Water Resources Research Institute STUDIES of the FOREST ENERGY BUDGET

SF E Studies of the Forest O Energy Budget by L. W. Gay H . R . Holbo

. 11 4li Water Resources Research Institute STUDIES OF THE FOREST ENERGY BUDGET

L . W . Gay and H . R . Holbo

Water Resources Research Institut e and Department of Forest Engineerin g Oregon State Universit y Corvallis, Oregon 97331 Summary

Application of the mean-profile energy budget analysis to forests i s discussed, with regard to limitations imposed by the forest environment . These limitations are primarily associated with the diffuse structure an d aerodynamically rough nature . of, forest canopies, which contrast with th e dense and relatively smooth canopies of agricultural crops . The results of the few forest energy budgets that have been tomplete d are reviewed . The country of origin and number . of articles reviewed include : Australia, 2 ; Belgium, 2 ; Canada, 5 ; Federal Republic of Grmany, 8 ; German Democratic Republic, 2 ; Soviet Union, 2 ; United Kingdom, I ; and United States , 2 . Results from these studies are compared with a clear weather energ y budget for a second-growth Douglas-fir site on . the Cedar River east of Seattle , Washington . On August 10, 1972, the energy budget components of the 28- m forest were (cal/cm 2 24-hours) : net, radiation, 399 ; change in stored energy , 2 ; latent energy, -240 ; convection, -161 . Budgets are also tabulated for a clear and far an overcast day at this site during the summer of 1971 . The results are based on the Bowen ratio model . The Cedar River results for August 10, 1972 are the mean of estimate s made from two towers, 100 m apart . The individual latent energy fluxes, fo r example, were 18 percent above and below the mean value of -240 cal/cm 2 24 - hours . Despite obvious limitations upon interpretation of variability fro m just two estimates, this appears to be the first forest study to replicat e the energy budget measurements . The variability found here confirms th e necessity for replication in future experiments . The literature review confirmed that among mean-profile methods, th e Bowen ratio was the most widely used model for forest energy budget analysis . Graphical, and statistical techniques for estimating the Bowen ratio ar e discussed . A few studies have attempted to compare the Bowen ratio result s with those obtained by different methods . The aerodynamic methods have no t generally appeared satisfactory-over rough forest canopies ; this was also the case at the Cedar River site . Tests and comparison of methods must continu e in order to improve instrumentation and experimental design . Acknowledgements

The work upon which this publication is based was supported in part by funds provided by the U .S . bepartthent of Interior, Office of Wate r Resources Research, as authorized u1der the Water Resources Research Ac t of 1964, and administered by the Water Feddardet Research Institute , Oregon State University ; and the %atib%al Stiehce Foundation Brant No . GB-36910X to the Coniferous Fbresf Blatt, V .S . Analysis of Ecosyttems s International Biological Program . This i-t contfibution no . 109 frof th e Coniferous Forest Biome . Portions of tht material in this report hav e been presented elSewhere 1 2 . Professor Leo J . Fritschen, Collet Of Forest Resources, University of Washington, cooperated id fdtiy phll- of this Work . Professor Janusz Paszynski, Institute of Geography of tit POlish Academy of Sciences , Warsaw, and Dr . Nicola Tarsia, ENtC Cettral Forest Experiment Station ; Rome, provided valuable Ataidtance duriii+g the 1972 field seasi n .

1 Gay, L . W . and J . B . Stewart . Energy balance studies in coniferou s forests . Unpublished paper presented to Swedish Coniferous Forest Biom e Seminar, JHdrags, Sweden, May 17, 1973 . 24 pp ., mimeo .

2Fritschen, L . J ., L . W . Gay and H . R . Hcflbo . Estimating evapotranspiratio n from forests by meteorological and lysimetric methods . Symposium Oh Evapotranspiration from Natural Terrain . American Geophysical Union, Fal l Annual Meeting, San Francisco, Calif ., Ddt . 12, 1973 .

Table of Content s

Summary Acknowledgement s List of Figure s List of Table s List of Symbol s

1. INTRODUCTION 1

2. ENERGY TRANSFER FUNDAMENTAL S 2

2 .1 Energy Budget Equation 2 .2 Radiation Exchange 2 .3 Latent and Sensible Energ y 2 .3 .1 Transfer Equation s 2 .3 .2 Bowen Ratio 2 .3 .3 Aerodynamic Method s 2 .3 .4 Eddy Correlation Methods 2 .4 Stored Energ y 2 .5 Photosynthesi s 2 .6 Remarks

3. RESTRICTIONS IN APPLICATION OF MICROMETEOROLOGICAL METHODS 6

3 .1 Steady-State Conditions , 6 3 .2 Surface Homogeneity 7 3 .3 Boundary Layer Integrity 7

4. THE SCOPE OF STUDIES CARRIED OUT ELSEWHERE 9

4 .1 Survey of Recent Forest Energy Budget Experiments 9 4 .1 .1 Australia 9 4 .1 .2 Belgium 9 4 .1 .3 Canada 10 4 .1 .4 Federal Republic of Germany 10 4 .1 .5 German Democratic Republic 11 4 .1 .6 Soviet Union 11 4 .1 .7 United Kingdom 12 4 .1 .8 United States 12 4 .2 General Comments 12

5. EXPERIMENTAL WORK IN DOUGLAS-FIR AT CEDAR RIVER 13

5 .1 Environmental Instrumentation 13 5 .2 Site Characteristics 14 5 .3 Field Measurements 14

iii

6 . RESULTS kid UISC J ON

6 . 1 Cedar River Results 1 8 6 . 2 Comparison with Other Fot#sjnergy Budget s 20 6 .3 Compa I geen of Methods 2 1 6 .3 .1 CogpatisOhs Between Iettbroio ±cal Met fides 2 1 6 .3 .2 Cdinparisons With Dire`ct Measureunl b 22 6 .3 .3 Tests by RepliQhtlOn 2 2

7 . PRACTICAL DIFFICIJITIES N ESTIMATING FORESt E\APOTOA SPIRA ON 2 3

7 .1 Problems of Dimension 1t 7 .1 .1 Canopy Dimension 2 7 .1 .2 Abrodyna iic Diit nsidiis 23 7 .2 R b1 ins of rnstrumetttatic ► 24 7 .3 DIrtIcultle$ in Analysis 24

8 . -CONCLUDING REMARKS 2 7

9 . LITERATURE CITED 29

APPENDIX I . Energy Flux Studies in a Cdtilferou Eoie"st t.ciigsteb . 3 5

List of Figure s

Figure 1 . The Cedar River site in 1971 .

Figure 2 . The Cedar River site in 1972 .

Figure 3 . The instrument hoist .

Figure 4 . The suspended instrument package .

Figure 5 . Energy budget components under clear skies .

Figure 6 . Energy budget components under overcast skies .

Figure 7 . Similarity between gradients of potential temperature an d . vapor pressure .

List of Table s

Table 1 . Energy budget totals at four coniferous forest sites .

Table 2 . Energy budget components at Cedar River, Washington .

.ti

List of Symbol s

Symbol Definition

C .carbon .dioxide concen.tirdtion, !ppm C . volumetric heat •capacU.ty df oir., fiioma s., ,ar .►soil, cal/cm 3OC C -epecifis heat of air., cal/tret DP effective displacement height, •D =d - zo , cm d actual displacement height where -u =%, , .and d = D + z o , cm E water vapour :flux, igm/oni 2 •mfn .e vapor pressure, mb es saturation vapor pressure, :mb G storage heat flux in air, biomass and cal/cm2 ►Vitt g acceleration :due to gravity., 980 cm%sec t H sensible -heat flux, ,cal/cm2 min ih height of trees, m xKC eddy diffusivity for carbon dioxide, cm 2 /sec IKE eddy diffusivity for water vapor, cm 2 /se c =Kg eddy diffusivity for sensible heat, cm2 /sec eddy diffusivity for momentum, cm 2 /sec K+ global radiation flux, cal/cm2 min KT reflected global radiation flux, cal/cm2 min K Viet global radiation flux, cal/cm2 min Ton Karmans constant, 0 .4 1 L4 longwave radiation flux, incoming, cal/cm 2 min Lt‘ 1.ongwave radiation flux, outgoing, cal/cm 2 min L laet longwave radiation flux, cal/cm2 min P net photosynthetic heat flux, cal/cm2 min p atmospheric pressure, mb Q net allwave radiation flux, cal/:cm2 min q specific humidity, (g water/g air) •fluctuations in specific humi ity RI Richardson number - `" T temperature, °C or ° K t time, sec ra aerodynamic resistance, sec/d m rs surface resistance, sec/cm u oindspeed, cm/se c w fluctuation in vertical wind, cm/sec z height above a reference plane, cm r z o roughness length, cm

coefficient in the stability correctdon Tfunict46 Bowen ratio exponent in the stability correction function an operator denoting a finite differenc e ratio of the mole weight cif water :to . Lr, 0 .l62 2 potential temperature, ° K latent heat of fusion, cal/-,g latent heat flux, cal/cm2 min heat of assimilation of carbon, cal/ g density of air, g/cm3 stability correction

1, INTRODUCTION

The key to a comprehensive theory of environment lies in a descriptio n of the transformation and transfer of energy at or near the surface of the earth . Most of the energy transformations take place at the interface . between the solid (or liquid) surface and the gaseous atmosphere, Sola r radiation readily reaches the surface through the relatively transparen t atmosphere . If the quantity of solar radiation is measured, the major - problem remaining is to evaluate the rate at which it is transformed into r If other forms of energy . The majority of these transformations involve forms of radiant , latent or sensible energy . The major effect of these transformations i s the creation and maintenance of the general circulation of the earth . The relatively minor fraction that is photochemically fixed by plants, thoug h small in quantity, does have a significant effect upon mans activities . The cycle of energy transformations at the earth-air interface is also A associated with a corresponding cycle of mass . For example, the mass o f

water cycled at this interface can readily be expressed as energy required I I to effect its transformation from one phase to another, Likewise, th e mass of carbon cycled between the biomass of vegetation and the atmospher e can be expressed as the equivalent energy required to effect a net chang e in phase from gas to solid during the photosynthesis process . Transfers of energy and mass are governed by physical principles . Ap- plication of these principles in the past few decades has proven useful i n studying such biological processes as growth and transpiration . It is no t possible to define all of the physical, chemical and biological factors tha t effect living organisms, but we can identify certain physical factors tha t are exceedingly important in the energy and mass cycles . In doing so, we shall exclude factors pertaining to chemical composition of the soil an d genetic characteristics of the organisms, despite their importance to th e growth processes in plants . The physical factors required to describe the cycles of energy an d mass between vegetated surfaces and the atmosphere are radiation, wind,t ;, , and precipitation, plus the temperature, moisture, and carbon content or the air, vegetation, and soil . We shall follow de Vries (1963) notin g that these factors may have a macro- or microenvironmental scale, dependin g upon whether or not there is evidence of a marked surface influence, suc h as that derived from the type of vegetation, or the stage of its develop - ment . This paper will consider primarily microenvironmental factors in r . its discussion of energy transformations in coniferous forests . The major 1 . question to be considered is : How is radiation transformed into othe r forms of energy at the surface of the earth ? The objectives of the Water Resources Research Institute project were . twofold : to evaluate evapotranspiration from a Douglas-fir forest durin g the summer periods of maximum water use ; and to compare evapotranspiration predictions from several different models that involve a number of climati c parameters . Consideration of these points in relation to the experimenta l work and analyses undertaken leads to a statement of the objectives of thi s paper : -

1. to review the importance of a physical theory of environment , 2. to review the application of such theory to forested areas , 3. to examine the results obtained from a Douglas-fir forest .

This will be done by referring to work carried out elsewhere on the rol e of forest vegetation in the surface energy balance, and then by evaluatin g the energy budget of the Douglas-fir forest .

2 . ENERGY TRANSFER FUNDAMENTAL S

The microenvironmental aspects of basic energy transfer have bee n reviewed in several other papers (Webb, 1965 ; Federer, 1970 ; Tajchman, 1971 ; and Stewart and Thom, 1973 ; among others) ; so derivations need not be repeate d here . Certain basic equations, however, are summarized below as an aid i n interpretation of results to be given later .

2 .1 Energy Budget Equation .

The energy budget equation is merely a statement of the principle o f conservation of energy : the sum of energy fluxes across the boundaries o f a system are equal to the amount of work done on the system . Let the system boundaries be simply a plane of infinite extent, positioned at th e top of the vegetation . The major fluxes between vegetation and the atmospher e can then be identified and summed in the widely used energy budget equatio n

Q + XE + H + G + P = O . (1)

The first three terms represent fluxes across the reference plane whic h serves as the system boundary . Q is the net exchange of radiation, AE i s the latent heat flux, and H is the sensible heat flux . The last two term s represent changes in energy within the system, i .e ., the volume beneath th e boundary plane . G is the change in sensible heat stored within the air , biomass and soil, while P is the net transformation of energy by photo - synthesis and respiration . The usual sign convention is adopted so tha t Q, JOE and H are positive when their sense is directed downward, and G and P are positive when they represent energy leaving storage . Each of the components in Equation (1) can be expressed in terms of the factor selected t o represent the physical environment .

2 .2 Radiation Exchange .

Net radiation is the net exchange of radiation crossing the referenc e plane over all wavelengths . It can be represented as the sum of the short - wave (K) and longwave (L) fluxe s

Q=K+-KT+L+-LT=K+L (2 )

where the arrows indicate direction, and () signifies net exchange . Net radiation is also equivalent to the net transformation of energy fro m radiant into non-radiant forms . Its measurement, therefore, represent s the amount of energy that is partitioned into latent, sensible and chemica l energy .

2 .3 Latent and Sensible Energy .

The latent and sensible heat fluxes crossing the reference plane ca n be evaluated by a number of methods . Those considered here are based upo n either time-averaged profile measurements, or turbulent fluctuations o f temperature, moisture concentration, and wind velocity .

2 .3 .1 Transfer Equations .

Basic transfer equations define the latent and sensible hea t fluxes as being proportional to the respective gradients of vapour concentration (approximated by Ae/Az) and potential temperature (A0/Az), o r XE = P A KE Ae/Az (3a) and

H = p C p KH AC/Az . (3b)

KE and KH are the diffusivity coefficients for vapor and for sensible heat , respectively . The other constants are defined in the symbols list . Equation s (3a, 3b) can not be applied directly because the diffusivities are not known . The transfer equations are the basis for the Bowen ratio and for the aerodynamic methods that have proven useful for evaluation of the fluxes .

2 .3 .2 Bowen Ratio .

The Bowen ratio equation is developed by combining Equations (1) , (3a) and (3b) to yield :

AE = - (Q* + G + P)/(1 + (C Pp/ea) AC/Ae )

= - (Q + G + P)/(l + (4a)

and

H = - (Q* + G + P) 13/(1 + S) = XER . (4b )

KE and KH will be assumed equal (Dyer, 1967), so they are eliminated in th e Bowen ratio formula .

