1067 
eddy-correlation measurements on site. The a estimate for WG was 1.47,which was close to the a value for 
the alfalfa crop. For most of the growing season, the test site was characterized by large expanses of hot, dry 
soil with sparse vegetation; thus, an equation accounting for the additional resistance induced by the exposed 
soil as a function of windspeed and surface temperature (Kustas et al. 1989) was introduced for the estimation 
ofr a 
Two different estimates of 
aerodynamic resistance r were used for 
J 3 
the calculation of these parameters. It 
was apparent that the water stress was 
overestimated using r without excess 
resistance and was more reasonable when 
the excess resistance was included. 
Values of Tr , Tr and ET for all 
a p a 
measurement days at the site were 
compared (Fig.5). As expected, the Tr a 
values were lower than ET values in 
a 
nearly all cases. The Tr p values were 
also lower than ET a , especially in the 
early growing season with lower 
vegetation amount. However, it is 
difficult to make any conclusions about 
the feasibility of the application to the 
rangeland site due to the questionable 
methods of deriving a . It may be 
possible to derive a in a more justifiable 
manner using simultaneous 
measurements of fAPAR, albedo and 
spectral reflectances made by Daughtry 
during 1990 at WG'92 (personal 
communication, 1993) 
4.3 Conclusion 
A new concept was proposed for evaluating the potential and actual daily transpiration from vegetation based 
on remote sensing. On the basis of the initial results presented here and the technique's sound theoretical 
foundation, it is likely that this approach will be successful for evaluating potential and actual daily transpiration. 
Several points require further investigation. 1) Behavior of the coefficient a will have to be examined in 
detail from experimental and theoretical points of view, because it may not always be consistent for various 
crops and sites. This a value is, by definition, determined as a function of full-cover reflectance and the slope 
of the fAPAR and SAVI relation as long as the proportionality between the absorbed solar radiation and the 
"potential" transpiration is assumed on a daily basis. From a practical point of view, nevertheless, both the 
full-cover reflectance and the slope are not expected to vary much for most crop canopies, and that is why a 
values for alfalfa and rangeland were similar (1.5 and 1.47, respectively). 2) Several questions related to the 
estimation of CWSI values for sparsely-vegetated sites or immature crops still need to be answered; a) Which 
formulation should be used to estimate r a in computation of ET p and CWSI ?, b) How can we calculate a CWSI 
value using t (composite surface temperature) rather than t £ (plant canopy temperature) ?, or how can we derive 
t £ from t s ? It is possible that the new stress index devised by Moran et al. (1993) may provide a good solution 
for such questions. Also, high-resolution infrared thermal imagery will be useful, providing pixel-by-pixel 
temperatures over the area of soil and plant mixture (Inoue 1990). We will be able to replace the present CS WI 
by future improved stress indices. 3) The estimates of Tr p and Tr a values should be compared with "true" values 
directly measured using the stem flow gauge. Since measurements of Tr a were not available in both experiments, 
instead, we used ET values in the well-watered full-cover canopies. Therefore, these results represent simply a 
demonstration, not a verification, of the approach. New experiments in which transpiration will be measured 
directly with stem-flow gauges will be conducted for further examination of the concept. 
The potential and actual values of daily transpiration rate may be estimated over wide vegetation areas 
with simple remote sensing measurements and basic meteorological parameters. The method may also provide 
information on biomass and stress status, because it accounts for the differences in both vegetation density and 
stress response of the plant canopy 
Day of year 1992 
Fig.5 Comparison of estimated values of potential and actual 
daily transpiration (Tr and Tr a ) with measured values of 
ET at Walnut Gulch site. 
a