak, Solar
forest ecotone in North America, Vegetatio, radiation characteristics of tree leaves,
82, 5558, 1989b. Ecology, 45(3), 646649, 1964.
Arris, L.L., and P.S. Eagleson, A water use Blum, B.M., Picea rubens Sarg., in Silvics
model for locating the boreal/deciduous of North America, vol. 1, Conifers,
forest ecotone in Eastern North America, coordinated by R.M. Burns and
Water Resources Res., 30(1), 19, 1994. B.H. Honkala, pp. 250259, Agriculture
Aylor, D.E., Y. Wang, and D.R. Miller, Handbook 654, U.S. Department of
Intermittent wind close to the ground within Agriculture, Washington, D.C., 1990.
a grass canopy, Boundary-Layer Meteorol., Bormann, F.H., and G.E. Likens, Pattern
66, 427448, 1993. and Process in a Forested Ecosystem,
Baker, F.S., Principles of Silviculture, Springer-Verlag, New York, 1979.
McGraw-Hill, New York, 1950. Boysen-Jensen, P., Die Stoffproduktion der
Pflanzen, Gustav Fischer Verlag, Jena,
Baker, J.B., and O.G. Langdon, Pinus taeda
Germany, 1932.
L., in Silvics of North America, vol. 1,
Conifers, coordinated by R.M. Burns and Brady, N.C., The Nature and Properties of
B.H. Honkala, pp. 497512, Agriculture Soils, 8th Edition, Macmillan, New York,
Handbook 654, U.S. Department of 1974.
Agriculture, Washington, D.C., 1990. Bras, R.L., Hydrology: An Introduction to
Baker, M.B., Jr., Hydrologic Regimes of Hydrologic Science, Addison-Wesley,
Forested Areas in the Beaver Creek Reading, MA, 1990.
Watershed, General Technical Report Bray, J.R., Primary consumption in three
RM-90, U.S. Department of Agriculture forest canopies, Ecology, 45, 165167,
Forest Service, Rocky Mountain Forest and 1964.
References 411
Briggs, L.J., Limiting negative pressure of Bunce, J.A., L.N. Miller, and B.F. Chabot,
water, J. Appl. Phys., 21, 721722, 1950. Competitive exploitation of soil water by
Briscoe, P.V., Stomata and the plant five Eastern North American tree species,
environment, PhD thesis, University of Bot. Gaz., 138, 168173, 1977.
Nottingham, U.K., 1969. Burdine, N.T., Relative permeability
Britton, N.L., When the leaves appear, Bull. calculations from pore size distribution data,
Torrey Bot. Club, 6, 235237, 1878a. Trans. AIME, 198, 7178, 1958.
Britton, N.L., When the leaves fall, Bull. Burns, R.M., and B.H. Honkala (technical
Torrey Bot. Club, 6, 211213, 1878b. coordinators), Silvics of North America,
vol. 1, Conifers, Agriculture Handbook 654,
Brix, H., The effect of water stress on the
U.S. Department of Agriculture,
rates of photosynthesis and respiration in
Washington, D.C., 1990a.
tomato plants and loblolly pine seedlings,
Physiol. Plant., 15, 1020, 1962. Burns, R.M., and B.H. Honkala (technical
Brooks, R.H., and A.T. Corey, Properties of coordinators), Silvics of North America,
porous media affecting fluid flow, J. Irrig. vol. 2, Hardwoods, Agriculture Handbook
Drain. Div., Am. Soc. Civil Eng., IR2, 654, U.S. Department of Agriculture,
6188, 1966. Washington, D.C., 1990b.
Brown, H.P., Growth studies in forest trees, Businger, J.A., J.C. Wyngaard, Y. Izumi,
I. Pinus rigida Mill, Bot. Gaz., 54, 386402, and E.F. Bradley, Fluxprofile relationships
1912. in the atmospheric surface layer, J. Atmos.
Sci., 28, 181189, 1971.
Brown, H.P., Growth studies in forest trees,
II. Pinus strobus L, Bot. Gaz., 59, 197241, Cannell, M.G.R., World Forest Biomass and
1915. Primary Production Data, Academic Press,
London, 1982.
Brown, K.W., and W. Covey, The
energy-budget evaluation of the Cannell, M.G.R., R. Milne, L.J. Sheppard,
micro-meteorological transfer processes and M.H. Unsworth, Radiation interception
within a cornfield, Agric. Meteorol., 3, and productivity of willow, J. Appl. Ecol.,
7396, 1966. 24, 261278, 1987.
Brutsaert, W.H., Evaporation into the Carbon, B.A., G.A. Bartle, and
Atmosphere, Reidel, Dordrecht, The A.M. Murray, Leaf area index of some
Netherlands, 1982. eucalypt forests in south-west Australia,
Budyko, M.I., The Heat Balance of the Austral. Forest Res., 9, 323326, 1979.
Earths Surface, Leningrad, 1956 (in Carbon, B.A., G.A. Bartle, and
Russian). Translated by N.A. Stepanova, A.M. Murray, Water stress, transpiration and
U.S. Weather Bureau Publication 131692, leaf area index in eucalypt plantations in a
Department of Commerce, Washington, bauxite mining area in south-west Australia,
D.C., 1958. Austral. J. Ecol., 6, 459466, 1981.
Budyko, M.I., Climate and Life, 1971 Carslaw, H.S., and J.C. Jaeger, Conduction
(in Russian). English translation edited by of Heat in Solids, 2nd Edition, Oxford
D.H. Miller, Academic Press, New York, University Press, New York, 1959.
1974. Cauchy, A., Memoire sur la rectification des
Budyko, M.I. (editor), Atlas of World Water courbes et la quadrature des surfaces
Balance, UNESCO, Paris, 1977. courbes, presente le 22 octobre, 1832,
412 References
Oevres compl`etes dA. Cauchy, Ire serie, Culf, A.D., Equilibrium evaporation
vol. 2, pp. 167177, Gauthier-Villars, Paris, beneath a growing convective boundary
1908. layer, Boundary-Layer Meteorol., 70,
Chabot, B.F., and D.J. Hicks, The ecology 3749, 1994.
of leaf life spans, Annu. Rev. Ecol. Syst., 13, Daily, J.W., and D.R.F. Harleman, Fluid
229259, 1982. Dynamics, Addison-Wesley, Reading, MA,
Chen, J.M., and T.A. Black, Defining leaf 1966.
area index for non-flat leaves, Plant, Cell Darwin, C., On the Origin of Species by
Envir., 15, 421429, 1992. Means of Natural Selection, John Murray,
Christensen, N.L., Vegetation of the London, 1859.
southeastern coastal plain, in North Davies, J.A., and C.D. Allen, Equilibrium,
American Terrestrial Vegetation, edited by potential and actual evaporation from
M.G. Barbour and W.D. Billings, cropped surfaces in Southern Ontario,
pp. 317363, Cambridge University Press, J. Appl. Meteorol., 12, 649657, 1973.
New York, 1988. Dawkins, R., The Selfish Gene, Oxford
Cionco, R.M., A windprofile index for University Press, New York, 1976.
canopy flow, Boundary-Layer Meteorol., Deans, J.D., Fluctuations of the soil
3, 255263, 1972. environment and fine root growth in a young
Cionco, R.M., Analysis of canopy index Sitka spruce plantation, Plant Soil, 52,
values for various canopy densities, 195208, 1979.