2 .3 .3 Aerodynamic Methods .

The aerodynamic equations can be written in a variety of ways t o express the fluxes in terms of the several factors, i .e ., potential temperatur e (0), vapor concentration (e), windspeed (u) and height (z) .

XE = e2A k2 (Ae Au/(0 ln(z-D) ) 2 ) (5a) P and

H = pCP k2 (A0 4u/ (0 In (z-D) ) 2 ) f (5b)

The measurement heights are scaled as z - D to account for the effec t of the forest in displacing the .exchange surface above a reference plan e at ground level . D is an "effective " height . The wind speed, u, goes t o zero at the level D + z o , where zo isthe aerodynamic roughness length . The zero-plane displacement height is thus d = D + z o . Other constants

are defined in the symbols list, but a new factor, (, warrants additiona l consideration . The term 4) is a correction for the effects of atmospheric stability near the surface . It represents the ratio of the transfer coefficient fo r heat (or vapor) to an analogous transfer coefficient for momentum, K M . When the temperature stratification in the atmosphere is adiabatic,, or neutral, the value of this ratio (and of c)) is unity . When a temperatur e gradient exists, however, the transfer of momentum will be enhanced i n unstable (lapse) conditions and suppressed in stable (inversion) condition s and the ratio will depart from unity . A correction is provided by the mode l

(I) = (1 - a Ri) 1 (5c )

where a and y are constants and Ri is the Richardson number (Richardson, 1920) . The Richardson number expresses the relative contributions to transfer o f the buoyancy effects due to heating, and the frictional effects due to wind , It is approximated by

Ri = (g/0) (A0/Az)/(Au/Az)2 (5d )

where 0 is the mean potential temperature of the air layer, and the othe r constants are defined in the symbols list . The value of a and y appear to be related to the heights of measuremen t and to the roughness of the exchanging surfaces ; a range of values have been proposed . For example, one formulation (Paulson, 1970) developed over a grassy surface suggests that a = 15 and y = 0 .75 . The values suitable fo r forests are not yet determined . The application of the aerodynamic metho d will be reviewed later ; its use is not recommended for forests at present .

2 .3 .4 Eddy Correlation Methods .

Eddy correlation methods are based upon turbulent fluctuations i n the atmosphere near the ground . The Bowen ratio and the aerodynamic model s depend upon time-averaged, mean profiles of temperature, vapor and win d speed . The eddy correlation method, in contrast, is based upon the fluctuations in these atmospheric properties about the mean value . The flux of latent energy can be estimated from measurement of th e short-term fluctuations (in the order of one second or less in vertica l wind movement (w ) and specific humidity (q ) . Over a horizontal site o f good uniformity, the upward flux of water vapor (Penman, Angus and Va n Bavel, 1967) is :

E _ (pw)q (6a)

where the bar signifies the mean of the product of the fluctuations . A similar equation yields the upward transfer of sensible heat

H = Cp (pw) T . (6b )

Since the individual determinations of the eddy fluxes are virtually instan- taneous, they must be integrated to yield estimates over the desired perio d of time .

The major technical drawbacks to application of the equations ar e associated with the design and construction of adequate, fast-respons e sensors . The theory has certainly been proven in agricultural applications , - but it remains almost untested for estimating the latent and sensible hea t components of the forest energy budget .

2 .4 Stored Energy . I The flux of stored energy actually represents the quantity of sensibl e energy gained or lostfrom storage beneath the reference plane during th e period of measurements . It can be eXpressed a s

.4 m G = E Ci- ATi Azi /At ( 7 ) 1=1 where the distance beneath the reference plane has been divided into m layers, each with a volumetric heat capacity of Ci and a thickness Az i . The term AT i is the mean temperature change of the ith layer during th e time interval At . Appropriate heat capacities and temperatures can b e used to determine the storage changes in the air, biomass and soil beneat h the reference plane . Stewart and Thom (1973) approximated the total change in stored energ y in a Scots pine forest by the empirical equatio n

G = k 1 AT + k2 Aq (8 ) where AT and Aq are the respective mean hourly changes in temperature (°C/hr . ) and specific humidity in the air layer between the reference plane and th e ground . When G has units of watts per square meter, their constants ar e kl = 18 and k2 = 17 .

2 .5 Photosynthesis . -

The flux of photosynthetic energy is small in forests, normally bein g a percent or less of Q . The flux of photosynthetic energy will be neglecte d in our analyses because of its small magnitude . Also, it is not often measured in the studies to be reviewed . If the gradient of C02 (AC/Az) is measured, net photosynthesis can then be written as a transfer equation, similar to those (Eqns . 3a, 3b) establishe d 8 for sensible and latent energy :

P = -pAK c AC/Az . (9 )

With the assumption that the eddy diffusivity of C O.2 equals those for sensibl e and latent energy (Kc = KH•= KE ), the photosynthetic flux can be written int o 4 a Boweq ratio model (Eqns . 4a, 4b) as was done by Denmead (1969) or into a n aerodynamic model analogous to Eqns- (5a, 5b) .

2 .6 Remarks .

The basic energy transfer-formulations have been thoroughly tested ove r areas that are bare, or else covered with low, relatively-smooth vegetation . The results have been reported by many authors in the past decade . Application to forest, however, has r-evealed problems that are associated with the siz e and scale of forest canopies`, We shall examine these problems, and then loo k at application difficulties after , a brief survey of experimental work under - taken in forests, and an initial examination of results obtained at our Douglas - fir site . 3 . RESTRICTIONS IN APPLICATION; OF MICROMETEOROLOGICAL METHOD S

The micrometeorological methods are used to indirectly estimate the energ y fluxes from.meesurements of the state of the atmosphere . This approach is now routinely used in agriculture, but applications to the forest remain strictly experimental . Therefore, it appears appropriate to discuss the restriction s that have heeo, .evident with experience in forest measurements . The basic micrometeorological methods (Bowen ratio, aerodynamic and eddy correlation) are each derived with certain fundamental assumptions . Tanner (1967} has rview}ec these assumptions, and discusses their effect upon a1p.licatio steady-state conditions, , (2) surface homogeneity, and, (3) boundary lave-r- integrity . It will be helpfu l to examine the effect of each of these requirements upon, the problem o.f- defining the forest energy budget . .

3 .1 Steady-St-ate Conditionss..

The surface energy budget is in a steady-state when the fluxes do no t change with respect to• time . Although the fluxes do change continuousl y throughout the day, the steady-state condition need prevail only during th e time required to sample the mean of the environmental variables . The actual time required depends upon the characteristics of the surface, and upon th e characteristics of the recorder and of the sensors . The interrelationship between these surface/recorder/sen,sar characteristics is not well-defined . It is also apparent that many transient effects occur in the environment , including the effects of clouds on the radiation flux ; the moving patter n of light and shadow at the floor of the fonest ; and, the variability gbserved in wind . Busch C19:73) suggests that steady-state requirements may seldom be met in nature to the degree required by aerodynamic (mean profile) methods . The same pessimistic view applies to the Bowen ratio method . The surface characteristics of forests tend to enhance mixing, which ma y reduce the time during which steady-state conditions prevail . For example , Dyer (1.9,63.) discusses the time required for a change in sensible heat flux to. be propagated up through the boundary layer above a grass surface . He reported that almost a half-minute was needed to -effect a 9 .0 percent adjustmen t of the profiles through a distance of a half-meter . Profile adjusttent should take place mare rapidly above a forest because of the higher intensity of - turbulence found near the rough canopy ; this would tend to shorten the perio d during which steady-state conditions must prevail . On the other hard, forest profile mmaaurements are commonly made at levels as high as 1Q 4 to 20 m abo. e the height of the zero plane displacement . Such large heights will obviously retu a len.ger periods for flux ch ges,to be propagated up through the profii e but Lq. definitive information is available on this problem . Stpdies are needO

on the effects of forest roughness and scale on steady-state requirements . The characteristics of the recorder/sensor system also are importan t in considering the .time required to sample for the mean the environmenta l variables . Two types of considerations apply : the first is the time needed for the data system to scan through the sensors, and the second i s the time required for a given sensor t o . .respond to a change in an environ - mental variable . Typical systems now used for aerodynamic and Bowen rati o studies require about 1 minute to sample 50 different sensors, while typical temperature sensors require 5 minutes to achieve 95 percent adjustment t o a step-change in temperature . The instruments that are now commonly in us e for gradient studies thus lengthen the period during which the steady-stat e condition must apply . A good estimate of the mean is also required in edd y correlation studies, because deviation from the mean is the primary variabl e used in computing eddy fluxes . Byrne (1970) comments upon some consideration s required in evaluating scanning rates and sensor responses ; more work i s needed .

3 .2 Surface Homogeneity .

A homogeneous surface contains uniform sources and sinks for laten t and sensible heat and for momentum . The concept of homogeneity applied t o agriculture concerns various scales of patchiness, e .g ., wet spots, row crops, etc . Apart from dense young plantations, forests can not really b e considered homogeneous . There is an extreme range of structural variabilit y associated with forest, with resulting influences on micrometeorologica l methods . Definitive criteria for homogeneity are lacking, however . Horizontal variability in agricultural surfaces contributes to selectiv e transfer of heat (or vapor) depending upon the scale and nature of th e inhomogeneity (Tanner, 1967) . This creates problems in spatial samplin g (mentioned by Byrne, 1970) and introduces questions regarding the applicatio n of all of the micrometeorlogical relationships . The difficulties appear to be even greater in forests, as the scale o f homogeneity must be extended to account for the appreciable depth of th e forest canopy . Denmead (1964) found that the sources Of heat and vapor an d the sink for momentum occurred at different depths within the canopy of a young pine plantation . Difference in the source/sink levels may be eve n more pronounced in mature forests with more diffusely distributed canopie s that extend through greater depths . If such were the case, it would then b e improper to consider temperature and humidity readings from a given psychromete r as representing energy exchanged by sensible and latent heat transfer from th e same level in the canopy . The extent and the consequences of different source / sink levels are not worked out for forests, but the effects may be considerable .

3 .3 Boundary Layer Int ,wity .

Boundary layer integrity prevails when the properties of the atmospher e derive primarily from exchange with the underlying surface . This layer , when in equilibrium with the surface, will be characterized by relativel y constant fluxes, e .g ., within 10 or 20 percent of surface values (Lumley an d Panofsky, 1964) . Within_sich a boundary layer developed over a homogeneou s surface, all of the net energy transfer will be in the vertical direction an d horizontal gradients will not exist .

7 Measurements made outside of this boundary layer will not properl y represent the exchanges taking place at the surface of interest regardles s of whether the aerodynamic, Bowen ratio, or eddy correlation technique i s used . It is common practice to consider the height of the sensors wit h respect to the thickness of the surface boundary layer, generally wit h reference to the fetch of the site . However, over forests one may inadvertently select measurement heights that are below the boundary laye r (e .g ., below the constant flux zone) . Measurements here would be in th e region dominated by local effects associated with the projecting crown s of the canopy, where horizontal gradients can not be neglected . Crowns of different size and shape, when sunlit, may act as source s of heat and vapor with resulting bouyant plumes that create horizonta l variability in the atmospheric properties . Shaded crowns, or space s 1 betieen crowns, may serve as sinks for sensible heat . Measurements within this region near the canopy are affected by inhomogeneity and horizonta l transport ; they may be as erroneous as measurements made above the upper limit of the surface boundary layer . One must thus consider two limits on sensor heights in order to insur e that measurements are made within the boundary layer of the surface o f interest . The only guidance available for selecting the appropriate minimu m height is Lettaus (1959) rule of thumb that sensors should be at leas t five times the average roughness length above the surface . With forests, th e actual exchange surface is ill-defined . One would presumably refer measurements to the level of the zero-plane displacement . Contradictions appear in the guidelines used to select the maximu m instrument height . A number of formulations developed for the growth o f boundary layer thickness ever smooth surfaces or low vegetation suggest tha t the ratios of fetch to instrument height (again, presumably measured above the zero-plane) are in the order of 100 or 500 to 1 . If such relationship s are extrapolated to forests, then the fetch requirements may reach the rang e of tens of kilometers . It is evident that none of the forest sites studie d to date can meet such stringent fetch requirements . It has been suggested that profile adjustment will be achieved mor e rapidly over a rough, forest surface than over lower, smoother vegetatio n (Webb, 1965 ; Pasquill, 1972) . Such an effect would mean that fetch requirements for forests are considerably less restrictive than commonly supposed . Again, this problem lacks a comprehensive treatment for forest conditions . 4 . THE SCOPE OF STUDIES CARRIED OUT ELSEWHERE

A brief examination ofthe forest energy budget work .underway .elsewh-re . will help place the results obtained here into perspective . When Reifsnyde r (1967) reviewed the status of forest energy budget research, he could cit e only two studies that were complete in the mid-1960s . . These were the pioneering work carried out in Germany during the summer of 1952 by Baumgartne r. (1956) and an International Geophysical Year-project in 1857-60 in Russi a ~(Rauner, 1958 ; Dzerdzeevski, 1963) . In the six years that elapsed since Reifsnyders review, resultshav e been published from a number of forest energy budget studies . These studie s have achieved varying degrees of success under a wide range of experimenta l conditions . It appears worthwhile at this point to examine the condition s under which these studies were conducted in order to better-evaluate the wor k that has been accomplished, and assess that which remains to be done . Most of the projects included in this survey have evaluated all compo - nents of the forest energy budget for a perio d o.f a day or more . Many excellent pape rs that discuss various restricted problems within the contex t of the energy budget have been excluded from the compilation in this section , although they will be referred to as appropriate throughout the remainder o f the paper . We will welcome information on additional citations that wer e either inadvertently omitted or else unknown to us at the time this surve y was compiled . The general conditions surrounding eac h- study will be examined in thi s section . Discussion of specific results will be deferred until Section 6 , when comparisons will be made with our results from Douglas-fir .

4 .1 Survey of Recent Forest Energy Budget Experiments .

The studies that have reported essentially complete energy budgets fo r forest sites were centered in Australia, Belgium, Canada, the Federal Republi c of Germany, the German Democratic Republic, the Soviet Union, the United Kingdo m and the United States . These studies will be described briefly, with respect t o the experimental conditions and methods used .

4 .1 .1 Australia .

Early studies were conducted within the canopy of a young Pinus radiata plantation (Denmead, 1964) ; these have been extended above a 7 .8 m high canopy (Denmead, 1969) . The Bowen ratio method was used to estimate the latent , convective, and photosynthetic energy fluxes between the plantation and th e atmosphere . The fluxes are tabulated as mean daily totals for the dayligh t periods (0600-1800) of four rather clear days in the spring . A taller (13 m ) P . radiata plantation was the site of Moore and Hicks (1973) eddy correlatio n study of energy exchange near Mt . Gambier ; South Australia . They tabulate d hourly means of the energy budget components for the period 0900 through 200 0 for an unspecified number of spring days . The experimental conditions included partly-cloudy skies .

