Agricultural and Forest Meteorology, 32 ( 1 9 8 4 ) 41 - -53 41 Elsevier Science Publ i shers B.V., A m s t e r d a m - - P r in ted in The N e t h e r l a n d s

EVALUATION OF CANOPY TEMPERATURE--EVAPOTRANSPIRATION MODELS OVER VARIOUS CROPS*

J.L. H A T F I E L D

USDA-ARS, Plant Stress and Water Conservation Research Unit, Route 3, Lubbock, TX (U.S.A.)

R.J. R E G I N A T O and S.B. IDSO

U.S. Water Conservation Laboratory, Phoenix, AZ (U.S.A.)

(Received May 2, 1 9 8 3 ; r e v i s i o n accep t ed F e b r u a r y 27, 1984)

A B S T R A C T

Hatf ie ld , J. L., Reg ina to , R. J. a n d Idso, S. B., 1984. Eva lua t ion of c a n o p y t e m p e r a t u r e - - e v a p o t r a n s p i r a t i o n mode l s over var ious crops. Agric. For. Meteorol . , 32: 41- -53 .

C a n o p y t e m p e r a t u r e s , w h e n m e a s u r e d r emo te ly , o f fe r a m e t h o d of e s t ima t ing evap- o t r a n s p i r a t i o n wi th surface energy ba lance models . E q u a t i o n s wh ich have been de- ve loped by o the r s have been eva lua ted on ly at a l imi ted n u m b e r o f l oca t ions and wi th a few crops. Our s t u d y was c o n d u c t e d at several loca t ions w i th weighing lys imete rs w i t h a var ie ty of c rops a r o u n d t he U n i t e d S ta tes : Brawley, CA; Temple , TX; Lincoln , NE; St. Paul, MN; Fargo , ND; Kimber ly , ID; and Davis, CA, to evalua te e v a p o t r a n s p i r a t i o n ut i l iz ing c a n o p y t e m p e r a t u r e as an i n p u t in to the surface energy balance. The resul ts show t h a t e v a p o t r a n s p i r a t i o n ca lcu la ted f rom the a e r o d y n a m i c res is tance fo rm of the surface energy ba lance was well co r re l a t ed w i th lys imete r m e a s u r e m e n t s at all loca t ions . The er rors us ing the surface energy ba lance were less t h a n 10% in all cases for full g round cover. The B a r t h o l i c - - N a m k e n - - W i e g a n d m e t h o d was m o r e closely coup led to ne t r ad i a t i on t h a n c a n o p y t e m p e r a t u r e .

U n d e r par t ia l c a n o p y cover, d i f fe rences b e t w e e n the two mode l s were apparen t . The B a r t h o l i c - - N a m k e n - - W i e g a n d m o d e l ove rp red i c t ed w h e n the ac tua l e v a p o t r a n s p i r a t i o n was above 200 W m -2 because of its insens i t iv i ty to surface t e m p e r a t u r e . However , t he surface energy ba lance m o d e l e x h i b i t e d on ly a sl ight ove rp r ed i c t i on above 200 W m -2 w h e n a weighed c o m p o s i t e surface t e m p e r a t u r e ( r ep resen ta t ive o f bare soil and c rop t e m p e r a t u r e ) was used. This small ove r p r ed i c t i on could be o v e r c o m e by cons ider ing the soil hea t f lux te rm. There was no loca t ion bias in t he surface energy ba lance mode l , wh ich shows t h a t i t s h o u l d work well a t o t h e r loca t ions .

I N T R O D U C T I O N

Evapotranspiration from a surface is a component of the partitioning of e n e r g y r e c e i v e d b y t h a t s u r f a c e . T h i s p r o c e s s c a n b e d e s c r i b e d b y t h e f a m i l i a r

energy balance equation (Monteith, 1973) with expanded sensible and latent

* C o n t r i b u t i o n f rom the Cal i fornia Agr icu l tu ra l E x p e r i m e n t S ta t ion , Pro jec t 3963-H, and t he USDA-ARS.