Boundary-Layer Meteorol., 15, 8193, Deans, J.D., Dynamics of coarse root
1978. production in a young plantation of Picea
Clary, W.P., M.B. Baker, Jr., P.F. OConnell, sitchensis, Forestry, 54, 139155, 1981.
T.N. Johnsen, Jr., and R.E. Campbell, de Bruin, H.A.R., A model for the
Effects of PinyonJuniper Removal on PriestleyTaylor parameter , J. Climate
Natural Resource Products and Uses in Appl. Meteorol., 22, 572578, 1983.
Arizona, 28 pp., Research Paper RM-128, Decker, J.P., Effect of temperature on
U.S. Department of Agriculture Forest photosynthesis and respiration in red and
Service, Rocky Mountain Forest and loblolly pines, Plant Physiol., 19, 679688,
Range Experiment Station, Fort Collins, 1944.
CO, 1974.
Denmead, O.T., Relative significance of soil
Colinvaux, P., Introduction to Ecology, John and plant evaporation in estimating
Wiley, New York, 1973. evapotranspiration, in Plant Response to
Cooley, K.R., Evaporation from Open Water Climatic Factors, edited by R.O. Slatyer,
Structures in Arizona, Agricultural pp. 505511, UNESCO, Paris, 1973.
Experiment Station and Cooperative Denmead, O.T., Temperate cereals, in
Extension Service Folder 159, University of Vegetation and the Atmosphere, vol. 2, Case
Arizona, Tucson, 1970. Studies, edited by J.L. Monteith, pp. 131,
Cowan, I.R., The interception and Academic Press, New York, 1976.
absorption of radiation in plant stands, Denmead, O.T., and E.F. Bradley,
J. Appl. Ecol., 5, 367379, 1968. Fluxgradient relationships in a forest
Crank, J., The Mathematics of Diffusion, canopy, in The ForestAtmosphere
Oxford University Press, New York, 1956. Interaction, edited by B.A. Hutchison and
References 413
B.B. Hicks, pp. 421422, Reidel, Dordrecht, permeable soils, Water Resources Res., 6(2),
The Netherlands, 1985. 478490, 1970.
Denmead, O.T., and E.F. Bradley, On scalar Eagleson, P.S., Dynamic Hydrology,
transport in plant canopies, Irrig. Sci., 8, McGraw-Hill, New York, 1970.
131149, 1987.
Eagleson, P.S., Climate, soil, and vegetation,
Denmead, O.T., and I.C. McIlroy, 1. Introduction to water balance dynamics,
Measurements of non-potential evaporation Water Resources Res., 14(5), 705712,
from wheat, Agric. Meteorol., 7, 285302, 1978a.
1970.
Eagleson, P.S., Climate, soil, and vegetation,
Denmead, O.T., F.X. Dunin, S.C. Wong, and 2. The distribution of annual precipitation
E.A.N. Greenwood, Measuring water use derived from observed storm sequences,
efficiency of eucalypt trees with chambers Water Resources Res., 14(5), 713721,
and micrometeorological techniques, 1978b.
J. Hydrol., 150, 649664, 1993.
Eagleson, P.S., Climate, soil and vegetation,
Dickinson, R.E., A. Henderson-Sellers,
3. A simplified model of soil moisture
C. Rosenzweig, and P.J. Sellers,
movement in the liquid phase, Water
Evapotranspiration models with canopy
Resources Res., 14(5), 722730, 1978c.
resistance for use in climate models: A
review, Agric. Forest Meteorol., 54, Eagleson, P.S., Climate, soil and vegetation,
373388, 1991. 4. The expected value of annual
evapotranspiration, Water Resources Res.,
Doley, D., Photosynthetic productivity of
14(5), 731740, 1978d.
forest canopies in relation to solar radiation
and nitrogen cycling, Austral. Forest Res., Eagleson, P.S., Climate, soil, and vegetation,
12, 245261, 1982. 5. A derived distribution of storm surface
Doorenbos, J., and W.O. Pruitt, Crop Water runoff, Water Resources Res., 14(5),
Requirements, FAO Irrigation and Drainage 741748, 1978e.
Paper 24, (revised 1977), United Nations, Eagleson, P.S., Climate, soil and vegetation,
Rome, 1977. 6. Dynamics of the annual water balance,
Drake, B.G., M.A. Gonz`alez-Meler, and Water Resources Res., 14(5), 749764,
S.P. Long, More efficient plants: A 1978f.
consequence of rising atmospheric CO2 ? Eagleson, P.S., Climate, soil, and vegetation,
Annu. Rev. Plant Physiol. Plant Mol. Biol., 7. A derived distribution of annual water
48, 609639, 1997. yield, Water Resources Res., 14(5),
Drozdov, A.V., The productivity of zonal 765776, 1978g.
terrestrial plant communities and the Eagleson, P.S., Ecological optimality in
moisture and heat parameters of an area, water-limited natural soilvegetation
Soviet Geography, 12, 5460, 1971. systems, 1. Theory and hypothesis,
Druilhet, A., A. Perrier, J. Fontan, and Water Resources Res., 18(2), 325340,
J.L. Laurent, Analysis of turbulent transfers 1982.
in vegetation: Use of thoron for measuring Eagleson, P.S., and T.E. Tellers, Ecological
the diffusivity profiles, Boundary-Layer optimality in water-limited natural
Meteorol., 2, 173187, 1972. soilvegetation systems, 2. Tests and
Dunne, T., and R.D. Black, An experimental applications, Water Resources Res., 18(2),
investigation of runoff prediction in 341354, 1982.
414 References
Ehleringer, J., Annuals and perennials of Feynman, R.P., The Feynman Lectures in
warm deserts, in Physiological Ecology of Physics, vol. 1, Addison-Wesley, Reading,
North American Plant Communities, edited MA, 1963.
by B.F. Chabot and H.A. Mooney, pp. Finnigan, J.J., Turbulent transport in flexible
162180, Chapman and Hall, New York, plant canopies, in The ForestAtmosphere
1985. Interaction, edited by B.A. Hutchison and
Eshbach, O.W. (editor), Handbook of B.B. Hicks, pp. 443480, Reidel, Dordrecht,
Engineering Fundamentals, John Wiley, The Netherlands, 1985.
New York, 1936. Finnigan, J.J., and M.R. Raupach, Transfer
Evans, G.C., and D.E. Coombe, processes in plant canopies in relation to
Hemispherical and woodland canopy stomatal characteristics, in Stomatal
photography and the light climate, J. Ecol., Function, edited by E. Zeiger,
47, 103113, 1959. G.D. Farquhar, and I.R. Cowen,
Eyre, S.R., Vegetation and Soils: A World pp. 385429, Stanford University Press,
Picture, 2nd Edition, Edward Arnold, Stanford, CA, 1987.
London, 1968. Fogg, G.E. (editor), The State and
Fakorede, M.A.B., and J.J. Mock, Leaf Movement of Water in Living Organisms,
orientation and efficient utilization of solar Part II, Water in the Plant, Society for
energy by maize (Zea mays L.), in Experimental Biology, London, 1965.
Symposium on the Agrometeorology of Forrest, W.G., Biological and economic
the Maize Crop, Iowa State University, production in radiata pine plantations,
1976, pp. 207230, WMO, Geneva, J. Appl. Ecol., 10, 259267, 1973.
1977.