4 .1 .2 Belgium .

A long-term study in Belgium has focused on the energy budget of a n

9 oak forest of some 16 to 21 m in height at Virelles in the Ardennes . Monthly means of the energy budget components for each hour (June through October , 1967) are tabulated by Galoux (1971) . The course of the fluxes throughou t the day are also available for clear sky conditions on July 7, 1967 (Grulois , 1968) and for relatively cloudless skies on August 12, 1967 (Galoux, 1968) . Grulois also tabulates the daily totals of the energy budget components . The analyses are based upon the Bowen ratio measured between the 16 and 26 m levels . The site and instrumentation are described by Galoux, Schnock an d Grulois (1967) .

4 .1 .3 Canada .

Results from Canadian forest energy budget research have evolve d from three different research areas . A micrometeorological study over a young, 7 .8 m high plantation of Douglas-fir near Vancouver extended over si x weeks during the summer of 1970 . A graph of the energy budget component s for each hour is available for the clear day of July 14, 1970 (Black an d McNaughton, 1971) ; the energy budget totals for this and 16 other individua l days are given by McNaughton and Black (1973) . The energy fluxes wer e developed from the Bowen ratio, based upon measurements at 8 .1 and 9 .1 m with reversing, diode psychrometers . Storr, Tomlain, Cork and Munn (1970) applied the Bowen ratio techniqu e to a high elevation spruce and fir forest at Marmot Creek, Alberta . The forest was 25 .8 m (85 feet) high ; the psychrometers were placed at 31 .8 m (105 feet) and 40 .9 m (135 feet) . They present 0700-1900 hr totals of the energy budget components for each day for the period July 8-26, 1967 . They concluded that the energy budget evapotranspiration for the period compare d well with an estimate based on the water budget . The energy budget was evaluated during the summer of 1970 at a Jack pin e (Pinus banksiana) stand of 7-10 m height, near Pinawa, Manitoba (Reimer and Desmaris, 1973) . Monthly means of the hourly values of energy budget components are presented graphically for each of the three summer months . In addition, monthly totals are given for latent energy term . The analyse s were based upon the Bowen ratio method, with dewpoint hygrometers place d at heights of 6 and 21 m .

4 .1 .4 Federal Republic of Germany .

Many of the forest energy budget studies available from German y have been carried out in a Norway spruce (Picea abies) stand near Munich . The height of this stand was 27 .2 m in 1965 . Tajchman (1967) has presente d a detailed report of the initial studies, with monthly sums of the fores t energy budget for the period May 1-September 30, 1965 . Some facets of this work are summarized in Tajchman (1971), and a graph of the hourly component s for a clear and a cloudy day are given in Tajchman (1972a) . Both the Bowen raio and aerodynamic methods were used to analyze psychrometric and win d measurements obtained at heights of 28 .9, 30 .9 and 35 .9 m. The small gradients above the forest made application of these methods difficult, and the mea n monthly results contain a considerable amount of interpolated data . A subsequent report by Strauss (1971a) summarizes monthly means of th e energy budget components at this site for the year 1969 . Additional detail s and daily sums of the components for the entire year are also availabl e (Strauss, 1971b) . Baumgartner (1971) tabulates the monthly evapotranspiration

component at this forest site for the May-September periods of 1965, 1966 . and 1967 . Tajchman has subsequently conducted an energy budget study over a 2 .8 m plantation of Scots pine (Pinus sylvestris) near Freiburg . His initial Bowen ratio analysis (1972b) gives the hourly and daily totals of the energy budge t components for a single clear day . He has recently (1973) reported good result s when comparing the Bowen ratio and,aerodynamicmethods at this site, using dat a collected at 2 .3, 2 .7, 3 .5, 4 .8 and 7 .5 m . Mean values are given for the 0800- 1600 hour period of four different days . - An intensive study of the energy budget of a mature, 26 m beech fores t was undertaken at Soiling as part of the German IBP contribution . The result s and analyses were reported by Kiese (1932) ; he tabulated mean energy budget components for each hour over the nearly six month period from May 3 through October . .20, 1970 . The Bowen ratio was the basic analysis nsed in this lon g term study . Temperature and humidity measurements were made at 25, 28 an d 35 in. Details of the instrumentation and the site characteristics were give n earlier by Kiese (1971) . We are aware that the energy budget of a spruce forest has been in - vestigated as part of the Soiling project, but we have not been able t o obtain a copy of the report s-for review .

4 .1 .5 German Democratic Republic .

Comparative measurements have been made in forest and meadow site s at the Eberswald site . L$tzke (1966) tabulates and graphs the energy budge t components of a 21 m pine forest with beech understory . The hourly data ar e the mean of three consecutive clear days in the autumn of 1964 . The Bowen ratio analysis is based upon measurements made at the 19, 24 .6 or 27 .7 m levels . A subsequent paper (L$tzke, 1969) deals with results from a 12 .6 m pine forest . Components are tabulated as monthly means for the period Apri l 1-September 30, 1 9.67, and as hourly means for a 10-day period in May .

4 .1 .6 Soviet Union .

A limited number of Russian publications have become available i n translation . For example, the results of energy budget studies in spruc e forests at the Valdai Hydrologic Laboratory have been. summarized by Fedorov (1970) . Monthly evapotranspiration means are tabulated for 4 summer month s at the forest, as estimated by Bowen ratio and aerodynamic methods . The other budget components are not given, however . F.orest,energy budget research at the Institute of Geography of th e Academy of Sciences USSR has been carried out at the Zagorsk study statio n since 1957 . Reifsnyder"s (1967) review cited several translations of earl y reports of this work . Another report of this early work is found in Raune r (1962) . Seasonal summaries of-mean daily energ y . budget components of a young oak forest from this area are presented by Rauner (1966) . The data collected in 1962-1963 over a young stand some 5-6 m in height .

1 Miess, M. 1968 . Vergleichen der Darstellung von meteorologische n Messergebnissen and WHrmehaushalts Untersuchungen an drei unterschiedliclen Standorten in Norddeutschland . Ber . Num . 2 . Inst . f . Meteorologie and Klimatologie, Tech . Universitgt, Herrenhauser Str . 2, Hanover .

- 11 - 4 .1 .7 United Kingdom .

Stewart and Thom (1973) have published the first energy budge t results from a continuing study in a Scots pine (P . sylvestris) plantation of 15 .8 m mean height . The experimental site is located near Thetford , Norfolk . They graphically present the hourly energy budget component s (0500-1900) for seven clear days during the summer of 1971 . The analyse s were made . with the Bowen ratio method ; the ratio was determined fro m profiles of temperature and specific humidity measured at 6 levels abov e the canopy . Sensors were psychrometers made with quartz crystal thermometers .

4 .1 .8 United States .

The hourly and daily energy budget for a 28 m Douglas-fi r (Pseudotsuga menziesii) stand at Cedar River, Washington, was evaluated b y Gay (1972b) for 1 clear and 1 overcast day in the summer of 1971 . The analysis was based upon the Bowen ratio, derived from measurements take n at the 26 .16, 28 .16 and 31 .16 m levels . Fritschen and Doraiswamy (1973 ) .have evaluated the hourly values of net radiation and latent energy in th e same stand for two consecutive clear days in .the spring of 1972 . Their latent energy values were measured directly with a weighing lysimeter tha t contained a 28 m Douglas-fir tree . Gay (1973) has also evaluated the hourly energy budget components o f a lodgepole pine (Pinus contorta) stand near Bend, Oregon . The mean height of the forest was about 7 m . The Bowen ratio method was used in analyzin g the measurements which were made at heights of 7 .07, 9 .57 or 12 .07 m. Results are presented for two clear days in late August, 1969 .

4 .2 General Comments .

The information available on forest energy budget experiments wa s surveyed on a geographical basis . It could as easily have been groupe d into two classes, depending upon the purpose of the studies . One clas s would represent long-term studies of the climatology of energy budge t components, while the other would be devoted to intensive studies of th e energy exchange processes . The year-long estimates presented by Strauss (1971b), or the growin g season totals given by Galoux (1971) certainly fit the first category . Studies of this nature are particularly helpful in consideration of seasona l trends in the magnitudes of the energy budget components . In contrast , Stewart and Thom (1973) give an excellent discussion of processes, bu t do not provide daily totals of the energy budget components . It has become evident that much effort is required to obtain hig h quality data for process studies . The long-term evaluations, on the other hand, can often minimize instrumental and observational errors b y averaging over weekly or monthly time periods . This summary of experimental conditions, methods used for analysis, an d temporal extent of results will be useful when making comparisons betwee n studies . We shall now look in greater detail at the experimental work under - taken in this study .

I - ti 16 - S ■ r t 1 i T I -M-Cl 1 1 n1 1 n 5 . EXPERIMENTAL WORK IN DOUGLAS-FIR AT CEDAR RIVER ti -

1 - A substantial field program was carried out in connection with this - y = study . Measurements were made during the summers of 1971 and 1972 at the" . ~,• - intensive study site of the U .S . Coniferous Forest Biome on the Cedar Rive r in the foothills of the Cascades to the east of Seattle, Washington . This (Pseudotsuga menziesii) stand approaches the idea l _ 4 second-growth Douglas-fir specifications of a level, uniform forest of infinite extent . Here the U .S . Coniferous Biome project is sponsoring a variety of inter-related studie s r at the Cedar River site . These involve ecologists, physiologists, soil - scientists, hydrologists and meteorologists . Especially close cooperation - exists between the work described here and the energy budget investigation s of Dr . Leo Fritschen, College of Forest Resources, Univeristy of Washington, . Seattle . Energy budget model testing has been carried out in conjunction with Dr . Fritschens lysimeter installation (Fritschen, 1972) . Similar comparisons may be made with his eddy flux system which began operation late i n 1972 on an experimental basis . The characteristics of the instrumentation, the site, and the field 1 1 measurement program will be described in preparation for a discussion of - results in a subsequent section .

5 .1 Environmental Instrumentation .., ,Y r .

A number of environmental data acquisition systems have been constructed 1 in recent years . Some of these systems have been mobile (Clayton and Merryman , 1960 ; Valli, 1966), but most have been fixed installations (Allen, 1970 ; Backlund and Perttu, 1971 ; Fritschen and Van Bavel, 1963 ; Reifsnyder, 1962) . Fritschen (1970) has reviewed specific requirements for data systems employe d in microclimate research . All of these systems have sought to combine the - attributes of precision and convenience in application, with ease in data , handling and processing . The performance of these and similar systems has continued to improv e with the development of new instruments and with increased experience i n field operation . The major improvements have been in data collection an d handling procedures, as digital data systems have now become common plac e for environmental studies . These systems can readily collect grea t quantities of data with a high degree of precision . The use of small on - line computers is just beginning to minimize the sometimes laborious, an d often slow, data processing tasks associated with environmental studies . The accuracy requirements for resolving very small gradients of T , e and u present the biggest problem in the design of an instrumentatio n system for forest energy budget studies . A discussion of the characteristics of the systems used at Cedar River should be helpful for those planning t o undertake similar studies elsewhere . The system that we used at Cedar River was developed at Oregon State ;~ . University for use in a variety of environmental studies . It has been described in detail by Gay (1973, Ch .2) . The system includes a digital r recorder, a truck-mounted laboratory, a power source, sensors, signal cable s and instrument supports . The design of these components has stressed th e concept of portability, in order to facilitate deployment at differen t experimental sites .

I . , I %

4 . . r, Iti - I l _

- 13 -

The digital recorder will scan up to 100 analog inputs at a rate of 5 per second with a resolution of 0 .001 percent (0 .1 microvolts on a 10 millivol t scale) . The recorder also contains a 12-channel digital scanner, used t o record the pulse output of up to 12 anemometers (Gay and Holbo, 1970) . The system output is on punched paper tape and a printed strip . The system use s ceramic wick, thermocouple psychrometers of the basic design of Laurence an d Pruitt (1969), modified by Gay (1972a) for measurement of temperature an d vapor pressure gradients . Wind speed measurements are made with Thornthwait e cup anemometers that use a photochopper circuit to generate a pulse output . The shortwave radiation components were measured with Kipp pyranometers, an d the allwave components with Lange pyrradiometers of Schultzes design . The system does not at this time include an on-line computer to facili - tate data collection and analysis . The data is processed through the OS U Computer Center, using a variety of programs described by Gay (1973, Ch .3 ) before analysis is completed with the appropriate meteorological models .

5 .2 Site Characteristics .

The meteorological study site is in the broad, flat valley of the Ceda r River, east of Seattle . The stand and soils characteristics at Cedar Rive r have been described by Fritschen (1972) . The Douglas-fir forest is about 3 5 years old, and originated as natural regeneration following logging . The density averages 570 trees per hectare, and the average spacing is about 5 . 7 in . The soil in a Barneston, gravelly, loamy sand originating from a glacia l outwash ; the root system is restricted to the upper metre of soil . The mean dbh is about 16 cm . and the average height of the stand is 28 in . The canopy is relatively level, with a uniform density . The general level of canop y a.losure was estimated to be about 22 in .

5 .3 Field Measurements .

The same basic pattern of measurements was followed during both summers . All sensors were sequentially recorded at intervals of five minutes durin g the day, and 10 minutes at night . The system was in operation at Cedar Rive r for 6 days in 1971, and for 20 days in 1972 . The observations took plac e a at the end of July and in early August during both years . The instrumen t configuration differed somewhat between the two field seasons, however . In 1971, a single "crankup" instrument package was used above th e forest . The sensors were mounted on a trolley that ran vertically up a 33 . 5 m (110 ft) tall, triangular TV tower about 0 .3 m (1 foot) in width . Thi s tower was adjacent to the lysimeter tree . The instrument configuration i s shown in Figure 1 . Wind, temperature and vapor pressure measurements wer e made at six levels, one meter apart, beginning at 26 .16 m, and ending a t 31 .16 m above the forest floor . The radiation budget components were measure d from a height of 30 .66 m above the floor . The tip of the tallest tree in the vicinity of the tower extended to 29 in, though the bulk of the crowns wer e below 24 in, and the general level of crown closure was about 20 to 22 m abov e the floor . Five soil heat flux discs were installed nearby, 2 cm below the top of the litter layer and just beneath the surface of mineral soil on th e forest floor . During the 1972 field season, a new instrument configuration (Figure 2 ) was used . Six- psychrometers and anemometers were mounted at lam intervals 3 Figure 1 . The Cedar River site in 1971 . The psychrometers are attached t o a track-mounted trolley that runs up the south face of the tower . The psychrometers are 1 m apart ; the lowest one is 26 .16 m above the fores t floor .

