1062 
2. THEORY AND MODEL 
2.1 Basic concept 
Plant water status is usually closely related to the soil water status. Even flooded rice plants, however, can not 
always transpire at the "potential" rate. Nevertheless, we can easily imagine some extreme situation such as 
potentially transpiring plants on a completely dry soil-surface or physiologically depressed (i.e., decreased 
transpiration) ones over a completely wet soil-surface. Although the term "potential" transpiration is often used, 
it is not well-defined, because it is determined by the interaction of the physiological status of plants and 
physical environmental factors. 
Therefore, it may be better to define a meteorological "potential" transpiration based on the available 
energy for vaporization at the plant leaves; that is, healthy plants without water shortage are transpiring at the 
potential rate which is mainly regulated by the available energy. This energy limited transpiration has often 
been estimated by a linear function of absorbed solar radiation (Makkink 1957) or absorbed net radiation 
(Priestley and Taylor 1972). Such "radiation methods" have been widely used for the estimation of canopy 
transpiration (Brutsaert 1982). Sakuratani (1987) reported a close correlation between absorbed solar radiation 
and transpiration of well-watered soybean plants. 
The total radiation energy absorbed is a function of the incoming solar radiation (Rs) and the interception- 
ability of the canopy. The intercepted solar radiation (IRs) has been analyzed as an exponential function of leaf 
area index (LAI) using the relationship (Monsi and Saeki 1953), 
IRs = Rs [1 - exp (-k LAI) ] (1) 
where k is an extinction coefficient that is a function of both canopy architecture and spectral features of leaf 
sections. Also, the intercepted or absorbed photosynthetically active radiation (APAR) has been related to LAI 
in a similar manner, APAR = PAR [1 - exp (- k LAI) ], where PAR is the incident photosynthetically active 
radiation (Asrar et al. 1984). 
On the other hand, a number of papers have reported a close relationship between LAI or biomass and 
remotely sensed spectral reflectance values such as vegetation indices (VI; e g., Wiegand et al. 1991). From a 
theoretical point of view, however, the radiation balance of a canopy may be estimated more directly from 
remotely sensed spectral data than from the leaf area index, because effects of other factors such as canopy 
architecture, the spectral signature of the leaf, and viewing/illumination geometry can be included in die spectral 
measurements. In fact, APAR was shown to be well estimated by spectral reflectance measurements (Asrar et 
al. 1984, Gallo et al. 1985). Furthermore, Pinter (1993) showed that the relationship between fraction APAR 
(fAPAR) and spectral indices derived from reflectance measurements was independent of solar zenith angle or 
time of day, though the actual values of these parameters varied with time of the day. His results suggest that 
the use of multispectral vegetation indices (such as SAVI, soil adjusted vegetation index) to estimate the 
proportion of incident solar energy absorbed by a plant c ommuni ty is promising and has biologically significant 
meaning. 
On the basis of the above consideration, we assume that [1 - exp (-k LAI)] can be expressed as a linear 
function of the spectral vegetation index, slapping the estimation of both k and LAI values; that is, 
fAPAR = 1 - exp (- k LAI ) oc VI (2). 
Considering the radiation balance of a canopy, (me can define the radiation absorbed by the canopy (ARs) 
as, 
ARs = Rs [ 1 -r - (1 - r 0 ) exp (-k LAI) ] (3) 
where r and r Q are the reflectances of the canopy and soil, respectively. The value of r was found to be related to 
the reflectance r f of a full-cover canopy and r Q of bare soil, respectively, 
r = r f -(r f -r„)exp(-k'LAI) (4) 
where k' is a constant (ET research group, 1967). The value of k' was found to be s imil ar to k. Moreover, the 
effects of both r Q exp (-k LAI) and r Q exp (-k' LAI) are small, and the difference between these two terms is 
negligible. Thus, we can obtain the following equation by substituting Eq.4 into Eq3. 
ARs = Rs (1 - r f ) [1 - exp (- k LAI) ] = Rs (1 - r f ) fAPAR = Rs a'SAVI (5)