Forrest, W.G., and J.D. Ovington, Organic
Farnsworth, R.K., E.S. Thompson, and matter changes in an age series of Pinus
E.L. Peck, Evaporation Atlas for the radiata plantations, J. Appl. Ecol., 7,
Contiguous 48 United States, 26 pp. with 177186, 1970.
maps, NOAA Technical Report NWS 33,
Fraser, D.A., Vegetative and reproductive
Office of Hydrology, National Weather
growth of black spruce (Picea mariana
Service, Washington, D.C., 1982.
[Mill.] BSP.) at Chalk River, Ontario,
Fechner, G.H., Picea pungens Engelm., in Canada, Can. J. Bot., 44, 567580,
Silvics of North America, vol. 1, Conifers, 1966.
coordinated by R.M. Burns and
Fritschen, L.J., J. Hsia, and P. Doraiswamy,
B.H. Honkala, pp. 238249, Agriculture
Evapotranspiration of a Douglas fir with a
Handbook 654, U.S. Department of
weighing lysimeter, Water Resources Res.,
Agriculture, Washington, D.C., 1990.
13(1), 1977.
Federer, C.A., and D. Lash, BROOK: A
Hydrologic Simulation Model for Eastern Fritz, S., Solar radiation during cloudless
Forests, Water Resources Research Center days, Heat. Vent., 46, 6974, 1949.
Report 19, University of New Hampshire, Frohlich, C., and R.W. Brusa, Solar
Durham, NH, 1978. radiation and its variation in time, Sol.
Feller, M.C., Biomass and nutrient Phys., 74, 209215, 1981.
distribution in two eucalypt forest Gaastra, P., Climatic control of
ecosystems, Austral. J. Ecol., 5, 309333, photosynthesis and respiration, in
1980. Environmental Control of Plant Growth,
References 415
Hawk, K.L., and P.S. Eagleson, Climatology Holmgren, P., P.G. Jarvis, and M.S. Jarvis,
of Station Storm Rainfall in the Continental Resistances to carbon dioxide and water
United States: Parameters of the vapor transfer in leaves of different plant
BartlettLewis and Poisson Rectangular species, Physiol. Plant., 18, 557573, 1965.
Pulses Models, with diskette, M.I.T. Horn, H.S., The Adaptive Geometry of
Department of Civil Engineering, R.M. Trees, Princeton University Press, Princeton,
Parsons Laboratory Report 336, Cambridge, NJ, 1971.
MA, 1992.
Horton, R.E., The role of infiltration in the
Heath, M.C., and P.D. Hebblethwaite, Solar
hydrologic cycle, Trans. Am. Geophys. U.,
radiation intercepted by leafless,
14, 446460, 1933.
semi-leafless and leafed peas (Pisum
sativum) under contrasting field conditions, Hughes, G., J.D.H. Keatinge, and S.P. Scott,
Ann. Appl. Biol., 107, 309318, 1985. Pigeon peas as a dry season crop in
Trinidad, West Indies, II. Interception and
Hendricks, D.W., and V.E. Hansen,
utilization of solar radiation, Trop. Agric.,
Mechanics of evapo-transpiration, J. Irrig.
Trinidad, 58, 191199, 1981.
Drain. Div., Am. Soc. Civil Eng., 88, IR-2,
6782, 1962. Illick, J.S., Pennsylvania Trees, 4th Edition,
Pennsylvania Commission of Forestry,
Hicks, B.B., and C.M. Sheih, Some
Harrisburg, PA, 1923.
observations of eddy momentum fluxes
within a maize canopy, Boundary-Layer Inoue, E., On the turbulent structure of
Meteorol., 11, 515519, 1977. airflow within crop canopies, J. Meteorol.
Hicks, D.J., and B.F. Chabot, Deciduous Soc. Japan, 41, 317326, 1963.
forest, in Physiological Ecology of North Inoue, K., and Z. Uchijima, Experimental
American Plant Communities, edited by study of microstructure of wind turbulence
B.F. Chabot and H.A. Mooney, pp. 257277, in rice and maize canopies, Bull. Nat. Inst.
Chapman and Hall, New York, 1985. Agric. Sci. Tokyo, Japan, Ser. A, 26, 188,
Hinckley, T.M., M.O. Schroeder, 1979.
J.E. Roberts, and D.N. Bruckeshoff, Effect Iqbal, M., An Introduction to Solar
of several environmental variables and Radiation, Academic Press, New York,
xylem pressure potential on leaf surface 1983.
resistance in white oak, For. Sci., 21, Jacquemin, B., and J. Noilhan, Sensitivity
201211, 1975. study and validation of a land surface
Hinckley, T.M., R.G. Aslin, R.R. Aubuchon, parameterization using the
C.L. Metcalf, and J.E. Roberts, Leaf HAPEX-MOBILHY data set,
conductance and photosynthesis in four Boundary-Layer Meteorol., 52, 92134,
species of the oak hickory forest type, For. 1990.
Sci., 24, 7384, 1978.
Jarvis, P.G., and J.W. Leverenz, Productivity
Hoerner, S.F., Fluid-Dynamic Drag, of temperate, deciduous and evergreen
Hoerner Fluid Dynamics, Brick Lawn, NJ, forests, in Encyclopedia of Plant
1958. Physiology, New Series, vol. 12D,
Holbrook, N.M., and M.A. Zwieniecki, Physiological Plant Ecology IV: Ecosystem
Embolism repair and xylem tension: Do we Processes, edited by O.L. Lange, P.S. Nobel,
need a miracle?, Plant Physiol., 120, 710, C.B. Osmond, and H. Ziegler, pp. 233280,
1999. Springer-Verlag, New York, 1983.
References 417
Jarvis, P.G., and K.G. McNaughton, Kim, C.P., and D. Entekhabi, Examination
Stomatal control of transpiration: Scaling up of two methods for estimating regional
from leaf to region, Adv. Ecol. Res., 15, evaporation using a coupled mixed layer and
149, 1986. land surface model, Water Resources Res.,
Jarvis, P.G., G.B. James, and J.J. Landsberg, 33(9), 21092116, 1997.
Coniferous forest, in Vegetation and the Kimball, H.H., and I.F. Hand,
Atmosphere, vol. 2, Case Studies, edited by Sky-brightness and daylight illumination
J.L. Monteith, pp. 171240, Academic measurements, Monthly Weather Rev., 49,
Press, New York, 1976. 481487, 1921.
Jasinski, M.F., and P.S. Eagleson, Estimation Klein, W.H., Calculation of solar radiation
of subpixel vegetation cover using and the solar heat load on man, J. Meteorol.,
redinfrared scattergrams, IEEE Trans. 5, 119129, 1948.
Geosci. Remote Sens., 28(2), 253267,
Knapp, A.K., and M.D. Smith, Variation
1990.
among biomes in temporal dynamics of
Jensen, M.E., and H.R. Haise, Estimating aboveground primary production, Science,
evapotranspiration from solar radiation, 291, 481484, 2001.
J. Irrig. Drain. Div., Am. Soc. Civil Eng.,
IR-4, 1541, 1963. Kohler, M.A., T.J. Nordenson, and D.R.
Baker, Evaporation Maps for the United
Jensen, P., and S. Pettersson, Nutrient
States, Technical Paper no. 37, U.S. Weather
uptake in roots of Scots pine, in Structure
Bureau, Washington, D.C., 1959.
and Function of Northern Coniferous
Forests: An Ecosystem Study, edited by Kondo, J., Relation between the roughness
T. Persson, pp. 229237, Ecological coefficient and other aerodynamic
Bulletin 32, Swedish Natural Science parameters, J. Meteorol. Soc. Japan, 49,
Research Council (NFR), Stockholm, 1980. 121124, 1971.