Figure 2 . The Cedar River site in 1972 . The sensors inthe suspended packages are 1 m apart . A . Tower 1 from the south, with the 1971 towe r in the right of the picture . The lowest psychrometer is 25 m above th e floor . B . Tower 2 from the southeast . The prevailing wind is from the west and northwest . The lowest psychrometer is at 22 m . - FigurFiguree 3 . The instrument hoist . An electric, geared winch wasused to hoist the instrument package during the 1972 field season .

Figure 4 . The suspended instrument package . A . The instrument packag e contains sensors to measure temperature, vapor concentration and win d speed at six levels, each 1 m apart . The central shaft serves as a common duct for aspiration of the psychrometers, and ad a mount fo r both psychrometers and anemometers . B . The instrument package sus - pended from tower 1, with the top psychrometer at 34 .m. The top of the 1971 tower is visible in the middle background .

-1616 -

on a vertical support mast 6 m in length . The hollow, rectangular suppor t mast served as a common aspiration duct for the psychrometers, which wer e ventilated by a single blower mounted at the lower end . This instrument package was suspended above the forest from a lightweight aluminum trus s which held the package approximately 4 m away from the face of the supportin g tower . Two instrument packages were used . One was suspended from a ne w tower, 37 m high, erected adjacent to the lysimeter tree (tower 1) and th e second from an identical tower (tower 2) located about 100 m away . The instrument packages were oriented with respect to the up- and down-valle y winds that prevailed at this site . The suspended package could be raise d or lowered with an electric winch for servicing as desired . Details ar e shown in Figures 3 and 4 . Pyranometers and pyrradiometers were mounted o n a boom extending 3 m away from the south face of each tower at the 36 m level . Soil flux discs and soil temperature elements were placed in th e ground near the base of each tower . In addition, the temperature and humidity of the air was measured near the base of tower 2 with two psychrometers that were one meter above ground level . The recording system worked reasonably well during the field measurement periods and a range of weathew conditions was sampled . Some of thes e results will be examined next .

I 1 , . et, ,_ I 1 I

- 17 -

6 . RESULTS AND DISCUSSIO N

The rather limited results now available from the measurements made a t the Cedar River site can be contrasted with results reported by some of th e studies reviewed in Section 4 . Some initial comparisons can also be mad e between the meteorological models used to estimate the components of th e forest energy budget .

6 .1 Cedar River Results .

The totals of the energy budget components at Cedar River are presente d in Table 1 for one overcast and two clear days . These are the condition s that provide the greatest contrasts in the energy budget . Measurements made under such contrasting conditions should improve understanding of the basi c mechanisms governing energy transfer between the forest and the atmosphere .

Table 1 . Energy Budget Components at Cedar River, Washington . Daily total s are subdivided into periods of daylight (0630-1930 hours) and night . Units are cal/c m2 . The 1972 H and XE fluxes are the mean from two towers .

31 July 71-overcas t 29 July 71-clear 10 August 72-clear day night total day night total day night tota l

R 141 - 7 454 -44 454 -55 134 410 39 9

G - 7 13 - 43 48 - 27 29 6 5 2

H -38 - 1 -143 8 -135 -203 42 - 39 -161

XE -96 - 6 -263 -17 -225 -15 -102 -280 -240

The totals in Table 1 are subdivided into daylight (0630-1930 hours , P .D .T .) and night periods . Such a separation facilitates future comparisons , here and elsewhere, among energy budget components measured on days wit h different lengths of daylight . The daily totals for the two clear days are quite similar, although the y were obtained in two different years . The contrasts that Gay (1972) identi - fied between the energy budgets of the clear and the overcast weather wil l therefore still apply . The sky conditions had no affect on the forest albedo , which remained at 9 percent on all three days . The diurnal trends in the energy budget components are shown for clea r weather in Figure 5 . An hourly plot for overcast day is given in Figure 6 . The major difference apparent between the two clear days are related t o phase shifts among the H and AE fluxes . Values of Q and G are similar o n u

4 6

E O.8 f N 2 0.6 v <0.4 v Q2 c0r W z o W -0.2 0 2 4 6

Figure 5 . Energy budget components under clear skies . A . 197 1 measurements . Gay (1972b) . B . 1972 measurements, which ar e the mean of two towers .

Figure 6 . Energy budget components under overcast skies . Gay (1972b) .

both days . However, in 1971, the sensible heat flux reached its maximu m about an hour before solar noon, while latent energy flux peaked about tw o hours after solar noon (Figure 5A) . These 1971 results match well with those obtained over young Douglas-fir by Black and McNaughton (1971) . The explanation proposed for such a phase shift by Black and McNaughton (1972 ) and by Stewart and Thom (1973) is related to the internal and externa l transfer resistances of coniferous needles . The interplay between the two resistances results in the latent energy flux being driven by the vapo r pressure deficit of the atmosphere, which generally peaks in mid-afternoon , rather than by the net radiation which peaks near solar noon . Such a concept has proven useful in interpreting apparently anomalous energy budge t results . The contrasts between the 1971 and 1972 clear weather measurement s at Cedar River are not yet fully explained, however . Stewart and Thom (1973) discuss the resistance concept in great detail , and apply the theory to their measurements made over Scots pine . Other applications of this concept to forests are based upon the measurements o f Tajchman over Norway spruce near Munich (Szeicz, Endr8di and Tajchman, 1969) . Preliminary calculations of resistances at Cedar River have been made fo r the 1972 measurements . Resistance calculations can not be made for 197 1 because of the variable quality of the wind profiles . Considerable effor t must still be spend in interpreting and reconciling our results ; these wil l be described in a subsequent report .

6 .2 Comparison with Other ForestEnergy Budgets .

The Cedar River energy budgets can be compared with the results of a young Douglas-fir plantation (Black and McNaughton, 1971) and a young Scot s pine plantation (Tajchman, 1972b), along with additional unpublished result s from the mature Scots pine plantation described by Stewart and Thom (1973) . These studies used good instrumentation, and analyzed the data with the Bowe n ratio . Daily sums of the energy fluxes from all four sites are tabulated i n Table 2 .

Table 2 . Energy Budget Totals (cal/cm 2 24-hours) at Four Coniferou s Forest Sites . Clear Weather .

Stand G AE H Date

2 .8 m Scots pin e 364 - 9 -163 -193 July 28, 197 0 (Tajchman, 1972)

15 .8 m Scots pine 2 471 -23 -167 -281 July 7, 197 1 Thetford, Norfolk

7 .8 m Douglas-fi r 379 0 -273 -106 July 14, 197 0 (McNaughton Black, 1973)

28 m Douglas-fir 3 399 2 -240 -161 Aug . 10, 1972 Cedar River, Washington

2Gay, L . W . and J . B . Stewart . Energy budget studies in coniferous forests . Unpublished paper presented to Swedish Coniferous Biome Seminar, JadraRs , Sweden, May 17, 1973 . Mimeo ., 24 pp .

3The latent and sensible fluxes are mean from two towers .

Energy budget results of the type presented in Table 2 have not ye t become available often enough to permit inferences to be drawn regardin g the. behavior of different forest species . A thorough interpretation of th e energy budget components requires some supplementary information on . the availability of water to the four experimental sites . Such data have no t been reported for the few forest energy budgets that have been completed . However, in the examples available, Douglas-fir demonstrates a tendency t o transpire at a somewhat higher rate than Scots pine . The process studie s of Stewart and Thom (1973) and McNaughton and Black (1973) show evidenc e that Douglas-fir has lower stomatal resistances than does the more xeri c Scots pine . Process studies of this type will contribute directly t o interpretation of energy budget results from different forest communities .

6 .3 Comparison of Methods .

Few of the published studies were suitable for testing accuracy of th e various methods of estimating the energy balance components . Most of the studies cited used the Bowen ratio technique, which has become widely accepte d as a standard for comparison .

6 .3 .1 Comparisons Between Meteorological Methods .

One study (Tajchman, 1973) reports good agreement between estimate s of the sensible heat flux by the Bowen ratio ( HB) and the aerodynamic metho d (Ha) over the young Scots pine plantation . The 2 .8 m plantation has a roughnes s length zo = 0 .26m = 0 .09 h, and an "effective" displacement of D = 2 m = 0 .7 1 h . The percentage difference between the two meteorological models (expresse d as (HB - Ha)/(HB x 100) was -0 .4%, 4 .1%, -0 .8% and 18 .7% over the period 0800 - 1600 hours on each of four days . The day with the largest discrepancy was characterized by light and variable winds, conditions under which the aero - dynamic method might be expected to fail . Stewart and Thom (1973), in contrast, concluded that the aerodynami c method was inappropriate in the vicinity of aerodynamically rough surfaces , such as their 15 .8 m high Scots pine plantation (D = 12 m = 0 .75 h ; zo = 0 .93 m = 0 .06 h) . Their aerodynamic fluxes Ma + Ha) consistently amounte d to only 40% of the available energy (Q + G) during daytime periods wit h Ri < O . They found no such discrepancy during the stable conditions (Ri > 0 ) present at night . Black and McNaughton (1972) examined the consistancy of several form s of the Bowen ratio, and tested the aerodynamic method as well . Their calculations showed X Ea to be 34 percent greater than AEB . However, th e aerodynamic estimates were derived from synthetic profiles, since th e original experiment took measurements at only one level for wind speed an d two levels for temperature and vapor . The calculations were also base d upon the values of D = 4 .8 m = 0 .62 h, and zo = 0 .82 m = 0 .11 h, as estimated by formulas developed elsewhere . Their comparisons are helpful in elucidatin g the exchange processes, but the results are not definitive tests of methods . Moore and Hicks (1973) did not compare their eddy correlation measurement s against another method, but they did systematically check consistancy with a technique that would also be applicable to aerodynamic analyses, although no t to the Bowen ratio model . Moore and Hicks defined a "recovery rati o " as (H + AE)/(Q + G) ; a similar concept was mentioned by Stewart and Thom (1973) . The eddy correlation recovery ratio during the daytime was about 1 .07, with a

- 21 - Standard error of 0 .09, for nine half-hour- runs with suitable fetch . This gives one confidence in their -regnlts . Sirtiildr tests would be desirable fo r the analyses that yield aerodynamic estimates .

6 .3 .2 Comparisons With Direct Measurements .

Several types of direc t, measurements can be used to check the at estimates derived from meteorological data . Storr, et al . (l970) compared their Bowen ratio results to those of a short-term water balance for thei r experimental basin and obtained good agreement . Meteorological estimate s of evaporation were 89 mm of water equivalent= ; the water balance est at e was 79 nun fot the 19 day period . The August 10, 1972, Boren ratio estimates at Cedar River averaged -24 0 cal/cnf2 24-hours for XE . This can be compared to the estimat e4 of -275 car/c m2 24-hours obtained by direct tfieasuremett of water loss from a lysimeter tre e adjacent to tower 1 . The weighing lysimeter contains a 28-m Douglas-fit tree ; details of its construction dhd use have been given by Fritschen, et al . (1973) and by Fritschen and Dotiswamy (1973) . The agreement between the two method s appears reasonable .

6 .2 .3 Tests by Replication .

The 1972 Bowen ratio results at Cedar River can also be examined _ with respect to the consistahcy of the estimates made at the two towers . The phase of the two latent energy estimates is similar throughout the day . The totals, however, were :

At-1 = -196 cal/cm2 day, and XE-2 = -285 cal/cm 2 day .

The variation between the two estimates is large, with a range of 18 percen t above and below the mean of 240 cal/cm 2 day . The cause of this discrepanc y is not obvious . It may well be associated with differences in the site, in positioning, or with instrumentation limitations . Despite obvious limitations on interpretation of results from only tw o towers, no other forest energy budget experiments, here or abroad, have reporte d results obtained from more than one location above the canopy . Studies of such variability appear essential for improving the experimental design and instrumentation to be used in future studies . The variability reported here clearl y dictates that some form of replication must be included in forest energy budge t studies . Several aerodynamic analyses have also been attempted with the forest dat a from the two towers . The results have not generally been satisfactory ; a genera l discussion of these analyses will be deferred until a later paper . The work wit h the aerodynamic models, even though incomplete at this time, and the comparison s between the lysimetric and Bowen ratio results have stimulated us to reexamin e the problems associated with forest energy budget analyses . We Shall therefor e discuss in the next section the problems apparent in application, based upo n our experience and that of, other researchers who have examined various aspect s of energy budget research .

4Fritschen, Leo J . Personal communication, December 1973 . 7 . PRACTICAL DIFFICULTIES IN ESTIMATING FOREST EVAPOTRANSPIRATIO N

Some problems in application have already been touched upon durin g our presentation of restrictions, recent work, and results . Applicatio n problems in agriculture have received much attention (for example, se e Tanner, 1967), and some aspects have been considered for forest s (Federer, 1970) . With the accumulation of experience, both at our sit e and elsewhere, there has developed an increased appreciation of practica l difficulties in the measurements of forest evapotranspiration . We shall attempt to summarize those that are particularly pertinent to the use o f micrometeorological and lysimetric techniques .

7 .1 Problems of Dimension .

The need for appropriate descriptions of forest structure has lon g been recognized in energy budget research (Miller, 1965), However , stand height remains the most commonly presented, stand characteristic . Leonard and Federer (1973) attempted to relate topographical measurement s of a rough forest canopy to the development of profiles in the boundar y layer . Their results confirm that mean stand height does not appear t o provide a suitable description of such variables as crown shape, stan d density, and variations in ground topography, Rauner and Ananjev (1971 ) have normalized profile measurements within a number of forest specie s with respect to the vertical distribution of the surface area of th e phytomass . Further developments in this area will be welcome .

7 .1 .1 Canopy Dimension .

Selection of the proper location for instruments within the boundary layer is difficult because of the diffuse nature of the exchangin g surfaces within the canopy . Tanner (1968) and McBean (1968) comment upo n problems of spatial variability, Federer (1968) and Droppo and Hamilto n (1973), however, report little difficulty due to spatial (horizontal ) variability of net radiation above forests, Droppo and Hamilton, in addition, report that statistical differences existed between three separat e profiles within the canopy of a deciduous forest of 18 m height, even thoug h the three towers were only 15 m apart . The results of our Bowen ratio estimates made at two locations abov e the forest were presented earlier ; the comparisons were not encouraging . The heights used at these two towers illustrate the problems of dimension . • Tower 1 estimates were based upon psychrometers at the 25, 28 and 29 m levels ; tower 2 levels were 25, 26 and 27 in, referenced to the base of th e appropriate tower . The mean level of the canopy is rather uniform ove r the study are, but the ground rises approximately 3 meters between the tw o towers, so the lowest psychrometer at tower 1 is actually 3 m nearer th e crowns than is the lowest level used at tower 2 . This may place the measurement levels at tower 1 below the boundary layer and thus contribute to th e differences found at the two towers .

7 .1 .2 Aerodynamic Dimensions .