42

heat terms, as

R n + G = p C p (Tc --Ta) +pCp [es(Tc)--ea] r a "y r a + r c

(1)

where R n is the net radiation (Jm -2 s -1 ), G the soil heat flux (Jm -2 s -1 ), p C~ the volumetric heat capacity of air (Jkg -1 °C-1), Te the canopy temperature (°C), ra the aerodynamic resistance (sm -1 ), 7 the psychrometric constant (kPa °C-1 ), es (Te) the saturated vapor pressure at Te (kPa), ea the actual vapor pressure (kPa), and re the canopy resistance to water vapor transfer (sm -1 ). This form of the energy balance equation has been often cited, but the difficulty in measuring canopy temperature had led to other approaches which decouple eq. 1 from a direct surface temperature measurement. The Penman--Monteith method, which includes a canopy resistance term, can be used, but it is difficult to assess the canopy resistance changes with soil water status (Monteith, 1973).

With the development of thermal infrared thermometers that are accurate and easily used, several investigators proposed the use of surface temperature in surface energy balance models to estimate evapotranspiration (Bartholic et al., 1970; Brown and Rosenberg, 1973; Jackson et al., 1977; Soer, 1980; and Sequin and Itier, 1983). These approaches have ranged from mani- pulation of eq. 1 to the use of empirical coefficients for deriving daily evapotranspiration from midday canopy measurements (Jackson et al., 1977). Although these approaches have been proposed only limited evaluations have been conducted. Stone and Horton (1974) compared the techniques suggested by Bartholic et al. (1970) and Brown and Rosenberg (1973) against the more traditional methods (Bowen ratio (Bowen, 1926), Penman (1948), and van Bavel (1966)) and concluded that both approaches would have promise. The authors suggested that the Bartholic--Namken-- Wiegand method may be better to use because it does not require knowledge of the aerodynamic resistance which is a function of windspeed and the canopy aerodynamic properties. Verma et al. (1976), after extensive error analysis, concluded that the aerodynamic resistance method given by Brown-- Rosenberg (1973) was more sensitive to errors in canopy temperature than errors in aerodynamic resistance. Blad and Rosenberg (1976) determined that either a resistance or mass transfer model utilizing remotely sensed canopy temperature would be useful in regional models.

Other comparisons over limited conditions have been conducted by Heilman and Kanemasu (1976), Heilman et al. (1976), Soer (1980), Sumayao et al. (1980), and Hatfield et al. (1983). Soer (1980) found that the evapotranspiration estimated by an energy balance model agreed 70% of the time with water balance measurements in watersheds. He stated, however, that regional estimation with remotely sensed canopy temperatures would allow for the standard error about the mean to approach zero and

43

that this would make the estimates comparable with those from a Penman model, with both utilizing the same meteorological data base. Heilman et al. (1976) used remotely sensed canopy temperatures and found the models agreed very well after the remotely sensed canopy temperatures were corrected for atmospheric at tenuation between aircraft and ground.

This experiment was implemented with the objective of evaluating evapotranspiration models which use remotely sensed canopy temperature as an input. Our report addresses the performance of the models compared to lysimeter evapotranspiration for various crops and locations in the United States.

MATERIALS AND METHODS

Description and equations

The energy balance of a surface given in eq. 1 (Monteith, 1973) can be rewritten as

pCp ( T c - - T a ) hE = Rn + G (2)

3' ra

where hE is the latent heat of vaporization (J kg -1 ). Sumayao et al. (1980) and Hatfield et al. (1983) showed that when the canopy was cooler than air hE would be larger than net radiation. In this method we calculate r a from

ra = [ln (z--d)/z o] 2/k2u (3)

with z being the height (m) of observation of air temperature, windspeed, and vapor pressure above the crop surface, d the displacement height (m), zg the roughness length (m), k yon Karmans constant (0.40), and u the windspeed (m s -1 ). Following Monteith's (1973) correction for stability, we corrected the aerodynamic resistance as

( n(z--d)g(Tc--Ta))' (4) rac = r a 1 Tu 2

where g is the acceleration of gravity (9 .8ms -2) and T the absolute temperature (K), taken as the mean of the canopy and air temperatures. Monteith (1973) suggested that a value of 5 for n would be appropriate for field conditions. The roughness lengths and displacement heights for the crops in this s tudy are given in Table I.