Jones, L.A., and H.R. Condit, Sunlight and Kondo, J., On a product of mixing length
skylight as determinants of photographic and coefficient of momentum absorption
exposure, J. Optic. Soc. Am., 38, 123178, within plant canopies, J. Meteorol. Soc.
1948. Japan, 50, 487488, 1972.
Jordan, C.F., Derivation of leaf-area index Kondo, J., and S. Akashi, Numerical studies
from quality of light on the forest floor, on the two-dimensional flow in horizontally
Ecology, 50(4), 663666, 1969. homogeneous canopy layers,
Kaimal, J.C., and J.J. Finnigan, Atmospheric Boundary-Layer Meteorol., 10, 255272,
Boundary Layer Flows, Oxford University 1976.
Press, New York, 1994. Kondratyev, K. Ya., Radiation in the
Karman, T. von, Mechanische Ahnlichkeit Atmosphere, Academic Press, New York,
und Turbulenz, Nachr. Ges. Wiss. Gottingen, 1969.
Math.-Phys. Kl., 5876, 1930. Kozlowski, T.T. (editor), Water Deficits and
Karman, T. von, Turbulence and skin Plant Growth, vol. 2, Plant Water
friction, J. Aeronaut. Sci., 1, 120, 1934. Consumption and Response, Academic
Kestemont, P., Production primaire de la Press, New York, 1968.
strate aboree dune hetraie a` fetuques, Bull. Kozlowski, T.T., and F.X. Schumacher,
Soc. Roy. Bot. Belgique, 106, 305316, Estimation of stomated foliar surface of
1973. pines, Plant Physiol., 18, 122127, 1943.
418 References
Kramer, P.J., Plant and Soil Water the suns beam, Agric. Forest Meteorol., 58,
Relationships: A Modern Synthesis, 7992, 1992.
McGraw-Hill, New York, 1969. Lang, A.R.G., and Yueqin Xiang, Estimation
Kramer, P.J., and W.S. Clark, A comparison of leaf area index from transmission of
of photosynthesis in individual pine needles direct sunlight in discontinuous canopies,
and entire seedlings at various light Agric. Forest Meteorol., 37, 229243, 1986.
intensities, Plant Physiol., 22, 5157, Lang, A.R.G., Yueqin Xiang, and
1947. J.M. Norman, Crop structure and the
Kramer, P.J., and J.P. Decker, Relation penetration of direct sunlight, Agric. Forest
between light intensity and rate of Meteorol., 35, 83101, 1985.
photosynthesis of loblolly pine and certain Lang, A.R.G., R.E. McMurtrie, and
hardwoods, Plant Physiol., 19, 350358, M.L. Benson, Validity of surface area
1944. indices of Pinus radiata estimated from
Kramer, P.J. and T.T. Kozlowski, Physiology transmittance of the suns beam, Agric.
of Trees, McGraw-Hill, New York, 1960. Forest Meteorol., 57, 157170, 1991.
Kriedeman, P.E., T.F. Neales, and Larcher, W., Physiological Plant Ecology,
D.H. Ashton, Photosynthesis in relation to translated by M.A. Biederman-Thorson,
leaf orientation and light interception, Springer-Verlag, New York, 1975.
Austral. J. Biol. Sci., 17, 591600, 1964. Larcher, W., Physiological Plant Ecology,
LaDeau, S.L., and J.S. Clark, Rising CO2 corrected printing of the 2nd Edition,
levels and the fecundity of forest trees, translated by M.A. Biederman-Thorson,
Science, 292, 9598, 2001. Springer-Verlag, New York, 1983.
Larcher, W., Physiological Plant Ecology,
Lamb, H., Hydrodynamics, 6th Edition,
3rd Edition, translated by J. Wieser,
Dover, New York, 1932.
Springer-Verlag, Berlin, 1995.
Landsberg, J.J., and P.G. Jarvis, A numerical
Lechowicz, M.J., Why do temperate
investigation of the momentum balance of a
deciduous trees leaf out at different times?
spruce forest, J. Appl. Ecol., 10, 645655,
Adaptation and ecology of forest
1973.
communities, Am. Nat., 124, 821842, 1984.
Landsberg, J.J., and A.S. Thom,
Ledig, F.T., A.P. Drew, and J.G. Clark,
Aerodynamic properties of a plant of
Maintenance and constructive respiration,
complex structure, Quart. J. Roy. Meteorol.
photosynthesis and net assimilation rate in
Soc., 97, 565570, 1971.
seedlings of pitch pine (Pinus rigida Mill.),
Lang, A.R.G., Leaf area and average leaf Ann. Bot., 40, 289300, 1976.
angle from transmission of direct sunlight,
Lee, X., and T.A. Black, Atmospheric
Austral. J. Bot., 34, 349355, 1986.
turbulence within and above a Douglas fir
Lang, A.R.G., Application of some of stand, I. Statistical properties of the velocity
Cauchys theorems to estimation of surface field, Boundary-Layer Meteorol., 64,
areas of leaves, needles and branches of 149174, 1993.
plants, and light transmittance, Agric. Forest Lemon, E.R., Energy and water balance of
Meteorol., 55, 191212, 1991. plant communities, in Environmental
Lang, A.R.G., and R.E. McMurtrie, Total Control of Plant Growth, edited by
leaf areas of single trees of Eucalyptus L.T. Evans, pp. 5578, Academic Press,
grandis estimated from transmittances of New York, 1963.
References 419
Lemon, E.R., Micrometeorology and the Little, E.L., Atlas of United States Trees,
physiology of plants in their natural vol. 1, Conifers and Important Hardwoods,
environment, in Plant Physiology: A U.S. Department of Agriculture
Treatise, edited by F. C. Steward, pp. Miscellaneous Publication 1141, U.S.
203227, Academic Press, New York, 1965. Government Printing Office, Washington,
Leonard, R.E., and C.A. Federer, Estimated D.C., 1971.
and measured roughness parameters for a Lotka, A.J., Contribution to the energetics
pine forest, J. Appl. Meteorol., 12, 302307, of evolution, Proc. Nat. Acad. Sci., USA, 8,
1973. 147154, 1922.
Levy, E.B., and E.A. Madden, The point Ludlow, M.M., and P.G. Jarvis,
method of pasture analysis, N. Z. J. Agric., Photosynthesis in Sitka spruce, I. General
46, 267279, 1933. characteristics, J. Appl. Ecol., 8, 925953,
1971.
Lewis, M.C., and T.V. Callaghan, Tundra, in
Vegetation and the Atmosphere, vol. 2, Case MacMahon, J.A., Warm deserts, in North
Studies, edited by J.L. Monteith, American Terrestrial Vegetation, edited by
pp. 399433, Academic Press, New York, M.G. Barbour and W.D. Billings,
1976. pp. 232264, Cambridge University Press,
New York, 1988.
Lhomme, J.-P., A theoretical basis for the
PriestleyTaylor coefficient, Magi, H., and J. Ross, Phytometric
Boundary-Layer Meteorol., 82, 179191, characteristics and photosynthetic
1997. productivity of the barley stand, II. Growth
dynamics of the assimilation area and
Lindeman, R.L., The trophic dynamic
increase of the dry-matter content. In
aspects of ecology, Ecology, 23, 399418,
Photosynthetic Productivity of Plant Stand,
1942.
pp. 144173, Institute of Physics and
Linder, S., Potential and actual production Astronomy, Academy of Sciences of the
in Australian forest stands, in Research for Estonian SSR, Tartu, 1969 (in Russian).