The problems of deriving precise measurements of a zero plan e displacement and roughness length have certainly not been solved . A common

-23 -

approach used over forests and other tall vegetation is to derive ggaphical estimates of these two parameters for periods of neutral stability . Stewart and Thom (1973) found that such estimates of D yielded a value of 12 m, or 0 .76 of the height of their stand . The value of d = D + zo was 12 .93 m, and they assumed that these values did ibt change the wind speed . Leonard an d Federer (1973) reviewed results from six forests and concluded that the relationship between D, z o , and wind seed was not well defined . Szeic2 , Endr8di and Tajchman (1969), Rauner and Ananjev (1971) and Konstantinov (1963, p .163), however, present data from studies over a number of forest s that shows D decreasing and Ze increasing with windspeed, with the ne t results that D + zo decreased with ndtpeed . Since the aerodynamic equations require appropriate values of the affective displacement height , it is clear that progress must be made in determining the constancy of thi s parameter .

7 .2 Problems of Instrumentation .

The problems involved in creating a suitable recorder/sensor syste m for agriculture are described by Tanner (1963) . This is a major topi c that can best be handled by reference to primary sources . However, it i s evident that the gradients above forests are an order of magnitude les s than those found over low vegetation with relatively denser canopies . In particular, temperature and vapor concentration sensors above forests mus t resolve differences approaching a feta hundredths of either a degree, or a millibar, per meter . Some studies have increased the distances between the measuremen t levels in order to increase the profile differences to a magnitude commensurat e with the sensitivity of the recorder/sensor system (for example : Galoux et al ., 1967 ; Reimer and Desmarais, 1973 ; Storr et al ., 1970) . Consideration must then be paid to the position of the instruments within the surfac e boundary layer, and to application of appropriate corrections to the measure d temperatures in order to obtain the desired, potential gradients of dry - and wet-bulb temperature . Small gradients above the forest Were successfully measuredby Blac k and McNaughton (1971) with a differential psychrometer that interchanged sensors bettaeen readings . The dynamic response of a sensor to transients that affect th e atmosphere variables above the forest may also introduce errors . ?4oor e and Hicks (1973) increased their eddy flux estimates above a pine .forest by 11 percent to account for the high frequency components that were no t picked up by their "fluxatron " . In yet another example, the dynami c effects of wind on lysimetric measurements of evapotranspiration has been found to be greater in forestry than in agriculture . During fairidy periods , only daily evapotranspiration values can be resolved from the Cedar Rive r lysimeter because of the noise incorporated into the output of the weighin g lysimeter .

7 .3 Difficulties in Analysis .

By far the greatest amount of forest energy budget analysis has bee n based oil application of the Bowen ratio . It is becoming evident tha t % Wed ratios derived from measurements at only two levels are subject to

biases among sensors, and . to the influence of surface homogeneities . . Much better results can be obtained if temperature and vapor concentratio n measurements are made at many levels . Tanner (1963) mentioned the advan- tage of plotting measurements of temperature against vapor concentratio n to insure that conditions of similarity are met . Gay (1972) reported goo d results in identifying malfunctioning psychrometers above .the forest_wit h this technique .

GA A . 29 JULY 02 HOUR (PDT) - - 0 1 2 3 4 567,11111 ])j 17,192021222324 _ 8 9 10 1112 13 14 15 16 18 -1 \

-0.2

-0404 - MH VAPOR PRESSURE, M B

Figure 7 . Similarity between the gradients of potential temperatur e and vapor pressure . A . Clear .weather on July 29, 1971 . Levels 1, 4 and 6 . B . Overcast weather .•on July 31, 1971 . Levels 1, 2 and 3 . Gay (1972b) .

As an example of this graphical approach, the similarity plots ko t July 29, 1971, are shown for levels 1, .4, 6 in Figure 7A, and for levels 1, 2, 3 on July 31 in Figure 7B . The potential temperature gradient (0 ) is on the ordinate, and the vapor pressure gradient (e) scale is shown i n 3B-. Since the plot for each hour has been normalized by subtracting th e 0 and e value at the bottom level from the observations at each•of th e other levels, the plots are actually the increments DO and ie taken wit h respect to the measurement at the bottom level, nearest the canopy . Though several points can be deduced from such similarity plots, th e most important conclusion concerns adequacy of data . The linearity a t the selected levels confirms their acceptability for the Bowen ratio model , although an unexplained offset is evident at level 2 during the 1000-120 0 hours on July 31 . Once the data is judged acceptable, one notes that th e slope of the lines is Lx0/txe ; this is directly proportional to 13, the Bowe n ratio, as shown in Equation (3) . Thus the relative slope of the similarit y plots is an index to the way that the surface is partitioning the net energ y supply into convection and evaporation . Further, the quadrant of each hourl y plot-indicates the sign of S, and the direction of the H and XE fluxes . The similarity plots can thus provide three things : a ready indication of the magnitude of the Bowen ratio ; an indication of the direction of the H and E fluxes ; and an identification of levels suited for the Bowen ratio analysis . Identification and elimination of malfunctioning psychrometers is only the first step, however . The Thetford projec t 5 now fits a linear regressio n of potential temperature on specific humidity, using data from 6 or 8 levels, to obtain standard errors for their estimates of the Bowen ratio . The 1972 energy budgets at Cedar River were based upon such estimate s using three or four levels of measurement . The reversing psychrometer s of Black and McNaughton (1971) should also provide improved accuracy in the determination of the Bowen ratio . Either of these techniques should be considered in the planning of future forest experiments .

ersonal communication . J . B . Stewart, Institute of Hydrology, Wallingford , Berkshire, U .K ., 1973 .

-26 -

8 . CONCLUDING REN1RK S

A number of , limitations have been identified in applying mean-profil e energy budge t , analyses to forests . These limitations are primarily associated with the diffuse and aerodynamically rough structure of forest canopies, whic h are in contrast with the dense and relatively smooth canopies of agricultura l crops . The criteria for suitable sites appear to become more stringent as th e height of the vegetation increases . The horizontal extent of an experimental site limits the inasclmum height at which the sensors can be placed . The minimum sensor height is affected by the uniformity and structure of th e canopy . , The mere height of forest vegetation can create mechanical difficultie s in positioning instruments . Support towers must be sturdy and safe to climb , yet they must not unduly disturb the environment . There are often practica l problems in orientating instruments . It is also difficult to ensure that the properties T, e and u are measured at the same effective level in the atmosphere . Psychrometers ordinarily measure T and e in the same air stream, bu t anemometers are necessarily placed some distance away . Tanner (1963) suggest s grap=hical tests for checking the assumption that all three properties ar e indeed being measured at the same level . These tests were found useful by Gay (1972) in analyzing forest measurements . _ In addition to these mechanical difficulties, the structure of the fores t canopy creates problems that are not apparent in energy budget studies ove r low, dense vegetation . Needle elements in coniferous canopies are onl y sparsely distributed throughout a considerable depth ; the canopy is thu s a diffuse source of sensible and latent heat . Surface roughness furthe r enhances mechanical mixing in the atmosphere near the forest canopy atd a s a result, the gradients of T and e are in the neighborhood of a few hundredths (°C or mb) per metre . This is an order of magnitude lower than the gradients encountered over low, dense vegetation, and there is a corresponding increase in the difficulties of obtaining the require d measurement precision . The effective upward displacement (D) of the profiles must also b e determined for the application of the aerodynamic equations (5a, 5b) t o forests . Normally it Is assumed that the same D applies equally to th e profiles of T, e and u . However, the sources and sinks of sensible an d latent heat, and momentum, may well occur at different levels in the considerable depth of a diffuse coniferous canopy . This would require that an appropriate displacement height be determined for each profile . In any case , the application of the theory in t-he field has not consistently yielde d reasonable values for displacement heights in tall vegetation . The fetch, or distance through which the wind travels ovet the vegetatio n surface under study before reaching the sensors, must be-sufficient to allow the development of an equilibrium layer in which the profiles of T, e and u are affected primarily by the experimental surface . A long fetch effectivel y eliminates horizontal advection, and assures that the fluxes are vertical . The relationships proposed in agricultural studies give ratios of fetch t o maximum instrument height (presumably measured above the effective displacemen t height, D) in the range of 100-500 to 1 . If this relationship is extrapolate d to forests, then a fetch of tens of kilometers appears needed . It is eviden t that the forest energy budget studies reported to date cannot meet such stringent fetch requirements . It seems likely that the profile equilibrium sill be

achieved more rapidly over atrough forest a rface than over lower., smoothe r vegetation (Pasquill, 1972 ; Webb, 1965) ; such an effect would .ee.se tae etc h requirements considerably . The uniformity and structure .of •the canopy affects the minimum hetiught at which the sensors -can lbe laced . The lowest sensor must be within th e equilibrium layer, and snot Abe unduly Influenced .by local features of th e canopy . Lettau a (1959) rule of •thunlb dot- •smooth surfaces and low .crops suggests that the lower •sensoar level should be at least five times-the height orf the average roughness element .(zo) . Presumably, in tall vegeta- w. tion the distance would be measured above the effective displacement height , D . Again, definitive est4maite;s . acre lacking for tall, rough vegetation- . The higher the instruments above the canopy, •thesmalier are the gradient quantities, and -the more difficult It becomes rto make the measurements . - A survey of forest me °bi dgeat --st;udies reveals that published .results are sparse . The quality of -the work imported appears variable, as many of the limitations on equipment and techniques are just becoming apparent . The information available in these publications is useful in providin g an insight into the magnitudes of the components of the forest energy :budget . Only a few (for examples, see Stewart and Thom, 1973 ; Black and McNaughton , 1972) examine the mechanisms that control the -exchange of energy . -Studie s and evaluations of these complex mechanisms should come soon, as more information is collected in and above the forest . Replication of the meteorological measurements at two points above the Douglas-fir site at Cedar River, Washington, revealed rather poor agreemen t for the Bowen ratio model . The individual latent flux estimates marled ±I8 percent about the mean of 240 cal/cm 2 day . The highest flux (-285 .cal/cm2 day) compared well with the .water loss from the lysimeter tree (-275 -cal/c-m2 day),-however . The discrepancy between two sites which were separat db y about 100 m confirms the necessity for replicating measurements of the energ y balance of forests . The data evaluated here showed that the young Douglas-fir forest transformed•about 60 percent of the net radiation into latent energy under clea r skies . The ratio was higher, nearly 75 percent, for the overcast day . These values are in reasonal agreement with Black and McNaughtons (1972) results i n British Columbia . Such agreement may be partly fortuitous, as t-here was no information available on the water status of these two experimental sites . The exchange of energy depends upon the physiological response of th e forest, as well as upon the amount of energy made available to the canopy . Information on the status of the vegetation needs to be incorporated !int o future studies . We expect that such information will be given in futur e studies of the forest energy balance, now that the basic procedures and applications have been proven for experimental vork . LITERATURE CITED

Allen, L . H . 1970 . An operational system for (1) sampling and sensin g micrometeorological elements, and (2) logging and processin g micrometeorological data . In : Powell, J, M ., and C . F . Nolasco • (Eds .), Proceed ., Third Forest Microclimate Symposium, Can . Forestry Service, Calgary . Pp, 91-116, March 1970 .

Backlund, B . and K . Perttu . 1971 . System for data logging at shor t intervals and processing of data about plant growth and climate . Royal College of Forestry, Stockholm . 47 pp .

Baumgartner, A . 1956 . Investigations on the heat- and water economy o f a young forest . Translated by E . Pichler, from : Untersuchunge n {fiber den Wdrme-- and Wasserhaushalt eines jungen Waldes . Ber . Deut . Wetterdienst 5 :4-53 . CSIRO Translation 3760, Melbourne , Australia . 1958 .

Baumgartner, A . 1971 . Wald als Austauschfaktor in der Grenzschich t Erde/AtmosphRre . Forstwiss . Centralblatt 90 :174-182 .

Black, T . A . and K . G . McNaughton . 1971 . Psychrometric apparatus for Bowen-ratio determination over forests . Bound . Layer Meteor . 2 :246-254 .

Black, T . A . and K . G . McNaughton . 1972 . Average Bowen-ratio method s of calculating evapotranspiration applied to a Douglas-fir forest . Bound . Layer Meteor . 2 :466-475 .

Busch, N . E . 1973 . The surface boundary layer (Part 1) . Bound . Layer Meteor . 4 :213-240 .

Byrne, G . F . 1970 . Data logging and scanning rate considerations i n micrometeorological experiments . Agric . Meteor . 7 :415-418 .

Clayton, W . H . and J . D . Merryman . 1960 . Design of the mobil e micrometeorological stations employed on Project Greenglow . Final Report AF19 (604)-04562 . Dept . Oceanog . Meteor ., A M College of Texas, College Station, Jan . 1960 . 83 pp .

Denmead, O . T . 1964 . Evaporation sources and apparent diffusivities in a forest canopy . J . Appl . Meteor . 3 :383-389 .

Denmead, O . T . 1969 . Comparative micrometeorology of a wheat field an d a forest of Pinus radiata . Agric . Meteor . 6 :357-371 .

de Vries, D . A . 1963 . The physics of plant environments . In : Evans , L . T . (Eds .) . Environmental control of plant growth . Pp . 5-22 . Academic Press, N . Y .

- 29 - Droppo, J . G ., Jr ., and H. L ., Hami1t a,. Jr .. 1973 . Experimental vari4bAlity in the determination of the energy balance of a deciduous forest ., J . Appl, Meteor . 12 :781-791 .

Dyer, A . J . 1963 . The adjustment of profiles. and eddy fluxes . Quart . J . Roy . Meteor . Soc . 89 :276-280 .

Dyer, A . J . 1967 . The turbulent transport of heat and water vapor in an unstable atmosphere . Quart . J . Roy . Meteor . Soc . 93 :501-50$ .

Deerdzeevski, B . L . 1963 . Study of the heat balance of the forest . Silva fenn . 113 :3-17 .

Federer, C . A . 1968 . Spatial variation of net radiation, albedo an d surface temperature of forests . J . Appl . Meteor . 7 :789-795 .

Federer, C . A . 1970 . Measuring forest evapotranspiration - theor y and problems . USDA Forest Service Research Paper NE-165 . Upper Darby, PA . 25 pp .

Fedrov, S . F . 1970 . The data from experimental studies of the elements: of the water and heat balances on small forested watersheds . Reports submitted by Soviet Scientists to the International Symposium on th e Influence of Forest on Environment . Pp . 200-215 . FAO Forest Mgt . and Forest Ind . Dept ., State Comm . Forestry and Council Minister s USSR, Moscow, 1970 .

Fritsehen, L . J . 1970 . Particular problems of instrument design an d overall development on systems for agricultural and climatologica l stations . Meteor . Monogr . 11 :346-351 .