Bartholic et al. (1970) rearranged eq. 1 to obtain the following equation

Rn+G E = (5)

(Ta - -Tc) 1 + 3 "

[es(Ta)--es(Tc)]

44

0

2-

o

0

£

~ 6 ~ o

• o

.N

N M

09

N

0

e', 0

0

> > > 1 > > 1 > ~

O 0 0 0 0 O 0

O ~ 0 ~ 0 ~ 0 O ~ 0 ~ 0 ~ o o o o o o o

O ~ 0 ~ 0 ~ 0 ~ 0 ~ O ~ O ~

0 0 0 0 0 0 0

~ 0 0 0 0 O 0

O 0 0 O 0 0 0

~ 0 ~ 0 ~ 0 0 ~ 0 ~ 0 ~ 0 ~

?T ? ?T 71

0

0 oO

0 c c~

x

x

o

r,

o

o

o

o

0 o 0

4.a

i ~

.o

~0 LO

C'xl Cxl 0 ~

0 0 0 0 0 0

~'~ 0 ~'~ C'-1 LO Cq 0

X X X X X X X

0 0 0 O0 ~0 O0 "~"

X X X X X X

C) 0 0 LO ~ 00

E -

45

This a p p r o x i m a t i o n sets the sur face and air a t the sa tu ra t ion v a p o r pressure which l imits the equa t i on to p o t e n t i a l e v a p o t r a n s p i r a t i o n f r o m an inf in i te ly we t surface .

EXPERIMENTAL PROCEDURES

Dur ing the s u m m e r of 1980 , an e x p e r i m e n t was c o n d u c t e d at several loca t ions in the wes te rn half o f the Uni t ed States . The loca t ions and c rop i n f o r m a t i o n are given in Tab le I. These sites were chosen because t h e y had lys imete rs wi th suf f ic ien t a ccu racy to give a rel iable h o u r l y measu re of evapo t r ansp i r a t i on . Table I I descr ibes the lys imete rs at each loca t ion . In all cases e x c e p t T e m p l e the f e t ch r e q u i r e m e n t s were m e t for these measure - m e n t s and at T e m p l e the ly s ime te r was s u r r o u n d e d by tal ler corn. The sites m o n i t o r e d also covered a wide range in la t i tude , c l imate and crops , which was d e e m e d necessary to achieve ou r e x p e r i m e n t a l object ives . At each e x p e r i m e n t a l site, m e a s u r e m e n t s were m a d e of c rop he ight and the values fo r roughness length and d i s p l a c e m e n t height were d e t e r m i n e d f r o m typ ica l values r e p o r t e d in the l i te ra ture . Fo r those loca t ions wi th less t han c o m p l e t e g round cover the roughness length was a s sumed to be larger t han r e p o r t e d values as suggested b y V e r m a and Barfield (1979) and Hat f ie ld (1982, unpub l i shed data) .

A t each loca t ion the e x p e r i m e n t a l p r o c e d u r e was the same, including the f r e q u e n c y o f obse rva t ion . Meteoro log ica l var iables were r eco rded at one- m i n u t e intervals by a c o m p u t e r - c o n t r o l l e d da ta acquis i t ion sys t em and the da t a r e c o r d e d o n t o f lexible discs. The sys t em was housed in a m o b i l e van which was p a r k e d near the l y s ime te r a t each loca t ion . Observa t ions r e c o r d e d on the r o o f o f the van were global solar r ad ia t ion and longwave rad ia t ion f r o m the a t m o s p h e r e . The p y r a n o m e t e r was an E p p l e y PSP p y r a n o m e t e r wi th a WG295 fi l ter. L o n g w a v e r ad ia t ion f r o m the a t m o s p h e r e was m e a s u r e d wi th a Swiss teco p y r a d i o m e t e r wi th a b l a c k b o d y cup. Pos i t ioned over the l y s ime te r was an inver ted E p p l e y p y r a n o m e t e r , a F r i t schen min i a tu re net r a d i o m e t e r , an asp i ra ted d ry -bu l b and we t -bu lb t he rmi s to r , and Gill l i g h t c h o p p e r a n e m o m e t e r . The p y r a n o m e t e r and ne t r a d i o m e t e r were pos i t i oned 5 0 c m above the u p p e r sur face o f the c a n o p y and, where appl icab le , ove r the p l an t row. The asp i ra ted dry- and we t -bu lb t h e r m i s t o r s and a n e m o m e t e r s were pos i t i oned 3 0 c m and 1 3 0 c m above the u p p e r surface o f the c a n o p y . These i n s t r u m e n t s were a t t a c h e d to the da ta acquis i t ion s y s t e m b y a m u l t i c o n d u c t o r shie lded cable and were also s ampled once pe r m i nu t e . The da t a were r edu ced and qua l i ty con t ro l l ed t h rough rou t ines desc r ibed b y Hat f ie ld et al. (1981) and r e p o r t e d as in tegra ted hou r ly values.