Forest Management, edited by
Malkonen, E., Annual Primary Production
J.J. Landsberg and W. Parson, pp. 1151,
and Nutrient Cycle in some Scots Pine
CSIRO, Melbourne, 1985.
Stands, Communicationes Instituti
Lindroth, A., Canopy conductance of Forestalis Fenniae (Helsinki) 84, Finnish
coniferous forest related to climate, Water Forest Research Institute, Helsinki,
Resources Res., 21, 297304, 1985. 1974.
Lindroth, A., Aerodynamic and canopy Marie-Victorin, Fr`ere, Les Gymnospermes
resistance of short-rotation forest in relation du Quebec, Contributions du Laboratoire de
to leaf area index and climate, Botanique de lUniversite de Montreal 10,
Boundary-Layer Meteorol., 66, 265279, University of Montreal, 1927.
1993. Martin, H.C., Average winds above and
Liou, K.-N., An Introduction to Atmospheric within a forest, J. Appl. Meteorol., 10,
Radiation, Academic Press, New York, 11321137, 1971.
1980. McCaughey, J.H., and J.A. Davies, Diurnal
List, R.J., Smithsonian Meteorological variation in net radiation depletion within a
Tables, 6th revised Edition, Smithsonian corn crop, Boundary-Layer Meteorol., 5,
Institution, Washington, D.C., 1951. 505511, 1974.
420 References
McNaughton, K.G., Evaporation and growing season, Black Rock Forest Paper 1,
advection, I. Evaporation from extensive 3044, 1936.
homogeneous surfaces, Quart. J. Roy. Monin, A.S., and A.M. Obukhov, The basic
Meteorol. Soc., 102, 181191, 1976. laws of turbulent mixing in the surface layer
McNaughton, K.G., and P.G. Jarvis, of the atmosphere, Akad. Nauk, SSSR Trud.
Predicting the effects of vegetation changes Geofiz. Inst., 24(151), 163187, 1954.
on transpiration and evaporation, in Water Monin, A.S., and A.M. Yaglom, Statistical
Deficits and Plant Growth, vol. 2, edited by Fluid Mechanics: Mechanics of Turbulence,
T.T. Kozlowski, pp. 147, Academic Press, vol. 1, MIT Press, Cambridge, MA, 1971.
London, 1983.
Monk, C.D., An ecological significance of
McNaughton, K.G., and W.T. Spriggs, An evergreenness, Ecology, 47, 504505, 1966.
evaluation of the Priestley and Taylor
Monsi, M., and T. Saeki, Uber den
equation and the complementary
Lichtfaktor in den Pflanzengesellschaften
relationship using results from a mixed layer
und seine Bedeutung fur die
model of the convective boundary layer, in
Stoffproduktion, Jap. J. Bot., 14(1), 2252,
Estimation of Areal Evaporation, edited by
1953.
T.A. Black, D.L. Spittlehouse, M.D. Novak,
and D.T. Price, pp. 89104, International Monteith, J.L., Gas exchange in plant
Association of Hydrological Sciences, communities, in Environmental Control of
Gentbrugge, Belgium, 1989. Plant Growth, edited by L.T. Evans,
pp. 95112, Academic Press, New York,
Mendham, N.J., P.A. Shipway, and R.K. 1963.
Scott, The effects of delayed sowing and
Monteith, J.L., Evaporation and
weather on growth, development and yield
environment, Symp. Soc. Exp. Biol., 19,
of winter oil-seed rape (Brassica napus), J.
205234, 1965.
Agric. Sci., Cambridge, 96, 389416, 1981.
Monteith, J.L., Light interception and
Milburn, J.A., Cavitation: A review past,
radiative exchange in crop stands, in
present and future, in Water Transport in
Physiological Aspects of Crop Yield, edited
Plants under Climatic Stress, edited by
by J.D. Eastin, pp. 89115, American
M. Borghetti, J. Grace, and A. Raschi,
Society of Agronomy, Madison, WI, 1969.
pp. 1426, Cambridge University Press,
Cambridge, U.K., 1993. Monteith, J.L., Principles of Environmental
Physics, Elsevier, New York, 1973.
Miller, H.G., J.D. Miller, and J.M. Cooper,
Biomass and nutrient accumulation at Monteith, J.L., Climate and the efficiency of
different growth rates in thinned crop production in Britain, Phil. Trans. Roy.
plantations of Corsican pine, Forestry, 53, Soc. Lond., B., 281, 277294, 1977.
2339, 1980. Monteith, J.L., Climatic variation and the
Miller, P.C., Leaf temperatures, leaf growth of crops, Quart. J. Roy. Meteorol.
orientation and energy exchange in quaking Soc., 107, 749774, 1981.
aspen (Populus tremuloides) and Gambells Mooney, H.A., The carbon balance of plants,
oak (Quercus gambellii) in central Colorado, Annu. Rev. Ecol. Syst., 3, 315346, 1972.
Oecol. Plant., 2, 3, 241270, 1967. Mooney, H.A., J. Kummerow,
Mitchell, H.L., Trends in the nitrogen, A.W. Johnson, D.J. Parsons, S. Keeley,
phosphorus, potassium and calcium content A. Hoffmann, R.I. Hays, J. Giliberto, and
of the leaves of some forest trees during the C. Chu, The producers: their resources and
References 421
Luftschicht, Beitrage z. Phys. d. freien Pike, J.G., The estimation of annual runoff
Atmos., 24, 163189, 1937. from meteorological data in a tropical
Parkhurst, D.F., and O.L. Loucks, Optimal climate, J. Hydrol., 2, 116123, 1964.
leaf size in relation to environment, J. Ecol., Poldmaa, V., On distribution of diffuse
60, 505537, 1972. radiation over the sky, in Investigations in
Pearson, G.A., The growing season of Atmospheric Physics, vol. 4, pp. 111119,
western yellow pine, J. Agric. Res., 29, Institute of Physics and Astronomy,
203204, 1924. Academy of Sciences of the Estonian SSR,
Tartu, 1963 (in Russian).
Penman, H.L., Natural evaporation from
Prandtl, L., Ueber die ausgebildete
open water, bare soil, and grass, Proc. Roy.
Turbulenz, Proc. 2nd Int. Cong. Appl.
Soc. Lond., A, 193, 120145, 1948.
Mech., 6275, 1926.
Penning de Vries, F.W.T., A.H.M.
Prandtl, L., Essentials of Fluid Mechanics,
Brunsting, and A.H. van Laar, Products,
Hafner, New York, 1952.
requirements and efficiency of biosynthesis,
a quantitative approach, J. Theoret. Biol., Priestley, C.H.B., and R.J. Taylor, On the
45, 339377, 1974. assessment of surface heat flux and
evaporation using large-scale parameters,
Perrier, A., Land surfaces: vegetation, in
Monthly Weather Rev., 100(2), 8192, 1972.
Land Surface Processes in Atmospheric
Pruitt, W.O., D.L. Morgan, and
General Circulation Models, edited by
F.J. Lourence, Momentum and mass
P.S. Eagleson, pp. 395448, Cambridge
transfers in the surface boundary layer,
University Press, Cambridge, U.K.,
Quart. J. Roy. Meteorol. Soc., 99, 370386,
1982.