Pritschen, L . J . 1972 . The lysimeter installation on the Cedar Rive r Watershed . In : J . F . Franklin, L . J . Dempster, and R . H . Waring (eds .) . Proceedings - research on coniferous forest ecosystems - a symposium . P . 255-260 . Pac . Northwest Forest and Range Exp . Sta ., Portland, Oreg .

4 Fritschen ,, I . J ., L . Cox, and R . Kinerson . A 28-meter Douglas-fir in- a weighing lysimeter . For . Sci . 15 :256-261.

Fritsc-hen, L . J . and P . Doraiswamy . 1973 . Dew : an addition to the hydrologic balance of Douglas-fir . Water Resources Res . 9 :91-8g4 .

Fritschen, L . J . and C . H . M . van EaveI . 1963 . Micromettorolo"gica 1 data handling system . J . Appl . Meteor . 2 :151-155 .

Gay, L . W . 1972x . On the consiriuction and use of-ceramic-wick psychro - meters . In : R . W . Brown and B . P . Van. Haveren (eds .), Psychro?metry in water relations research . Pp . 2251-258 . Utah Agric . Exp . gta . , Logan . 338 pp . Gay, L . W . 1972b . Energy flux studies in a coniferous forest ecosystem . In : J . F . Franklin, L . J . Dempster, and R . H . Waring (eds .) , Proceedings - research on, coniferous forest . ecosystems - a symposium , p . 24 3 L253, illus :~ Pac . -Northwest Forest , and Range Exp . Sta . , Portland, Oreg .

Gay, L . W . 1973 . Energy exchange studies at the earths surface . I . Energy budgets of desert, meadow, forest and marsh sites . OSU Dept . Atmos . Sciences Tech . Rep . No . 73-1 . Jan . 1973 . Corvallis .

Gay, L . W ., and H . R . Holbo . 1970 . A counter and interface for use wit h digital data acquisition systems . Proceedings : Conference on electronic instrumentation . Pp . .91-115 . Depts . Phys . and Elec . Engr ., Univ . Idaho, Moscow .

Galoux, A . 1968 . Flux denergie et cycles de matieres en tant qu e processus ecologigues . 1Association Nationale des Professeurs d e Biologie de Belgique 14 :167-202 .

Galoux, A . 1971 . Flux et transferts denergie au niveau des ecosysteme s forestiers . Pp . 21-40, Productivite des ecosystemes forestiers , Actes Coll . Bruxelles, 1969 . (Ecologie et Conservation, 4) . UNESCO .

Galoux, A ., G . Schnock and J . Grulois . 1967 . Les installations eco- climatologiques .. Station de Recherches des Eaux et Forets, Travaux - Serie A, No . 12 . Groenendaal-Hoeilaart, Belgium . 52 pp .

Grulois, J . 1968 . Flux thermiques et evapotranspiration au cours dun e journee sereine . Bull . Soc . roy . Botanique Belgique 102 :27-41 .

Kiese, O . 1971 . S . The measurement of climatic elements which determin e production in various plant stands . Methods and preliminary results . In : Ellenberg, H . (ed .) Ecological Studies, Analysis and Synthesis , Vol . 2, pp . 132-142 . Springer-Verlag .

. Kiese, O . 1972 . Bestandsmeteorologische Untersuchungen zur Bestimmun g des Wdrmehaushalts eines Buchenwaldes . Ber . Num . 6 . Inst . f . Meteorologie and Klimatologie, Tech . Universitdt, Herrenhauser Str . 2 , Hanover .

Konstantinov, A . R . 1963 . Evaporation in Nature . Transl . from : Isparenie v prirode . Gidrometeorologichesbol izdate l stvo, Lenningrad, 1963 . IPST Cat . No . l529 . U .S . Dept . Commerce, Clearinghouse for Federal Scientific and Technical Information, Springfield, VA .22151 . 1966 . 523 pp .

Leonard, R . E . and C . A . Federer . 1973 . -Estimated and measured roughnes s parameters for a pine forest . J . Appl . Meteor . 12 :302-307 .

Lettau, H . 1959 . A review of research problems in micrometeorology . Dept . Meteorology, Univ . Wis . Final Rep ., Contract DA-36.-039-SC-80063 . Lourence, F . J . and W . 0 . Pruitt, 1969 . A psychrometer system for micro - meteorology profile determination . J . Appl . Meteor . 8 :492-498 .

Lumley, J . L . and H . A . Panofsky . 1964 . The Structure of Atmospheri c Turbulence . Vol . 12, Interscience Monographs and Texts in Physic s and Astronomy . Wiley, New York . 239 pp .

Laze, R . 1966 . Vergleichende Energie umsatzmessungen im Walde and au f einer Wiese . Archie Forstwesen 15 :995-1015 .

Ltitze, R . 1969 . Wasserverbrauch and Energieumsatz eines Kiefernbestandes and einer Wiese . Archie Forstwesen 18 :921-927 .

McBean, G . A . 1968 . An investigation of turbulence within the forests . J . Appl . Meteor . 7 :410-416 .

McNaughton, K . G . and T . A . Black . 1973 . A study of evapotranspiratio n from a Douglas fir forest using the energy balance approach . Water Resources Res . 9 :1579-1590 .

Miller, D . H . 1965 . The heat and water budget of the earths surface . Advan . Geophys . 11 :176-277 .

Monteith, J . L . 1965 . Evaporation and environment . Symp . Soc . Exp . Biol . 19 :205-234 .

Moore, C . J . and B . B . Hicks . 1973 . The heat budget of a pine forest . Proceedings : First Australian Conference on Heat and Mass Transfer . Section 3, pp . 58-64 . Monash Univ ., Melbourne, Australia . May 23-25, 1973 .

Oliver, H . R . 1971 . Wind profiles in and above a forest canopy . Quart . J . Roy . Meteor . Soc . 97 :548-553 .

Pasquill, F . 1972 . Some aspects of boundary layer description . Quart . J . Roy . Meteor . Soc . 98 :469-495 .

Paulson, C . A . 1970 . The mathematical representation of wind speed an d temperature profiles in the unstable atmospheric surface layer . J . Appl . Meteor . 9 :857-861 .

Penman, H . L ., D . E . Angus and C . H . M . Van Bavel . 1967 . Microclimati c factors affecting evaporation and transpiration . In : Hagen, R. M . , H . R . Haise and T . W . Edminster (eds .) . Irrigation of Agricultural Lands . Agronomy 11 :483-505 .

Rauner, Yu . L . 1958 . Thermal-balance observations in a deciduous forest . Izv . Akad . Nauk SSSR (Ser . Geogr .) 5 :79-86 . Translated from Russian , IPST Cat . No . 1608, U .S . Dept . Commerce, 1966 .

Rauner, Yu . L . 1962 . The heat balance of forest and its role in th e formation of the microclimate of wooded and treeless landscapes o f the Moscow region . Soviet Geography (English transl .) 3 :40-47 .

J - 11 f l 17

isell _1 .11 1 1

Rauner, Yu . L . 1966 . Methodik and einige Ergebnisse der experimenteller Erforschung des WHrmehaushalts der WaldbestHnde ., Angew . Meteorol . 4 _ 5 :157-165 .

Rauner, Yu . L . and I . P . Ananjev . 1971 . Merkmale der Atmosphdrische n Turbulenz in Waldbestdnde . Translated from : Charakteristik i atmosfernoj turbulentnosti v uslovijach lesa . Izv . Akad . Nauk SSSR (Ser . Geogr .) 2 :70-79 . Ber . Eidg . Anstalt foretl . Versuchswesen (Birmensdorf, CH), Num . 98, 1973 . 23 pp .

Reifsnyder, W . E . 1963 . Techniques of measuring, recording and prd essin g micrometeorological variables in forest energy-budget studies . Symposia on Water Balance of the Soil and Forest Meteorology . Pp . 115-120 . Meteorol . Branch . Dept . Transport, Canada . Toronto .

Reifsnyder, W . E . 1967 . Forest meteorology : the forest energy balance . Inter . Rev . Forestry Res . 2 :127-179 .-. • _

Reimer, A . and R . Desmaris . 1973 . Micrometeorological energy budge t methods and apparent diffusivity for boreal forest and grass site$ - at Pinawa, Manitoba, Canada . Agric . Meteor . 11 :419-436 .

Richardson, L . F . 1920 . The supply of energy to and from atmospheric . eddies . Proc . Roy . Soc . London, A . 97 :354-373 .

Stewart, J . B . and A . S . Thom . 1973 . Energy budgets in pine forests . ! Quart . J . Roy . Meteor . Soc . 99 :154-170 .

Storr, D ., J . Tomlain, H . F . Cork and R . E . Munn . 1970 . An energy budge t study above the forest canopy at Marmot Creek, Alberta, 1967 . Water-IH r Resources Res . 6 :705-716 .

Strauss, R . 1971a . Energiehaushalt eines Fichtenwaldes, 1969 . Wiss . Mitt . Nr . 21, pp . 17-19, Meteorol . Inst ., Univ . M{lnchen .

Strauss, R . 1971b . Energiebilanz and Verdunstung eines Fichtenwalde s im Jahre 1969 . Wiss . Mitt . Nr . 22, Meteorol . Inst ., Univ . MJnchelp , 66 pp .

Szeicz, G ., G . Endr8di and S . J . Tajchman . 1969 . Aerodynamic and surfac e factors in evaporation . Water Resources Res . 5 :380-394 .

Tajchman, S . J . 1967 . Energie- and Wasserhaushalt verschiedene r Pflanzenbestdnde bei Munchen . Wiss . Mitt . Nr . 12, Meteorol . Inst . , Univ . M{lnchen . 95 pp .

Tajchman, S . J . 1971 . Evapotranspiration and energy balances of fores t and field . Water Resources Res . 7 :511-523 .

Tajchman, S . J . 1972a . The radiation and energy balances of coniferou s and deciduous forests . J . Appl . Ecology 9 :359-375 .

Tajc?Ymati, S . J . 1972b . Messui gen zum WHrmehaush;alt tber eine r Kiefsrnsfihhnun,g it Trookengebiet des Oberrheins . Allg . Forst- 4 : J . 2tg. 141 :35138 .

Tgjchmain, 5 . 3 : 1371, . 01 vertical profiled of Mtt-eorologidal parameters abtve a laeit of rough vegetation : J . iph}os . . Res . 78 :63.81-6385 .

Tdtltltr, G : B . 1961 . Basic lad trUMeb ation and measurements for plan t of jittnment and micrometeta `61ogy Soi1t Bull . 6 ; various pag. Jto& . Maiisat ; Univ ., Wise

Tanner, C : $ : 1967 . MeasuraMant of evapotranspiration . l : R . 1i . IMF , A H . R . Raise, and T . W . Minister, teas .) irrigation of Aggri titUre Lands . Amer . Soc . A. omt Monogr . 11 :534=04 . . Tanher C , E . 1968 . Evaptoratiai tf watt rtnit plants and soil . In a Koa lbtoaki, T ., t . (ed .) . Water Deficits and Plant ( owthl Vol . I . DeVelo nt , Control ahd MdaduttMaant . Chap . 4, pp . 13=1 Academic Ptess ; I .Y .

V . J . 1966 . A bibfettorological data logging system fot Agricultuta L search . Ga . Agr : E . Sta ., Univ : Ga ., Mimeo . Series L : g ; 244 . PP

Webb, E . K . 1965 . Aerial mitroolihate . In ; Waggoner ; P ._ (ed .) . Agficultural Meteorology . Chap . 2 . Meteor . Motiogr . 6 g 75t : APPENDIX I

Energy Flux Studies in a Coniferous Forest Ecosystem

L . W. Gay

Reprinted from : Franklin, J . F ., L . J . Dempster, and R . H . Waring (eds .) . Proceedings - Research on Coniferous Fores t Ecosystems - A Symposium . Pp . 243-253 . Pacific Northwes t Forest and Range Experiment Station . 1972,

Proceedings-Research on Coniferous Forest Ecosystems-A symposium Published in 1972 by Bellingham, Washington-March 23-24 i 1972 . Pacific Northwest Forest and Range Experiment Station For sale by the Superintendent of Documents, U .S . Government Printing Ofce Washington, D.C ., 20402 - Price $2 .50 Stock Number 0101-0233 Forest Service, U .S. Department of Agriculture Portland, Orego n

Edited by Energy flux studies in a Jerry F . Franklin L. J. Dempster coniferous forest ecosystem Richard H . Waring

Lloyd W. Gay Associate Professor of Forest Climatology Department of Forest Engineering Oregon State University Corvallis, Orego n

Abstract The fluxes of thermal energy between the atmosphere and a young Douglas-fir forest were measured durin g two contrasting summer days, one cloudless and one overcast . The energy budget components were evaluate d by the Bowen ratio method, with ceramic-wick psychrometers at the 26 .16 m, 28 .16 m, or 31 .16 m levels. The maximum height. of the tallest trees was 28 m, and the general level at the lop of the closed canopy was abou t 22 m. Daily totals of the energy budget components (cal/cm 2) under cloudless shies on July 29, -197J, were : solar radiation, 584 ; net radiation, 410 ; change in storage, 5 ; convection, -135 ; and latent energy, -280. The albedo was 0.09 on both the clear and the overcast day . Analysis of the overcast conditions of July 31, 1971 , yielded the following values : solar radiation, 171 ; net radiation, 134 ; change in storage, 6; convection, -39 ; and - . latent energy, -102. Problems of measurement and analysis-are discussed. These include the storage term in the biomass, and th e small gradients of potential temperature and vapor pressure above the canopy . Clear day gradients at noon, for example, were in the order of -0.03°C m-r and -0.03 mb m- . Techniques are presented for minimizing measurement errors.