C a n o p y t e m p e r a t u r e s were m e a s u r e d over the l y s ime te r wi th an inf rared t h e r m o m e t e r w i th a 4 ° fov and 8 - - 1 4 p m waveband . This uni t has an accu racy o f + 0 .5°C and a r e so lu t ion of 0 .1°C. M e a s u r e m e n t s were m a d e of the c r o p f r o m each card ina l d i r ec t ion a t a b o u t a 30 ° angle f r o m the

46

horizontal. Where the ground cover was not complete, nadir views of the soil between the rows were made along with temperatures of individual plant leaves. These measurements were made at half-hourly intervals, from 15 min before sunrise to 15 min after sunset. In the analysis procedure, the canopy temperatures were averaged to match the hourly meteorological values. In cases where incomplete ground cover was present composite scene temperature was determined by computing a weighed average based on the fraction of bare soil relative to the crop cover and their respective tempera- tures. This could be improved by monitoring composite scene temperature with a nadir looking angle infrared thermometer.

Lysimeter values were recorded at the same half-hourly intervals as canopy temperatures. These data were then plotted and smoothed with a three-term running average from which hourly values of evapotranspiration could be determined. It was felt that this time resolution would be adequate for all lysimeters and would insure the greatest degree of precision.

The meteorological, canopy temperature, and evapotranspiration data formed the data set from which all subsequent analyses were conducted. Evapotranspiration was calculated from eq. 2 with the inclusion of eq. 4 for the surface energy balance and eq. 5 for the Bartholic--Namken-- Wiegand method. The Penman--Monteith equation was given as

(Rn + G) + pC, (es (Ta) --ea)/ra ET = (6)

A + 7

where A is the slope of the saturation vapor pressure curve (kPa °C-1 ) and was used with a canopy resistance term of zero in the calculations. This method provided an estimate of potential evapotranspiration for our study.

RESULTS AND DISCUSSION

For each location the Bartholic--Namken--Wiegand model (eq. 5), the surface energy balance model (eq. 2), and Penman--Monteith combination equation (eq. 6), formulas for the estimation of potential evapotranspiration, were calculated with the hourly data set. The hourly estimates were then compared with measured evapotranspiration values for each location. On each graph the integrated daily total of evapotranspiration is given as a point of reference for the reader. Only one representative day is shown for each site.

Diurnal trends of net radiation, measured evapotranspiration and the calculations of potential and actual evapotranspiration are shown in Fig. 1 for alfalfa at Kimberly. These results were for a clear day with moderate windspeeds, and the evapotranspiration values were quite high. The Bartholic--Namken--Wiegand model underestimated actual evapotran- spiration values and the other models throughout the day. Evapotranspiration estimated by the surface energy balance (eq. 2) closely followed the

47

I O 0 0 E

8 0 0

6OO

2 0 0

0

I I I I I I I I I I q I T I KIMBERLY-DAY 208 ALFALFA

- - NET RADIATION DArLY TOTAL - - ____ LYS IMETE R ET ET = 84ram

O. - -O PENMAN E3-- .O S - N - W ~ S U R F A C E

ENERGY BALANCE

6 a I0 12 14 16 18

T I M E ( H O U R S I

~ I000 'E

i 800 600 400 2O0 0 20

' l l [ I r ] l l l [ ]

LINCOLN-DAY 179 SOYBEAN - - N E T RADIATION OAJLy TOTAL

. . . . LYSIMETERET E T ' 3 8 m m ~- -OB-N-W ~,----E~SURFACE

ENERGY BALANCE

N PENMAN

6 8 I 0 12 14 16 P8 ~

T I M E ( H O U R S )

I 2 O

Fig. 1. Diurnal trend of net radiation and measured evapotranspiration for alfalfa and calculated evapotranspiration by the surface energy balance, Bartholic--Namken--Wiegand and Penman--MonLeith combination model for Kimberly, Idaho.

Fig. 2. Diurnal trend of net radiation and measured evapotranspiration for soybeans and calculated evapotranspiration by the surface energy balance method, Bartholic-- Namken--Wiegand model, and Penman--Monteith combination model for Lincoln, Nebraska.

lysimeter values throughout the day. The differences between the surface energy balance and the measured values for Kimberly were largest on this day, although they were generally less than 10%. Estimates from the Penman--Monteith approach provide an estimate of potential evapotran- spiration and are shown for the benefit of the reader. Since the purpose of the study is to compare actual evapotranspiration with the canopy temperature methods, the estimates of potential evapotranspiration will not be discussed in the remainder of the paper.