1973.
Philip, J.R., The theory of infiltration, IV.
Raison, R.J., B.J. Myers, and M.L. Benson,
Sorptivity and algebraic infiltration
Dynamics of Pinus radiata foliage in
equations, Soil Sci., 84, 257264, 1957.
relation to water and nitrogen stress, I.
Philip, J.R., General method of exact Needle production and properties, Forest
solution of the concentration-dependent Ecol. Mgmt, 52, 139158, 1992.
diffusion equation, Austral. J. Phys., 13(1),
Rauner, J.L., Deciduous forests, in
112, 1960.
Vegetation and the Atmosphere, vol. 2, Case
Philip, J.R., The distribution of foliage Studies, edited by J.L. Monteith,
density with foliage angle estimated from pp. 241264, Academic Press, New York,
inclined point quadrat observations, Austral. 1976.
J. Bot., 13, 2, 357366, 1965. Raupach, M.R., Canopy transport processes,
Philip, J.R., Plantwater relations: some in Flow and Transport in the Natural
physical aspects, Annu. Rev. Plant Physiol., Environment: Advances and Applications,
17, 245268, 1966. edited by W.L. Steffen and O.T. Denmead,
Philip, J.R., The theory of infiltration, in pp. 95127, Springer-Verlag, Berlin, 1988.
Advances in Hydroscience, vol. 5, edited by Raupach, M.R., Drag and drag partition on
V.T. Chow, pp. 215296, Academic Press, rough surfaces, Boundary-Layer Meteorol.,
New York, 1969. 60, 375395, 1992.
Philip, J.R., A physical bound on the Bowen Raupach, M.R., Simplified expressions for
ratio, J. Climate Appl. Meteorol., 26, 8, vegetation roughness length and zero-plane
10431045, 1987. displacement as functions of canopy height
References 423
and area index, Boundary-Layer Meteorol., Ripley, E.A., and R.E. Redmann, Grassland,
71, 211216, 1994. in Vegetation and the Atmosphere, vol. 2,
Raupach, M.R., and J.J. Finnigan, Case Studies, edited by J.L. Monteith,
Single-layer models of evaporation from pp. 349398, Academic Press, New York,
plant canopies are incorrect but useful, 1976.
whereas multilayer models are correct but Rodrguez-Iturbe, I., D.R. Cox, and
useless: discuss, Austr. J. Plant Physiol., V. Isham, Some models for rainfall based on
15, 705716, 1988. stochastic point processes, Proc. Roy. Soc.
Raupach, M.R., and R.H. Shaw, Averaging Lond., A, 410, 269288, 1987.
procedures for flow within vegetation Rodrguez-Iturbe, I., D.R. Cox, and
canopies, Boundary-Layer Meteorol., 22, V. Isham, A point process model for rainfall:
7990, 1982. further developments, Proc. Roy. Soc. Lond.,
A, 417, 283298, 1988.
Raupach, M.R., and A.S. Thom, Turbulence
in and above plant canopies, Annu. Rev. Rogers, R.W., and W.E. Westman, Seasonal
Fluid Mech., 13, 97129, 1981. nutrient dynamics of litter in a subtropical
eucalypt forest, North Stradbroke Island,
Raupach, M.R., A.S. Thom, and I. Edwards,
Austral. J. Bot., 25, 4758, 1977.
A wind tunnel study of turbulent flow close
to regularly arrayed rough surfaces, Rogers, R.W., and W.E. Westman, Growth
Boundary-Layer Meteorol., 18, 373397, rhythms and productivity of a coastal
1980. subtropical Eucalyptus forest, Austral.
J. Ecol., 6, 8598, 1981.
Raupach, M.R., R.A. Antonia, and
Rogerson, T.L., Throughfall in pole-sized
S. Rajagopalan, Rough-wall turbulent
Loblolly pine as affected by stand density, in
boundary layers, Appl. Mechs. Rev., 44,
International Symposium on Forest
125, 1991.
Hydrology, edited by W.E. Sopper and
Reader, R., J.S. Radford, and H. Lieth, H.W. Lull, pp. 187190, Pergamon Press,
Modeling important phytophenological New York, 1967.
events in Eastern North America, in Romell, L.G., Vaxttidsundersokningar a tall
Phenology and Seasonality Modeling, och gran, Meddel. fran Statens
edited by H. Lieth, pp. 329343, Skogsforsoksanstalt, 22, 45124, 1925.
Springer-Verlag, New York, 1974.
Rosen, R., Optimality Principles in Biology,
Rees, L.W., Growth studies in forest trees: Plenum Press, New York, 1967.
Picea rubra Link., J. Forest., 27, 384403,
Rosenzweig, M.L., Net primary productivity
1929.
of terrestrial communities: prediction from
Relf, E.F., Discussion of the Results of climatological data, Am. Nat., 102, 6774,
Measurements of the Resistance of Wires, 1968.
with some Additional Tests on the Resistance Ross, J., Radiative transfer in plant
of Wires of Small Diameter, 5 pp., communities, in Vegetation and the
Aeronautics Research Council of Great Atmosphere, vol. 1, Principles, edited by
Britain Reports and Memoranda No. 102, J.L. Monteith, pp. 1355, Academic Press,
HMSO, London, 1914. New York, 1975.
Richards, L.A., Capillary conduction of Ross, J., The Radiation Regime and
liquids through porous mediums, Physics, 1, Architecture of Plant Stands, Dr. W. Junk,
318333, 1931. Boston, MA, 1981.
424 References
Ross, J., and Nilson, T., The extinction of Salvucci, G.D., and P.S. Eagleson, A Test of
direct radiation in crops. In Questions on Ecological Optimality for Semiarid
Radiation Regime of Plant Stand, pp. 2564, Vegetation, Ralph M. Parsons Laboratory
Institute of Physics and Astronomy, Report No. 335, Department of Civil
Academy of Sciences of the Estonian SSR, Engineering, MIT, Cambridge, MA, 1992.
Tartu, 1965 (in Russian). Salvucci, G.D., and D. Entekhabi, Hillslope
Rouse, H., Elementary Mechanics of Fluids, and climatic controls on hydrologic fluxes,
John Wiley, New York, 1946. Water Resources Res., 31(7), 17251739,
1995; corrected in Water Resources Res.,
Russell, G., P.G. Jarvis, and J.L. Monteith,
33(1), 277, 1997.
Absorption of radiation by canopies and
stand growth, in Plant Canopies: Their Saugier, B., Sunflower, in Vegetation and the
Growth, Form and Function, edited by Atmosphere, vol. 2, Case Studies, edited by
G. Russell, B. Marshall, and P.G. Jarvis, J.L. Monteith, pp. 87119, Academic Press,
pp. 2139, Cambridge University Press, New York, 1976.
Cambridge, U.K., 1989a. Schlichting, H., Boundary Layer Theory,
Russell, G., B. Marshall, and P.G. Jarvis translated by J. Kestin, McGraw-Hill, New
(editors), Plant Canopies: Their Growth, York, 1955.
Form and Function, Cambridge University Schmidt, W., Der Massenaustausch bei der
Press, Cambridge, U.K., 1989b. ungeordneten Stromung in freier Luft und
seine Folgen, Sitzber. Kais. Akad. Wissen.