Introduction thesis (Baumgartner 1965) . The advantages include sensitivity, mobility, and the benefits The level of biological activity at the sur- to be gained by use of a nondestructive tech- face of the earth is closely associated wit h nique. The application of these techniques t o cycles of energy and mass . The cycles of mas s the forest ecosystem appears feasible and use - and of energy are virtually interchangeabl e • ful. A number of studies have already bee n concepts,. Indeed, the transpiration and reported, but in general the effects of forest s photosynthetic components of the mass cycl e upon the cycles are not yet well know n can be studied through examination of th e (Baumgartner 1971, Tajchman 1971) . Conif- cycle of energy, with the energy required t o erous forests, as a class, are good absorbers o f change the phase of H2 O and CO 2 serving as solar radiation . The roughness of coniferous the connecting link . crowns also appears to effectively enhance The magnitudes and phase relationships o f mixing in the atmosphere near the top of the the mass and energy cycles are affected by th e canopy. These factors, combined with the characteristics of the surface, and by the stat e large surface area of canopies, make forest s of the atmosphere. The properties of vegeta- into very efficient exchange surfaces for wate r tion, particularly of low, cultivated eco- vapor, carbon dioxide, and energy . systems, have been investigated thoroughly i n Studies of the fundamental cycles of energ y a variety of studies that have clearly demon- and mass have begun at the Cedar River site in strated many advantages for energy budge t the Coniferous Forest Biome . A variety o f evaluations of transpiration and llhoLosyn- interrelated studies are planned in coopera-

213 - 37 -

Lion with physiologists, soil scientists, hy- net radiation is transformed into either ;,I: drologtsl ;, :DUI ►neteorolot;isl,s . The wor k rhai►ge in stored heat. in the soil and biomass, reported are is of preliminary research into ( ; ; a flow oI convective (sensible) energy be- the exchange of thermal energy between a tween the forest and the air, H ; or a flow of young Douglas-fir forest and the atmosphere . latent heat, XE, that is associated with a flux The energy exchange processes have bee n of water vapor E . The amount of energy evaluated during two contrasting conditions : transformed in photosynthesis, P, is of great the first during clear, warm weather charac- importance in productivity . However, since it terized by a large energy input to the forest , amounts to only a few percent in terms of th e and the second during cool, overcast condi- net radiation flux over the forest (Denmea d tions with a relatively small energy input to 1969), P will be neglected in this study . the forest . The objectives are to define th e The rate and direction of energy transfer energy transfer characteristics of the youn g depends upon the relative energy states of th e forest under conditions of both high and lo w canopy and the atmosphere and upon the rates of transpiration Study of these contrast- availability of radiant energy which is derive d ing conditions will help us to understand th e primarily from the sun. The state of energy is processes that control the exchange of energ y determined by temperature and by vapor con- and water vapor between the forest canop y centrations. Motion of the atmosphere ma y and the atmosphere . also enhance energy transfer, Energy transfe r models thus use measurements of temperature, vapor concentration, wind, and radian t -Experimental Methods energy in order to determine the flow o f energy between the forest and the atmosphere. The evaluation of thermal energy exchange The "Bowen ratio" model was selected fo r in the forest is simple in concept, requirin g use in this study because of relative simplicit y that appropriate boundaries be defined abou t in analysis and application, and because of its the forest site, that the quantities of energ y general acceptance based upon tests ove r crossing the boundaries be identified, and tha t other types of vegetation . The method wa s appropriate measurements be made at peri- first derived by Bowen (1926), and has bee n odic intervals. The periodic samples can the n adapted for energy transfer studies by a be combined to yield estimates of energy ex- number of workers (Fritschen 1965, Tanner change for the desired time interval . The 1960) . selection of an appropriate model is an im- The Bowen ratio model has been thor- portant step in evaluating the thermal energ y oughly described elsewhere, but a short dis- exchange. cussion here will help to place its applicatio n into perspective . First, we must sum th e Energy Transfer Models thermal energy fluxes active in the forest i n accordance with the principle of conservation Several review articles have discussed the of energy, to obtain application of various energy transfer model s to the problem of evaluating the exchange be- Q+G+H+XE+P=O . (1 ) tween the forest and the atmosphere The polarity convention considers fluxes t o (Baumgartner 1965, Federer 1970, Fritsche n 1ui0) . the surface as positive . After neglecting The basic problem concerns the transfor- photosynthesis, equation 1 can be solved fo r latent energy to yield mation of energy by the forest from radian t to nonradiant forms . The total energy thus XE = - (Q + G) / (1 + /3) (2) transformed is called net radiation, Q ; this equals the quantity of radiation of all wave - where 13 is the Bowen ratio of convective heat lengths absorbed at the surface, minus the to latent heat (H/XE). The Bowen ratio can be radiation lost by reflection and emission . The written

244 -38 -

=H/AE=yA,a De (3) Field Measurements

where y is the psychrometric constan t Net radiation, stored heat, and temperatur e (y 0.66 mb/°C at sea level) and an d and vapor pressure measurements were ob - Ae represent the measured differences i n -tained with a mobile data acquisition that has potential temperature and in vapor pressure a t been described by Gay . The truck-mounted two levels in the atmosphere above, but near, system includes sensors and supports, cabling , the surface . and a digital data logger with resolution o f The Bowen ratio model thus provides a 0.001 percent (0 .1 microvolt on a 10 millivol t rationale for partitioning the measured suppl y scale). Ceramic wick, wet-bulb psychrome_ter s of thermal energy into convection and laten t of the basic design of Lourence and Pruit t heat, based upon measurements of tempera- (1969), as modified by Gay (n.d.), were used ture and vapor concentration at only two for the temperature and vapor pressure levels above the canopy surface . Wind meas- measurements. urements are not required for the analysis , The sensors were mounted on a 33 .5 m although they are often useful in interpreta- (110 ft) tall, triangular TV tower about 0 .3 m tion of the results . The supply of thermal (1 foot) in width . Wind, temperature, an d energy (Q + G) is measured, so limits are vapor pressure measurements were made a t placed on the estimates of convection an d six levels, respectively, 26 .16, 27 .16, 28 .16 , Went heat with this method . • 29.16, 30.16, and 31 .16 m above the .forest The disadvantages of the model are pri- floor. The radiation budget components were marily associated with instrumentation ; measured from a height of 30 .66 m above th e accurate measurements are required in order floor. The tip of the tallest tree in the vicinity to measure the small gradients in temperatur e of the tower extended to 28 m, though th e and vapor found near the forest canopy . In bulk of the crowns were below 24 m, and th e addition, the basic relationship in equation 2 general level of crown closure was about 20 t o becomes undefined whenever 13 =-1 .. This 22 m above the floor . Five soil heat flux disk s normally occurs infrequently, and ordinarily were installed at the -2 cm level, just beneat h only for short periods near dawn or dusk the surface of mineral soil on the forest floor . when the amount of available energy i s Observations began on July 27 and con- limited. The magnitude of H and XE will tinued through August 1 . The sensors wer e normally be small at such times . sampled at 5-minute intervals during the day , and at 10-minute intervals at night . The obser- vation period spanned a range of weather con- Site Condition s ditions that included one clear and one com- The energy exchange studies were carried pletely overcast day . The data acquisitio n out in the broad, flat valley of the Cedar system performed well during this period . River, near Seattle, Washington . The soil and stand conditions have been described b y Fritschen (1972) . The stand is second-growt h Problems in Forest Energy Douglas-fir approximately 35 years old, wit h Budget Analyses an average of 570 trees per ha . The stan d canopy is relatively level ; the height of th e A variety of measurement problems are en-. tips of the tallest trees in the vicinity of th e countered in energy budget studies . In addi- experimental site is about 28 m . The site is tion, forests have unique characteristics o f adjacent to the lysimeter tree described by scale and mass that affect the application o f Fritschen (1972) . A series of energy transfe r studies are planned at this site in the future , L. W . Gay . An environmental data acquisitio n using the lysimeter tree, meteorologica l system . National Conference on the Forest, Weather , and Associated Environment, Atlanta, Georgia, Ma y towers, and a variety of models-for evaluatin g 1819, 1971 Mimen, 8 p Abstracted : Bull. Am . energy exchange processes . Mcaeor Soc 12 : 2112 20 : 1

245 - 39 -- the basic energy Ir .ulgel, model given in equa- a relatively large distance (10 m) in order t o tion 2 . The gradients ()I 1,4111IMYaLlIn ;111( 1 increase the differences 1,11 :11, are being ine :rs- vapor above I .he rough, por•oras furer;t. canop y ured to a level commensurate with the sensi- are very slight, and measurement difficultie s tivity of the data system (( ;alorix el, al . 1967 , increase with small gradients . Another majo r Storr et al . 1970) . In the second, reversin g problem is related to the difficulty of measur- sensors have been used to cancel the effect o f ing changes in stored energy in the biomass o f any small biases that may exist betwee n the forest. These problems will be placed into sensors (Black and McNaughton 1971) . perspective at the Cedar River site, for the y A large vertical separation of the sensor s s. affect the analyses and the subsequent inter- introduces questions of the representativenes s pretation of the results. of the measured gradients which should represent the effect of the underlying surface . It is e quite possible that a sensor placed 10 m abov e Evaluation of Gradients the canopy may be measuring properties of Gradients of temperature, vapor concentra- the atmosphere that are derived from a sur- tion, and wind are small above fores t face other than the one under investigation . canopies, even though the transfer of energy In contrast, a pair of reversing sensors, place d and mass may be proceeding at high rates . near the canopy, offers an excellent means fo r The small gradients result from mixing in- evaluating small gradients. However, th e duced by the mechanical turbulence create d reversing mechanism introduces new problem s by the rough canopy surface . In addition, th e of design for operation and support, in tal l exchange surfaces are distributed through a forests. considerable canopy depth so that the source s A graphical approach was used in this stud y of heat and vapor are diffuse, rather tha n to minimize sensor errors in the gradient , being concentrated as in the dense canopies o f measurements used for the Bowen rati o crops or other low vegetation . The smal l analysis . Initially, the mean hourly values o f gradients above the forest require tha t potential temperature were plotted agains t extreme care be taken in the development of the associated values of vapor pressure t o suitable instrumentation, and in the experi- yield 24 plots (1 per hour) for each day, wit h mental design controlling the deployment of each plot containing six points (one for eac h sensors. measurement level) . The plots will be straigh t The Bowen ratio model assumed that lines if similarity exists between the gradient s steady State conditions prevail, i .e ., the values of potential temperature and vapor concentra- of the variables do not change with respect t o tion, providing there are no errors of measure- time during the period of analysis . This is ment (Tanner 1963) . Since similarity is a n partially satisfied by averaging the values ove r assumption of the Bowen ratio method, thes e the period of an hour before applying th e plots were used to separate out the instru- model . Integration into hourly means als o ment levels that exhibited small offset s reduces the smalls random component of erro r throughout the day and to identify thos e s associated with the measurements. It does levels appropriate for use in the Bowen rati o Hart., however, reduce biases that are intro- analysis . duced by small differences among sensors . The similarity plots for July 29 are show n Such biases can be a source of serious error , for levels 1, 1, 6 in figure 1A, and for levels 1 , particularly with small gradients that exis t 2, 3 on July 31 in figure 1.1~. The potential above the forest . Bias errors must be handle d temperature (e, ) scale on the ordinate differs by techniques other than averaging . between the 2 days . The vapor pressure (e ) Two approaches have been used to cop e scale is shown in the legend . Note that the plo t with the problems involved in the measure- for each hour has been normalized by subtract- ment of small gradients above forests . In the ing the ~ and e value at the bottom level fro m first approach, the two levels of measuremen t the observations at each of the other levels . required for the Bowen ratio are separated by Therefore each plot actually shows the incre -

246 - 40

0.4 A . . 29 JULY 0.2 HOUR (PDT) --~ 8 17 4 v 0 0 I234567 11111 j 19202122232 O 8 9 10 II 12 13 14 15 16 o 1 8 w -0.2 H a -0.4 aW. w -0.6 ~ 0.0 4 B. 31 JULY

e e O y Ip e o 1202.1 22I . - - t 0 15 4 2 3 4 5 6 7 8 9 :io i i 1213 14 16 17 18 19 23 2 HOUR (PDT) -0.0 2 1 0ue -0.04 M6 VAPOR PRESSURE, M 8

Figure 1 . Similarity between gradients of potential temperature and vapor pressure . A. Clear weather on Jul y 29, 1971 . Levels 1, 4, and 6 . B . Overcast weather on July 31, 1971 . Levels 1, 2, and 3 .

meat of and e with respect to the firs t of the H and XE fluxes . measurement level near the canopy . For example, as plotted in figure 1A an d Though several points can be deduced from 1B, both fluxes will have the same sense in such similarity plots, the most important con- quadrants I and III where the slope is positiv e clusion concerns adequacy of data . The lin- (13>0) . In quadrant I, H and XE are both di- earity at the selected levels confirms thei r rected away from the surface and are negativ e acceptability for the Bowen ratio model , by the sign convention adopted earlier . In although an unexplained offset is evident a t quadrant III, H and XE are both positive, as level 2 during the 1000-1200 hours on July their sense is toward the surface . The slope 31 . and the coefficient are negative in quadrants Once the data is judged acceptable, on e II and IV . In quadrant II, H is toward the notes that the slope of the lines is A/Ae; this surface (advection) while XE is directed away , is directly proportional to f3, the Bowen ratio, while the fluxes in quadrant IV have th e as shown in equation 3 . Thus the relativ e opposite sense. The similarity plots can thu s slope of the similarity plots is an index to th e provide three things : a ready indication of th e way that the surface is partitioning the ne t magnitude of the Bowen ratio; an indication energy supply into convection and evapora- of the direction of the H and E fluxes ; and an tion . Further, the quadrant of each hourly identification of levels suited for the Bowen plot indicates the sign of R, and the direction ratio analysis.

- 41 - 247

Evaluation of Stored Hea t storage (-G) during the morning hours so tha t the daily integral approximately balanced . Application of the Bowen ratio model t o The final result is a much better estimate of forest measurements reveals a problem i n hourly changes in stored energy than could be measuring the stored heat term in equation 2 obtained by direct measurement of the soi l that is a direct consequence of the nature o f storage component alone . The magnitude of the forested surface . The scale of the fores t the stored energy change will be investigated elements makes the change in heat storage i n further during future studies. the biomass difficult to measure, and man y The major problems in application of th e studies (Baumgartner 1956) have treate d Bowen ratio model to the forest appear to b e these changes as negligible. This is certainly associated with adequate precision of th e true on a daily basis, but an appreciabl e measurements . The similarity tests demon- amount of energy appears to move int o strated here appear useful in eliminating storage (-G) in the biomass in the early morn- errors from the data . Further, the changes i n ing, and out again (+G) in the early evening . stiired energy can be estimated by an indirec t These amounts ordinarily will balance when method, based upon gradient measuremen t totaled over the day, but comparisons amon g and knowledge of the time at which th e hourly totals may be in error unless the stored heat flux reverses sign . energy storage changes are estimated for th e biomass as well as for the soil . The estimates of biomass storage change The Forest Energy Budge t were derived here by an indirect method tha t Energy budget analyses were developed for may prove useful in other situations as well . 2 days that represent quite different amount s First, application of the Bowen ratio model to of available energy . The solar energy input to data collected during the early evening hour s the forest was large on July 29, 1971, a day frequently resulted in positive estimates o f characterized by clear skies and a warm mea n latent energy (condensation) when the vapor temperature (24 .4" C). In contrast, July 31 , gradient clearly indicated that evaporatio n 1971, was completely overeast with reduce d was taking place . This apparent anomaly i n levels of incoming solar radiation and a rela- the Bowen ratio estimate of XE could occu r tively cool mean temperature. (17.4°C) . as a consequence of underestimating th e Examination of the energy budgets unde r storage term G . Ern order to obtain the prope r such contrasting conditions will improve ou r sign on XE under these conditions, the esti- understanding of the basic processes that mate of G must be increased to a positiv e govern energy transfer between the forest and value that exceeds the absolute value of th e the atmosphere . negative net radiation . The diurnal energy budget will first be Estimates of the minimum probable value examined with respect to daily totals ; a dis- of G were obtained in this manner for th e cussion of the relationships among hourl y hours between sunset and the time near mid- values of the energy budget components wil l night when the vapor gradient changed direc- then follow . tion, indicating the beginning of condensation . As a second step, the crossover point s The Diurnal Energy Budge t between release ({G) and uptake (-G) o f stored energy were then estimated from m y The daily energy budget totals are pre- experience and that of others (Grulois 1968 ) sented for the separate periods of daylight (1 3 as being about 1 hour after sunrise and 3 hours) and night (11 hours) and for 24-hou r hours before sunset . The release of stored totals in table 1 . Dividing the daily totals into heat (+G) is then defined by the two cross - daylight and night portions enhances futur e over points and by the magnitude during the comparisons that may be made with data early evening hours . The third and final ste p collected under different daylengths at Ceda r was to estimate the magnitude of the gain in River or elsewhere.