In comparing evapotranspiration estimated from the measurement of net radiation directly or from the calculation of net radiation components, differences were always less than 5%; there appeared to be no bias in the differences. For this reason, only the calculation of net radiation, as used by Soer (1980) in his evaluations, will be discussed for the remainder of the paper. In regional applications a direct measure of net radiation may not be feasible and for this reason it was felt this approach was closer to our objectives.

For the full ground cover alfalfa at Kimberly, the evapotranspiration was substantially reduced. This is illustrated by Fig. 2 for soybeans at Lincoln. Data for this day show that the conditions were generally clear with some clouds in the afternoon. Of more interest is the behavior of the measured evapotranspiration rates and the two models. The Bartholic--Namken-- Wiegand model tracks the net radiation curve very closely but greatly overestimates the measured evapotranspiration rates while the surface energy balance method more closely approximates the measured values throughout the day.

In the partial canopy-cover cases at Lincoln (Fig. 2), Fargo, and Brawley (Fig. 3), the surface temperature had to be adjusted to approximate the

48

I000 'IE 800 I I I I I [ I I [ I I I I BRAWLEY -DAY r55 COTTON

~ NET RAD4ATION DAILY TOTAL - ' - - LYSIMETER E T ET ; 4 7 r a m

~ SURFACE ENERGY

600 BALANCE H P E N M ~ N

4OO

200

0 6 8 I0 r2 14 16 18

TIME (HOURS)

Z

oZ 800

i 600

i 400

o 2OO

o 20

i ] , i I

ST PAUL DAY 83 ALFAL;A

- - N E T R~Ct~TIO~ 0~ILY TOTAL - - - - - - uYSIMETER ET E T : 78mm

/ , - - , ~ : ~URRACE ENERGY B&L~'WCE

6 8 PO 12 14 r6 18 TIME (HOURS)

Fig. 3. Diurnal trend of net radiation, measured evapotranspiration and calculated evapotranspiration from the surface energy balance, Bartholic--Namken--Wiegand, and Penman--Monteith combination models for Brawley, California.

Fig. 4. Net radiation, measured evapotranspiration and calculated evapotranspiration by the surface energy balance, Bartholic--Namken--Wiegand and Penman--Monteith combination models for alfalfa at St. Paul, Minnesota.

composi te scene temperature . This was done by averaging the soil surface area of the lysimeter they represented. Using only crop canopy temperatures, which were much cooler than the soil surface, the predicted evapotran- spiration values were too high because the model assumes that the entire surface area is at this temperature . This adjustment was made for each location before the calculation of evapotranspiration with any of the methods. Soil heat flux was not measured in this s tudy and would have to be accounted for if the model is to be applicable over large areas. The deletion of this componen t does not appear to introduce a large error, but it does int roduce a bias in the data. Composite scene canopy temperatures would be of more use than individual plant or leaf temperatures in the estimation of evapotranspiration.

Alfalfa evapotranspiration at St. Paul exhibited a condit ion in which the measured values were larger than the estimated values by the Bartholic-- Namken--Wiegand method (Fig. 4). The measured and calculated values from the surface energy balance me thod agreed very closely. For this day the measured evapotranspiration was above net radiation for the early and late part of the day and below net radiation during midday. At all times the surface energy balance model responded to the environmental conditions to estimate the measured evapotranspiration values very closely.

At Temple, the estimated evapotranspiration with the surface energy balance agreed very closely with measured values (Fig. 5). As was found in the other locations the Bartholic--Namken--Wiegand method closely followed the net radiation curves. This effect is also evident at Davis (Fig. 6) where the measured values and those predicted by the surface energy balance agreed very closely th roughout the day while the Bartholic--Namken--Wiegand method, in responding predominant ly to net radiation, was not affected by the late a f te rnoon sea breeze which caused an increase in the evapotran- spiration rate. This increase in evapotranspiration was affected by the

49

~ IOOO q ! I I , i I [ ] I ~ I ! 2 'e DAVIS-DAY 217 TOMATO z TEMPLE-DAY 167 GRAIN SORGHUM

B - - NET RADIATION ~ DAJLY TOTAL 800 - - NET RADIATION DAILEY TOTAL oo ___ LYSIMETER E I ET: 16ram - - - - LYSI~ETER ET ET= 53ram

l ....... ........... ~--CB-N-W ~---'~ SURFACE ~ SURFACE ENERGY

ENERGI 600 BALANCE 600 H p N

yj 400 ~ 4 0 0

o ! I I I I I I I 0 6 8 ~O 12 14 16 18 2 0 z 15 8 IO ~2 14 16 18 2 0

T I M E {HOURS) T I M E {HOURS)

Fig. 5. Diurnal trend of net radiation, measured evapotranspiration, and calculated evapotranspiration from the surface energy balance, Bartholic--Namken--Wiegand, and Penman--Monteith combination models for Temple, Texas.