Rutter, A.J., Water consumption by forests,
Wien [2a], 126, 757804, 1917.
in Water Deficits and Plant Growth, vol. 2,
Plant Water Consumption and Response, Scott, R.K., and E.J. Allen, Crop
edited by T.T. Kozlowski, pp. 2384, physiological aspects of importance to
Academic Press, New York, 1968. maximum yields in potatoes and sugar beet.
In Maximum Yield of Crops, edited by
Sakai, A., Low temperature exotherms of
J.K.R. Gasser and B. Wilkinson, pp. 2530,
winter buds of hardy conifers, Plant Cell
HMSO, London, 1978.
Physiol., 19, 14391446, 1978.
Seginer, I., P.J. Mulhearn, E.F. Bradley, and
Sakai, A., Freezing avoidance mechanism of J.J. Finnigan, Turbulent flow in a model
primordial shoots of conifer buds, Plant Cell plant canopy, Boundary-Layer Meteorol.,
Physiol., 20, 13811390, 1979. 10, 423453, 1976.
Sakai, A., and C.J. Weiser, Freezing Sellers, P. J., Vegetationcanopy spectral
resistance of trees in North America with reflectance and biophysical processes, in
reference to tree regions, Ecology, 54, Theory and Applications of Optical Remote
118126, 1973. Sensing, edited by G. Asrar, pp. 297335,
Salvucci, G.D., A test of ecological John Wiley, New York, 1989.
optimality for semiarid vegetation, SM Sellers, P.J., Y. Mintz, Y.C. Sud, and
thesis, Department of Civil Engineering, A. Dalcher, A simple biosphere model (SiB)
MIT, 1992. for use within general circulation models,
Salvucci, G.D., An approximate solution J. Atmos. Sci., 43(6), 505531, 1986.
for steady vertical flux of moisture through Sharkey, T.D., Photosynthesis in intact
an unsaturated homogeneous soil, leaves of C3 plants: Physics, physiology and
Water Resources Res., 29(11), 37493753, rate limitations, Bot. Rev., 51(1), 53105,
1993. 1985.
References 425
Shaw, R.H., and A.R. Pereira, Aerodynamic Stanhill, G., A simple instrument for the
roughness of a plant canopy: A numerical field measurement of turbulent diffusion
experiment, Agric. Meteorol., 26, 5165, flux, J. Appl. Meteorol. 8, 509513, 1969.
1982. Stone, E.C., Dew as an ecological factor,
Shuttleworth, W.J., Evaporation, Institute of Ecology, 38, 407422, 1957.
Hydrology Report 56, Institute of
Sun, R.J., and J.B. Weeks, Bibliography of
Hydrology, Wallingford, U.K., 1979.
Regional Aquifer-System Analysis Program
Shuttleworth, W.J., Micrometeorology of of the U.S. Geological Survey, 197891,
temperate and tropical forest, Phil. Trans. U.S. Geological Survey Water Resources
Roy. Soc. Lond. B, 324, 299334, 1989. Investigation Report 91-4122, U.S.
Shuttleworth, W.J., and I.R. Calder, Has the Government Printing Office, Washington,
PriestleyTaylor equation any relevance to D.C., 1991.
forest evaporation? J. Appl. Meteorol., 18, Swinbank, W.C., Long-wave radiation from
639646, 1979. clear skies, Quart. J. Roy. Meteorol. Soc.,
Sinnott, E.W., Plant Morphogenesis, 89, 339348, 1963, with discussions 90,
McGraw-Hill, New York, 1960. 488493, 1964.
Slatyer, R.O., The influence of progressive Szeicz, G., Solar radiation for plant growth,
increases in total soil moisture stress on J. Appl. Ecol., 11, 617636, 1974.
transpiration, growth and internal water
Taylor, F.G., Jr., Phenodynamics of
relationships of plants, Austral. J. Biol. Sci.,
production in a mesic deciduous forest, in
10, 320336, 1957a.
Phenology and Seasonality Modeling,
Slatyer, R.O., Significance of the permanent Ecological Studies: Analysis and Synthesis,
wilting percentage in studies of plant and vol. 8, edited by H. Lieth, pp. 237254,
soil water relations, Bot. Rev., 23, 585636, Springer-Verlag, New York, 1974.
1957b.
Thom, A., An Investigation of Fluid Flow in
Slatyer, R.O., PlantWater Relationships, Two Dimensions, Aeronautical Research
Academic Press, New York, 1967. Committee Reports and Memoranda 1194,
Slatyer, R.O., and I.C. McIlroy, Practical HMSO, London, 1929.
Micrometeorology, CSIRO, Melbourne, Thom, A.S., Momentum absorption by
Australia, 1961. vegetation, Quart. J. Roy. Meteorol. Soc.,
Smith, J.W., Phenological dates and 97, 414428, 1971.
meteorological data recorded by Thomas Thom, A.S., Momentum, mass and heat
Mikesell between 1873 and 1912 at exchange of plant communities, in
Wauseon, Ohio, Monthly Weather Review, Vegetation and the Atmosphere, vol. 1,
Supplement 2, Part II, W.B. No. 558, Principles, edited by J.L. Monteith,
pp. 2193, U.S. Government Printing pp. 57109, Academic Press, New York,
Office, Washington, D.C., 1915. 1975.
Solot, S., Computation of depth of Thornthwaite, C.W., and J.R. Mather,
precipitable water in a column of air, Instructions and tables for computing
Monthly Weather Rev., 67, 100103, 1939. potential evapotranspiration and the water
Spurr, S.H., and B.V. Barnes, Forest balance, Drexel Institute of Technology, Lab.
Ecology, 3rd Edition, John Wiley, Climatol., Pub. Climatol., 10, 181311,
New York, 1980. 1957.
426 References
Tolsky, A.P., Trans. For. Exp. St. Petersburg, U.S. Army Corps of Engineers, Snow
47, 1913 (Abstract by R. Zon For. Quart., Hydrology, North Pacific Division, Portland,
12, 277278, 1914.) OR, 1956. (Available from U.S. Department
Tranquillini, W., Die Bedeutung des Lichtes of Commerce, Clearinghouse for Federal
und der Temperature fur die Scientific and Technical Information as PB
Kohlensaureassimilation von Pinus cembre 151660.)
Jungwachs an einem hochalpinen Standort, U.S. Weather Service, Normals, Means and
Planta, 46, 154178, 1955. Extremes 1974, Climatological Bulletin,
Trewartha, G.T., An Introduction to Climate, Washington, D.C., 1974.
3rd Edition, McGraw-Hill, New York, Valentine, H.T., Budbreak and leaf growth
1954. functions for modelling herbivory in some
Tubbs, C.H., and D.R. Houston, Fagus gypsy moth hosts, Forest Sci., 29, 607617,
grandifolia Ehrh., in Silvics of North 1983.
America, vol. 2, Hardwoods, coordinated Van Bavel, C.H.M., Potential evaporation:
by R.M. Burns and B.H. Honkala, The combination concept and its
pp. 325332, Agriculture Handbook 654, experimental verification, Water Resources
U.S. Department of Agriculture, Res., 2(3), 455467, 1966.