248 - 42 -

Table 1.-Diurnal energy budget components 1

July 29-clear July 31-overcast Period @ G H AE @ G H AE

cal/cmZ t-.-- -- . .

Daylight 454 43 -14 3 -263 -263 141 - -7 -38 -9 6

Night -44 48 8 -17 -7 13 - .1 1 . 1 ; - ,

Daily total 410 5 -135 -280 134 6 -39 -102

Totals are given for the daylight hours, 0630-1930 PDT ; night hours, 1930-0630 PDT; and the full day , 0000-2100 PDT .

There is a large difference in the radiatio n Baumgartner (1971), forests are effective supply on the 2 days, as net radiation totale d absorbers of solar radiation . 410 cal/cm 2 under the clear skies of July 29 , The changes in stored energy (G) tabulated ~ and only 134 cal/cm 2 for the overcast condi- in table 1 for the 24-hour period are nea r tions of July 31 . These totals include a stead y zero, which is in accord with observations net loss of radiation at night, amounting t o reported elsewhere . This term represents th e -44 cal/cm 2 under clear skies, and -7 cal/cm2 changes in heat storage of both the biomas s under the overcast conditions. and the soil. Most of the storage changes are The net radiation term represents th e attributed to the biomass ; the indirec t energy converted from radiative to nonradia- methods used to estimate the changes i n tive forms by the forested surface . The short - storage have been described in an earlier sec- wave radiation from the sun makes up th e tion . largest component of the net radiation . Dur- I estimate that -43 cal/cm2 went into stor - ing the 13 hours of daylight on the clear day , age during the daylight hours on the clear da y the forest received 584 cal/cm 2 of solar radia- and that 48 cal/cm 2 came out of storage dur- tion and reflected 55 cal/cm 2 . Under overcast ing the night. The storage changes on th e skies, the forest received 171, and reflecte d cloudy day proceeded in a similar direction , 16, calories/cm 2 . The albedo was 0 .09 on but the magnitudes were much smaller . both days . The storage term at Cedar River appears Since the gain and loss of longwave radia- large because of the large quantity of the bio- tion also enters into the supply of radiant mass there . The biomass is as yet unmeasured , energy, it is not helpful to calculate a short- however. Attempts will be made to measure wave/net radiation ratio as an index of effi- the storage flux directly in future experiments . ciency of conversion . However, the lo w The convective flux for the clear da y albedo value (0 .09) emphasizes the efficiency totaled -135 cal/cm 2 , directed away from the with which the Douglas-fir canopy absorb s forest into the atmosphere . A slightly larger solar radiation . This low reflectivity is similar amount, -143 cal/cm 2 , was lost during day - to values reported for other coniferous can- light, but 8 cal/cm 2 was gained by the canop y opies (Stewart 1971), and is much lower tha n at night when the canopy temperatures tli 0.2-0 .2fi albedo values that commonl y dropped below that of the air . Under overcast , prevail over crops and other low vegetatio n sky conditions, -38 cal/cm were lost over the (Monteith and Si.eicz 1961) . As noted by full day.

43 -

Latent eucrgy war; the largest dissipatio n aid of figure 2 which : :bows Ihr• hourly vahws Jena on each of the days,, totaling 28 0 on .July 29, ;Hid figure 3 which shrew :; the call( nr fur July 29 ;aid 102 cill/cni on Jul y hourly valuer, on July 31 . Kach plotted pbbll. 31 . There Mt., a ~tct Toss oI latent energy b y represents the midpoint, of an hourly mean . night, as well as by daylight, for both days . The daytime, night and daily totals in table 1 The evaporation equivalent of the laten t were integrated from the rates shown in thes e energy total was about 0 .5 cm on July 29 , two figures . The 2 days exhibit different char- and 0 .18 cm on July 31 . acteristics, so they will be discussed separately . The Bowen ratio (PH/XE) is a measure o f The symmetry of the bell-shaped net radia- i how the surface partitions the energy suppl y tion curve on July 29 confirms that the skies between sensible and latent heat . The mea n were cloudless on that day . The maximu m daily value of Q was 0.48 for the clear day , intensity occurred during the hour centere d and 0 .38 for the overcast day . The difference on 1300 hours PDT, which closely corre- in between days is not large, but it suggest s sponded with solar noon . The net radiation that the forest partitioned more of the energy values became positive about 1 hour after sun - supply into convection on the clear day than rise and remained positive until shortly befor e on the overcast day . From another viewpoint , sunset, indicating the hours in which there the ratio of ,\ E/Q was 0 .67 on the sunn y was a surplus of energy that might be dissi- day, and 0 .76 on the overcast day . This is i n pated through the other energy budget com- the direction that one might expect for a ponents . The net radiation was negative stand of vegetation that receives a large inpu t throughout the night, as the surface lost of energy . energy to the atmosphere . The greatest ne t radiation loss occurred at 2200, about 1 hou r Hourly Energy Budget s after sunset at a time when the forest wa s rapidly losing the absorbed solar radiatio n The phase relationships among the energy that had been stored during the day . budget components can be examined with the The phase of the fluxes is also of interest,

4 6 8 10 12 14 16 TIME, HR (PDT )

figure 2 . Energy budget components under clear skies . Symbols: net radiation, Q ; change in heat storage of soil and biomass, G ; convection, H ; latent energy, XE .

250 - 44 -

4 6 8 10 12 14 16 18 20 22 24 TIME, HR (POT)

Figure 3. Energy budget components under overcast skies . Symbols: net radiation, Q ; change in heat storage of soil and biomass, G ; convection, H ; latent energy, XE .

as G, H, and XE all lag behind Q . Let us ing, but latent energy was apparently favore d consider the stored heat flux first . It reaches at the expense of sensible energy during the its peak flow into the biomass and soil (-G) in afternoon . A similar pattern is evident i n midmorning, and reverses to flow out of the measurements. over a ybung Douglas-fir forest biomass (+G) in the late afternoon and earl y near Vancouver, B .C. (Black and McNaughto n evening. The change in stored heat appears to 1971), and over a mixed hardwood forest provide a significant source of energy to the (Grulois 1968) . surface throughout the night . The phase shift in latent energy into th e The sensible heat flux, H, reaches its maxi- afternoon is probably related to a vapor pres- mum about two hours after G, but still an sure deficit which has an afternoon maximu m hour before solar noon . Sensible heat i s on clear days . Stewart. and Thom2 have con- directed away from the surface during day - cluded that the latent energy flux from thei r light (-H), but reverses in direction as convec- pine forest site in England is controlled mor e tion begins t,o provide energy (+11) to th e by the vapor pressure deficit than by th e surface during the night . During this period , supply of available energy . This conclusion i s the canopy cools below air temperature du e based upon their evaluation of the interplay to longwave emission . between the relatively large internal resistanc e The latent energy flux reached its maxi - to transfer and a small external resistance ; the mum about two hours after solar noon . ratio for the pine site was in the order o f Evaporation continued well into the night; 20 :1 . only during the early morning hours did a rather small amount of condensation take place. Conclusions The marked phase shift between sensible and latent energy is of interest, as many The observations reported . here are an studies have shown these two fluxes to be in , initial contribution toward the problem o f phase with net radiation (Baumgartner 1956 , evaluating the flow of energy and mass be- Denmead 1969, Rauner 1960) . The Douglas- tween the atmosphere and the young Douglas : fir forest, in contrast, partitioned the energ y fir forest at the Cedar River site . available at the surface into se.nsible and latent energy on a preferential basis . Thi 2 .1 B . Stewart and A . S . Thom . Energy budgets i n s pine forest . Inslil .ute of Hydrology, Wallingford , partition was on a I :1 basis during the .morn - I3erk4lire, U . K Unpublished mauiuscripl, 1972

25 1 45 -

Forested surfaces arc generally considere d O Of f ice of Water Resources Research, as . to be effective energy exchange surfaces . The authorized under the Water Resources Re - results confirm that this young stand has a search Act of 1964, and administered by th e high absorptivity for solar radiation, with a n Water Resources Research Institute, Orego n albedo of 0 .09 for both clear and overcas t State University ; and the National Scienc e conditions . This high absorptivity contributes Foundation Grant No . OB-20963 to th e to the large net radiation values that were Coniferous Forest Biome, U .S. Analysis o f measured under clear skies . Ecosystems, International Biological . Program. The role of the forest in dissipating the ne t This is Contribution No . 40 to the Coniferou s radiation is of particular interest . The porous , Forest Biome and Paper 8.44, Forest Research aerodynamically rough canopy is effective i n Laboratory, School of Forestry, Oregon Stat e transferring sensible and latent energy int o University. the atmosphere . The large quantity of fores t biomass may also involve an amount of store d thermal energy that is of significance durin g Literature Cited short periods, even though the daily totals ar e small . Summed over a 24-hour period, evapo- Baumgartner, A . 1956 . Investigations on the transpiration was about 280 cal/cm 2 min (0 .5 heat- and water economy of a young forest . cm water equivalent) or about two-thirds of Translated from Ber . Deut . Wetterdienst 5 : the net radiation that was transformed under 4-53. Commonwealth Sci . 1nd . Res . clear skies . Evapotranspiration was relativel y Organ . Translation 3760 . Melbourne , larger on an overcast day, about three- Australia, 1958 . quarters of net radiation, although the tota l 1965 . The heat, water and car- amount of latent energy (102 cal/cm 2 min, or bon dioxide budget of plant cover : meth- 0.18 cm water equivalent) was considerably ods and measurements . In F . E . Eckard t lower . (ed.), Methodology of plant eco-physi- These results provide initial estimates o f ology : Proceedings of the Montpellier the amounts of energy transferred during ex- Symposium, p . 495-512 . Paris : UNESCO . treme conditions under cloudless and unde r 1971 . Wald als Austauschfaktor overcast skies. The exchange of energy an d in der Grenzschicht Erde/Atmosphere. mass depends not only upon the energy inpu t Forstwiss . Cbl. 3 : 174-182. to the forest, however, but also upon th e Black, T . A ., and K. G. McNaughton . 1971 . physiological response of the vegetation. .Now Psychrometric apparatus for Bowen-rati o that the instrumentation system and th e determination over forests . Bound . Layer analysis model have been tested at this site , Meteorol . 2 : 246-254 . subsequent research will include a range o f Bowen, I. S. 1926 . The ratio of heat losses by environmental conditions . Instrumentation conduction and by evaporation from an y development and model testing will continue water surface . Phys . Rev . 27 : 779-787 . in cooperation with the lysimeter installatio n Denmead, O. T. 1969 . Comparative micro - and eddy flux system of cooperating investi- meteorology of a wheat field and a fores t gators. Ultimately, analysis and interpretatio n of Pin us radiate. Agric. Meteorol . 6 : of the energy transfer studies must includ e 357-371 . physiological as well as physical characteris- Federer, C . A. 1970 . Measuring forest evapo- tics of the stand . transpiration-theory and problems . USD A Forest Serv . Res. Pap . NE-165, 25 p . Northeast . Forest Exp . Stn ., Upper Darby , Acknowledgments Pa. Fritschen, L . J...1965 . Accuracy of evapotran- Th.e work upon which this publication i s spiration determinations by the Bowen based was supported in part by funds pro- ratio . method . Bull. Int. Assoc. Sci. Iidiol. vided by the U .S. Department of Interior, 1.0 : . 38-48 .

252 46 -

1970. Evapotranspiration .and Meteorol . 8 : 492-498 . meteorological methods of estimation a s Monteith, J. L., and G . Szeicz. 1961 . Tlie applied to forests. In J. M . Powell and C . radiation balance of bare soil and vegeta- F. Nolasco (eds .), Proceedings of the Third tion. Quart. J. Roy. .Meteorol. Soc. 87 : Forest Microclimate Symposium, p . 8-27. 159-170. Can. For. Serv., Calgary, Alberta. Rauner, Yu. L. 1960. The heat balance of for- 1972. The lysimeter installatio n ests. Izvestiya Akademii Nauk SSSR, Seriya on the Cedar River Watershed . In Jerry F . Geograficheskaya . 1 : 49-59 : Translated Translated by Franklin, L. J. Dempster, and Richard H. Israel Prog. Sci. Transl . No . 1342. U.S. Waring (eds.), .Proceedings-research o n Dep. Commerce, Washington, D .C. coniferous forest ecosystems-a sym- Stewart, J. B. 1971 . The albedo of a- pine for - posium, p . 255-260, illus. Pac . Northwest est. Quart. J. Roy. Meteorol. Soc. 97 : Forest Range Exp . Stn ., Portland, Oreg . 561-564. Galoux, A ., G . Schnock, and J . Grulois. 1967. Storr, D ., J. Tomlain, H . F. Cork, and R . E. Les installations eco-climatologiques . Munn . 1970. An energy budget study Travaux Stat . Rech. Eaux et Forets , above the forest canopy at Marmot Creek , Groenendaal-Hoeilaart, Serie A, 12, 52 p. Alberta, 1967 . Water Resour. Res. 6 : Gay, L. W. (n .d .) On the construction and use 705-716 . of ceramic-wick psychrometers . In R . W. Tanner, C . B. 1960. Energy balance approach Brown and B. P. Van Haveren (eds .), to evapotranspiration from crops . Soil Sci. Psychrometry in water relations research. Soc . Am. Proc. 24 : 1-9 . Utah Agric . Exp. Stn. (In press. ) 1963. Basic instrumentation an d Grulois, J. 1968 . Flux thermiques et evapo- measurements for plant environment an d transpiration au cours dune journee micrometeorology . Soils _Bull. 6, various sereine . Bull. Soc . r. Botanique Belg . 102: pagination . Madison : Univ . Wis. 27-41 . Tajchman, S. J. 1971. Evapotranspira i z d Lourence, F . J ., and W. O. Pruitt . 1969. A energy balances of forest and field . Water psychrometer system for micrometeor- Resour. Res. 7 : 511-523 . ology profile determination . J. Appl.

-47 - 253