Fig. 6. Diurnal trend of net radiation, measured evapotranspiration, and calculated evapotranspiration from the surface energy balance, Bartholic--Namken--Wiegand, and Penman--Monteith combination models for Davis, California.

increase in w indspeed wi th on ly a slight modera t ion in air temperature and vapor pressure deficit .

For all locat ions wi th at least 80% ground cover, hour ly measured evapotranspirat ion rates were compared against values calculated from the Bartho l i c - -Namken- -Wiegand m e t h o d or the surface energy balance. For each hour the aerodynamic resistance was stabil ity corrected via eq. 4. The comparisons for the Bartho l i c - -Namken- -Wiegand mode l are s h o w n in Fig. 7 for Kimberly , Temple , St. Paul, and Davis. There is generally a g o o d fit about the 1 : 1 l ine, but the standard error is a lmost 100 J m -~ s -~ . The predict ions

',.o I 0 0 0 -:S 'E C - NAMKEN -

~_ 800

6 0 0 ~ - ~ x ~ ~ , _ x -

<z I | ~ /oo.;

400 o "o -- x x$ e

x • .•

F- xo

• O.Bl (ET)

~- 2 0 0 4 0 0 6 0 0 BOO I 000

- 2 -I MEASURED EVAPOTRANSPIRATION(Jm s )

I 000 i w

8 0 0 z o

~ 6 0 0 a_

z

4 0 0

o 2 0 0 Lu F-

a

Q.

I I i I /

/ SURFACE ENERGY BALANCE

• KIMBERLY -

- x ,x, F " x _ ~ . o t . x

- - - - ~ : ~ O ~ PRFDICTE D " 9 . 83+099 (ET )

~ r 2 • 0 92 n = 85

200 400 600 800 ,ooo MEASURED EVAPOTRANSPIRAT ION ( J m ' 2 s "1)

Fig. 7. Measured evapotranspiration from lysimeters at Kimberly, Indiana; Temple, Texas; St. Paul, Minnesota; and Davis, California, compared with calculated evapotran- spiration by the Bartholic--Namken--Wiegand model.

Fig. 8. Measured evapotranspiration from lysimeters at Kimberly, Indiana; Temple, Texas; St. Paul, Minnesota; and Davis, California compared to calculated evapotran- spiration by the surface energy balance model.

50

could also be grouped by location -- at St. Paul and Kimberly the model tended to underpredict , while at Temple and Davis there was an over- prediction. Because net radiation is the main driving factor in this model, predictions of evapotranspiration are insensitive to canopy temperature , as shown by Hatfield et al. (1983).

Measured and predicted evapotranspiration by the surface energy balance (eq. 2) showed an extremely good fit for all of the locations with full- ground cover (Fig. 8). There was not any location bias, and the fit was good for the entire range of data. From the good agreement between predicted and measured values, the stability adjustment in aerodynamic resistance appears adequate. The lack of location bias indicates that there is no need for a local wind or crop factor, which makes this model more widely applicable. The standard error about the line was about 5 0 J m -2 s -1 , which would be an acceptable error over the range of the data. There also was not an overestimation at the higher evapotranspiration rates, as was found by Heilman et al. (1976). These results show that either the surface energy balance model with canopy tempera ture and net radiation, either directly measured or calculated from its components , would provide a reliable and accurate estimate of the evapotranspiration of a cropped surface.

During the early stages of crop growth, ground cover is minimal for most crops and in many locations, full ground cover is not obtained. As seen in Table I, three locations we visited had only 15% ground cover, so we were able to evaluate the two evapotranspiration models for the partial canopy cover condit ion. Results for the Bartholic--Namken--Wiegand model are given in Fig. 9. As shown in Fig. 2, this model overestimates the measured evapotranspiration rate at all values above 2 0 0 J m -2 s -1 . Below this level the measured values are very closely tied to the net radiation. At early morning and late evening hours the evapotranspiration is closely tied to radiation availability and the soil and canopy temperatures are relatively close to air temperature . During midday, however, the evapotranspiration process over the entire surface is low or negligible for the soil as compared to the crop and the relative disagreement decreases as the fraction of soil cover decreases.

