g/m 2 /VI unit) in terms of the 
photosynthetic size of the canopy 
(Wiegand et al., 1989; Wiegand and 
Richardson, 1990a,b). The numerical 
value of ey is similar for the 
indices GVI and PVI, and for NDVI and 
TSAVI became those VI pairs have 
similar magnitudes. It is evident from 
the discussion of individual terms in 
Eq. [3] that ey will be site 
dependent because the second and third 
right side terms are. 
For given production areas, the Y(zVI) 
relation can be calibrated using data 
from years of both good and poor 
growing conditions and be used to 
predict yields in other years. Landsat 
data have been available since 1972 and 
can be paired with historical yield 
observations to develop the 
calibrations. 
Fig. 4 displays the seasonal pattern in 
observed PVI and FPAR for an experiment 
conducted in 1989 at Weslaco, TX, with 
maize planted at three densities: 7.7, 
5.4, and 3.1 plants/m 2 (Wiegand et 
al., 1991). The two higher planting 
densities absorbed about 80% of the 
incident PAR and achieved maximum PVI's 
on DOY 129 as tasseling began. We did 
not measure L because Eq. [1] shows it 
is not necessary. From emergence 
through tasseling FPAR(PVI) was 
expressed by 
FPAR=-0.015+0.036(PVI), r 2 =0.956. [4] 
The daily cumulations of PVI and FPAR 
values from plant emergence to 
physiological maturity of the grain 
(Fig. 4), the observed grain yield (Y), 
and total aboveground phytomass 
increase (ADM) were used to obtain the 
values of each term of Eq. [3] shown in 
Table 1. Because the functional 
relation for all terms in Eq. [3] were 
linear when the VI used was PVI, the 
slopes of the functional relations 
i.e., the efficiency terms ey, ea 
and ec, respectively, are closely 
approximated by the ratio of the 
dependent variable treatment mean to 
the independent variable treatment mean 
as in the column headings of Table 1. 
In Table 1 ey is in the same order as 
ea. This same finding was reported 
by Wiegand et al. (1989) for rice. In 
that study the variation in the ZAPAR(E 
PVI) term was about 15% of the mean 
value for all treatments, compared with 
5% variation among treatments for the 
other two right side terms, so that it 
dominated the Y(EPVI) term. However, 
use of EVI as a practical estimator of 
per field or per production area yield, 
once the calibrations are developed, 
does not hinge on which right side term 
affects it most. That can be 
determined, to satisfy academic 
interests, in intensive plot studies 
such as the corn and rice studies 
reported here. If the site 
dependencies of the right side terms 
were known, the left side site 
dependencies would be predictable and 
would not have to be calibrated by 
production areas (Wiegand et al., 
1991) . 
Evapotranspiration and Végétation 
Ludica 
The most limiting factor in crop yields 
for important production areas of the 
world, such as the Great Plains of the 
U. S. and Canada, is water. Cumulative 
seasonal évapotranspiration (E ET, mm) 
can be related to spectral observations 
(Wiegand and Richardson, 1990a) by 
E ET(Z VI )= E ET(DM) xDMfeVI). [5] 
This equation recognizes the dependence 
of seasonal aboveground dry matter (DM) 
on seasonalE ET and the sensitivity of 
foliage production, observable by VI, 
to water stress. [For portions of the 
season, such as drying cycles, the 
equation can be stated in terms of A ET 
and A DM corresponding to differences in 
cumulative ET and in DM between the 
start and end of the time interval]. 
In the simplest terms, Eq. [5] is 
plausible because the photosynthesis 
and transpiration processes are both 
functions of the same live tissue that 
VI measures and both are driven by 
solar radiation. Eq. [5] is not 
applied in this paper because we lacked 
the water use data, but it is presented 
for completeness. 
Assumptions and Applicability of SCA 
Eqs. [1], [2], [3], and [5] constitute 
spectral components analysis (SCA). 
SCA assumes implicitly, in agreement 
with agronomic and crop modeling 
experience, that (a) plant stands 
integrate the growing conditions 
experienced and express the net 
assimilation through the canopies 
achieved, (b) stresses severe enough to 
affect economic yield will be 
detectable through their effects on the 
development and persistence of 
photosynthetically active tissue in the 
canopies, (c) high economic yields 
cannot be achieved unless plant 
canopies are achieved that fully 
utilise available solar radiation just 
prior to and during the reproduction 
period, (d) vegetation indices 
calculated from remote observations in 
appropriate wavelengths effectively 
measure the photosynthetic size of the 
canopies, and (e) commercial producers 
use locally recommended agronomic 
practices (cultivars, planting rates 
and configurations, fertiliser and