Washington, D.C., 1990. Van Keulen, H., J. Goudriaan, and
Uchijima, Z., Studies on the micro-climate N.G. Seligman, Modelling the effects of
within plant communities, I. On the nitrogen on canopy development and crop
turbulent transfer coefficient within plant growth, in Plant Canopies: Their Growth,
layer, J. Agric. Meteorol. (Nogyo Kisho) Form and Function, edited by G. Russell,
Japan, 18, 19, 1962. B. Marshall, and P.G. Jarvis, pp. 83104,
Cambridge University Press, Cambridge,
Uchijima, Z., Maize and rice, in Vegetation
U.K., 1989.
and the Atmosphere, vol. 2, Case Studies,
edited by J.L. Monteith, pp. 3364, Vertessy, R.A., R. Benyon, and
Academic Press, New York, 1976. S.R. Haydon, Melbournes forest
catchments: Effect of age on water yield,
Uchijima, Z., and J.L. Wright, An
Water (Australian Water and Wastewater
experimental study of air flow in a corn
Association), 21(2), 1720, 1994.
plantair layer, Bull. Nat. Inst. Agric. Sci.
Japan (Nogyo Gijutsu Kenkyusho Hokoku), Walter, H., Die Wasservergorgung der
A11, 1966, 1964. Wustenpflanzen, Scientia, Bologna, 59, 17,
1962.
Uchijima, Z., T. Udagawa, T. Horie, and
K. Kobayashi, Studies of energy and gas Walter, H., Vegetation of the Earth in
exchange within crop canopies, VIII. Relation to Climate and the
Turbulent transfer coefficient and foliage Eco-Physiological Conditions,
exchange velocity within a corn canopy, Springer-Verlag, New York, 1973.
J. Agric. Meteorol. (Nogyo Kisho) Japan, Walter, H., and S.-W. Breckle, Ecological
25, 215227, 1970. Systems of the Geobiosphere, vol. 1,
Udagawa, T., Z. Uchijima, T. Horie, and Ecological Principles in Global Perspective,
K. Kobayashi, Studies of energy and gas translated by S. Gruber, Springer-Verlag,
exchange within crop canopies, III. Canopy New York, 1985.
structure of corn plants, Proc. Crop Sci. Walter, H., and S.-W. Breckle, Ecological
Japan, 37, 589596, 1968. Systems of the Geobiosphere, vol. 2,
References 427
Agriculture Forest Service and Soil of computing vertical transfer within a crop,
Conservation Service in cooperation with Agron. J., 59, 427432, 1967.
the Arizona Agricultural Experiment Zahner, R., Water deficits and growth of
Station, U.S. Government Printing Office, trees, in Water Deficits and Plant Growth,
Washington, D.C., 1967. edited by K.K. Kozlowski, pp. 191254,
Wilson, C.C., Fog and atmospheric carbon Academic Press, New York, 1968.
dioxide as related to apparent Zim, H.S., and A.C. Martin, Trees: A Guide
photosynthetic rate of some broadleaf to Familiar American Trees, Golden Press,
evergreens, Ecology, 29, 507508, 1948. New York, 1987.
Wilson, N.R., and R.H. Shaw, A higher Zwieniecki, M.A., and N.M. Holbrook,
order closure model for canopy flow, Diurnal variation in xylem hydraulic
J. Appl. Meteorol., 16, 11971205, 1977. conductivity in white ash (Fraxinus
Wit, C.T. de, Photosynthesis of Leaf americana L.), red maple (Acer rubrum L.)
Canopies, 57 pp., Agricultural Research and red spruce (Picea rubens Sarg.), Plant
Report No. 663, Center for Agricultural Cell Envir., 21, 11731180, 1998.
Publications and Documentation, Zwieniecki, M.A., P.J. Melcher, and N.M.
Wageningen, The Netherlands, 1965. Holbrook, Hydrogel control of xylem
Wright, J.L., and K.W. Brown, Comparison hydraulic resistance in plants, Science, 291,
of momentum and energy balance methods 10591062, 2001.
Author index
Abramowitz, M., 28, 178, 179, 298, 362 Blum, B.M., 311
Addicott, F.T., 52 Bormann, F.H., 169, 295, 404
Ahlgren, C.E., 52 Boysen-Jensen, P., 244, 248
Akashi, S., 106 Bradley, E.F., 31, 82, 110
Albrektson, A., 250 Brady, N.C., 319
Allen, C.D., 163 Bras, R.L., 181
Allen, E.J., 250 Bray, J.R., 312
Allen, L.H., 31, 33, 101, 112 Breckle, S.-W., 31, 54, 401
Allen, L.H. Jr., 61, 239, 248, 280, 289, 312 Briggs, L.J., 256
Amiro, B.D., 22, 31, 32, 86, 103 Briscoe, P.V., 162
Anderson, M.C., 57, 61, 62, 63 Britton, N.L., 53,
Anderson, T.C. Jr., 55, 196, 197, 204, 228, 229, 289, Brix, H., 181, 258
312, 313, 354 Brooks, R.H., 172, 183, 360
Aronsson, A., 169 Brown, H.P., 55
Arris, L.L., xviii, xix, 53, 218, 316, 317, 318 Brown, K.W., 33, 81, 113, 129
Aylor, D.E., 33 Brusa, R.W., 37
Brutsaert, W.H., 82, 84, 85, 87, 90, 101, 102, 128,
Baker, F.S., 14, 18, 75, 76, 78, 79, 196, 211, 212, 163, 227
216, 226, 231, 238, 280, 282, 290, 291, 304, 312, Budyko, M.I., 186, 289, 349, 350
315, 316, 398
Bunce, J.A., 181
Baker, J.B., 277
Burdine, N.T., 183
Baker, M.B. Jr., 197, 204, 290, 353, 354
Burns, R.M., 277, 280
Baldwin, H.I., 52, 54, 55
Businger, J.A., 129
Barnes, B.V., 319, 320
Bartholomew, W.V., 312
Benecke, U., 181 Calder, I.R., 163, 164
Benney, D.J., xix, 298 Callaghan, T.V., 407
Billings, W.D., 275 Cannell, M.G.R., 21, 32, 250, 296, 312
Birkebak, R., 47 Carbon, B.A., 217
Black, R.D., 176 Carslaw, H.S., 184
Black, T.A., 20, 31, 103 Cauchy, A., 103
429
430 Author index
Houston, D.R., 278, 311 Lemon, E.R., 33, 61, 97, 112, 239, 248, 280,
Hughes, G., 250 289, 312
Leonard, R.E., 101
Illick, J.S., 55 Leverenz, J.W., 249
Inoue, E., 33, 86, 113 Levy, E.B., 21
Iqbal, M., xxxiv, 39, 289 Lewis, M.C., 407
Lhomme, J.-P., 163
Jacquemin, B., 31, 33 Likens, G.E., 169, 280, 295, 404
Jaeger, J.C., 184 Lindeman, R.L., 252
James, J.B., 107 Linder, S., 249, 250
Jarvis, P.G., 20, 22, 30, 31, 61, 78, 79, 84, 86, 101, Lindroth, A., 31, 32
104, 107, 114, 115, 118, 119, 120, 124, 125, 139, Liou, K.-N., 37
159, 168, 218, 225, 228, 239, 246, 248, 249, 250, List, R.J., 38
280, 289, 301 Little, E.L., 317
Jasinski, M.F., 197 Lotka, A.J., 2
Jensen, M.E., 52, 349, 350, 351 Loucks, O.L., xvi, 2
Jensen, P., 169 Ludlow, M.M., 246
Jones, L.A., 49 Lyon, J.L., 52
Jordan, C.F., 21
Page numbers given in bold type refer to plates; those in italics refer to figures.
434
Subject index 435