Agreement between measured and calculated evapotranspiration by the surface energy balance method, when composi te surface tempera ture was used, was much bet ter than the Bartholic--Namken--Wiegand method for partial canopy cover (Fig. 10). When only crop canopy tempera ture was used for these sites the slope was 1.7 with R 2 = 0.51. There is still considerable scatter about the 1 : 1 line, and the model tends to overpredict the measured amounts. This overpredict ion could be accounted for in eq. 2 if the soil heat flux term were included in the analysis. It is possible that this problem could be overcome by using the change in the soil surface tempera ture from one hour to the next to approximate a change in heat storage. This aspect will require fur ther investigation before the model can be applied th roughout a growing season. Measured composi te temperatures would possibly improve the model over the weighted average which was utilized in this study.

51

I000

8OO

G: E_ 600

400

w u~ 2OQ

I-

c3 # o (3_

I I I I / B : B R : : Ok L IC-NAMKEN-WIE GAND /

"05o • O/ PREDICTED =23 21+152(ET}

x~eX I Sle. = 97.55 X n =6#

200 400 600 800 1000

MEASURED EVAPOTRANSPIRATION (Jrn-2s -I)

f ~- ,ooo 'E

z 800

E m 6 0 0

4OO

2OO

C

[ I I ] / " I SURFACE ENERGY B A L A N C E /

t: ,Z- x FARGO o LqNCOLN

RE DICTED = 5578+ ,92 (E

r 2" 0 55

n = 6~ x xo r?2

o

200 400 600 800 I000

MEASURED EVAPOTRANSPIRATION (Jrn-2s -I)

Fig. 9. Measured evapotranspiration from partial canopy cover crops on lysimeters at Brawley, California; Fargo, North Dakota; and Lincoln, Nebraska compared to calculated evapotranspiration by the Bartholic--Namken--Wiegand model.

Fig. 10. Measured evapotranspiration from partial canopy cover crops on lysimeters at Brawley, California; Fargo, North Dakota; and Lincoln, Nebraska compared to calculated evapotranspiration by the surface energy balance model.

SUMMARY AND CONCLUSIONS

Canopy tempera ture as an input to surface energy balance models provides a me thod for estimating actual evapotranspiration from a surface. Surface energy balance models proposed by Brown and Rosenberg (1973) and Soer (1980) per formed well for all locations and crops in this study. The best agreement between measured and predicted evapotranspiration rates was for full canopy cover, although a composi te surface temperature (soil and plant) improved the fit between predicted and actual evapotranspiration values. Diurnal plots of the data showed that the Bartholic--Namken--Wiegand method was driven mainly by net radiation and had a large error in estimating actual evapotranspiration, particularly in situations with partial canopy cover.

Inclusion of canopy tempera ture with other meteorological data does not make a complete remote sensing approach but it does allow for dynamic coupling between the plant and atmosphere. Ground-based net radiation, air temperature and windspeed are still needed as model inputs. Any col- lection of surface tempera ture f rom an air or spacecraft would represent an instantaneous evapotranspiration rate and this technique would have to be applied over a complete day to improve the utility of the information for agricultural management. Jackson et al. (1983) proposed a method based on latitude, t ime-of-day and t ime-of-year which adjusts one time-of- day measurements to daily totals. This approach, along with an estimation of seasonal changes in the canopy aerodynamic properties, will have to be evaluated over a complete growing season to provide a rigorous test of the

52

m o d e l ' s l i m i t a t i o n s . T h e s e r e s u l t s i n d i c a t e t h a t t h e s u r f a c e e n e r g y b a l a n c e w i t h c a n o p y t e m p e r a t u r e i n p u t s is an a c c u r a t e a n d r e l i a b l e m e t h o d o f o b t a i n i n g a c t u a l e v a p o t r a n s p i r a t i o n f r o m a c r o p su r f ace .

ACKNOWLEDGEMENTS

We e x p r e s s o u r g r a t i t u d e t o t h e f o l l o w i n g i n d i v i d u a l s f o r t h e i r c o o p e r a t i o n a n d a s s i s t a n c e : C. E h l i g ( B r a w l e y ) , J . T. R i t c h i e ( T e m p l e ) , E. T. K a n e m a s u ( M a n h a t t a n ) , B. L . B l a d ( L i n c o l n ) , D. G. B a k e r (S t . P a u l ) , L. J . B r u n ( F a r g o ) , J . K. A a s e ( S i d n e y , M o n t a n a ) , J . L . W r i g h t ( K i m b e r l y ) , a n d W. O. P r u i t t (Dav i s ) . W i t h o u t t h e h e l p o f t h e s e i n d i v i d u a l s a n d t h e i r a l l o w i n g us t o use t h e i r f a c i l i t i e s , t h i s w o r k w o u l d n o t have b e e n p o s s i b l e .

REFERENCES

Bartholic, J. F., Namken, L. N. and Wiegand, C. L., 1970. Combination equations used to calculate evaporation and potential evaporation. USDA-ARS-Bull. 41- -170 .14 pp.

Blad, B. L. and Rosenberg, N. J., 1976. Evaluation of resistance and mass transport evapotranspiration models requiring canopy temperature data. Agron. J., 68: 764--769.

Bowen, I. S., 1926. The ratio of heat losses by conduction and by evaporation from any water surface. Phys. Rev., 27: 779--787.

Brown, K. W. and Rosenberg, N. J., 1973. A resistance model to predict evapotran- spiration and its applicat ion to a sugar beet field. Agron. J., 65 : 341--347.

Hatfield, J. L., 1983. Evapotranspiration obtained from remote sensing methods. In: D. E. Hillel (Editor), Advances in Irrigation, Vol. 2, Academic Press, New York, pp. 395--416.

Hatfield, J. L., Smietana, P. J., Jr., Carroll, J. J., Flocchini, R. G. and Hamilton, R. H., 1981. Solar Energy: Implementat ion of a research and training site at Davis, California. Land, Air and Water Resources Paper 10005. University of California. 259 pp.

Hatfield, J. L., Perrier, A. and Jackson, R. D., 1983. Estimation of evapotranspiration at one-time-of-day using remotely sensed surface temperature. Agric. Water Manage., 7: 341--350.

Heilman, J. L. and Kanemasu, E. T., 1976. An evaluation of a resistance form of the energy balance to estimate evapotranspiration. Agron. J., 68: 607--611.

Heilman, J. L., Kanemasu, E. T., Rosenberg, N. J. and Blad, B. L., 1976. Thermal scanner measurement of canopy temperatures to estimate evapotranspiration. Remote Sensing Environ., 5: 137--145.

Jackson, R. D., Reginato, R. J. and Idso, S. B., 1977. Wheat canopy temperatures: a practical tool for evaluating water requirements. Water Resour. Res., 13: 651--656.

Jackson, R. D., Hatfield, J. L., Reginato, R. J., Idso, S. B. and Pinter Jr., P. J., 1983. Estimation of daily evapotranspiration from one time of day measurements. Agric. Water Manage., 7 : 351--362.

Monteith, J. L., 1973. Principles of Environmental Physics. American Elsevier, New York, 243 pp.

Penman, H. L., 1948. Natural evaporation from open water, bare soil, and grass. Proc. R. Soc. London, Ser. A, 193: 120--145.

Sequin, B. and Itier, B., 1983. Using midday surface temperature to estimate daily evaporation from satellite thermal IR data. Int. J. Remote Sensing., 4: 371--384.

Soer, G. J. R., 1980. Estimation of regional evapotranspiration and soil moisture condit ions using remotely sensed crop surface temperatures. Remote Sensing Environ., 9: 27--45.

53

Stone, L. R. and Horton, M. L., 1974. Estimating evapotranspiration using canopy

temperatures: field evaluation. Agron. J., 66: 450--454. Sumayao, C. R., Kanemasu, E. T. and Brakke, T. W., 1980. Using leaf temperature to

assess evapotranspiration and advection. Agric. Meteorol., 22: 153--166. Van Bavel, C. H. M., 1966. Potential evaporation: the combination concept and its

experimental verification. Water Resour. Res., 2 : 455--467. Verma, S. B. and Barfield, B. J., 1979. Aerial and crop resistances affecting energy

transport. In: B. J. Barfield and J. F. Gerber (Eds.), Modification of the Aerial En- vironment of Crops. ASAE, pp. 230--248.

Verma, S. B., Rosenberg, N. J., Blad, B. L. and Baradas, M. W., 1976. Resistance--energy method for predicting evapotranspiration: Determination of boundary layer resistance and evaluation of error effects. Agron. J., 68: 776--782.