645 
called ’model initialization’, crop parameters are estimated from optical remote sensing and ’fed’ into a growth 
model as input or forcing function. Mostly, crop parameters that have been used succesfully so far are measures 
for the fractional light interception by the canopy, namely LAI and soil cover (Steven et al., 1983; Kanemasu 
et al., 1984; Maas, 1988; Bouman & Goudriaan, 1989). However, parameters like LAD and leaf optical properties 
can also be used as input in more elaborate growth models. 
In the second method, called ’model calibration’, crop growth models are calibrated on time-series of remote 
sensing measurements. Maas (1988) presented a method in which crop growth model parameters were adjusted 
in such a way that simulated values of LAI by the growth model matched LAI values that were estimated from 
reflectance measurements. Bouman (1991, 1992) developed a procedure in which remote sensing models (a.o. 
optical reflectance) were linked to crop growth models so that canopy reflectance was simulated together with 
crop growth. The growth model was then calibrated to match simulated values of canopy reflectance to measured 
values of reflectance. The calibration in this procedure is governed by the parameters which link the crop growth 
model and the remote sensing model (LAI, LAD and leaf optical properties). 
2.2 Models 
2.2.1 SUCROS. In this study the used crop growth model was SUCROS (Simple and Universal CROp growth 
Simulator; Spitters et al., 1989). It is a mechanistic growth model describing the potential growth of a crop as 
a function of irradiation, air temperature and crop characteristics. The light profile withina crop canopy is computed 
on the basis of the LAI and the extinction coefficient. At selected times during the day and at selected depths 
within the canopy, photosynthesis is calculated from the photosynthesis-light response of individual leaves. Integration 
over the canopy layers and over time within the day gives the daily assimilation rate of the crop. Assimilated 
matter is used for maintenance respiration and for growth. The newly formed dry matter is partitioned to the 
various plant organs. An important variable that is simulated is the LAI, since the increase in leaf area contributes 
to next day’s light interception and thus rate of assimilation. 
2.2.2 SAIL Model. The one-layer SAIL radiative transfer model (Verhoef, 1984) simulates canopy reflectance 
as a function of canopy variables (LAI, LAD and leaf reflectance and transmittance), soil reflectance, ratio 
diffuse/direct irradiation and solar/view geometry (solar zenith angle, zenith view angle and sun-view azimuth 
angle). Recently, the SAIL model has been extended with the hot spot effect (Looyen et al., 1991). 
2.2.3 PROSPECT Model. The PROSPECT model, as developed by Jacquemoud & Baret (1990), is a radiative 
transfer model for individual leaves. It is based on the generalized plate model of Allen et al. (1969,1970), which 
considers a compact theoretical plant leaf as a transparent plate with rough plane parallel surfaces. An actual 
leaf is assumed to be composed of a pile of N homogeneous compact layers separated by AM air spaces. The 
discrete approach can be extended to a continuous one where N need not be an integer. PROSPECT allows to 
compute the 400-2500 nm reflectance and transmittance spectra of very different leaves using only three input 
vanables: leaf mesophyll structure parameter N, chlorophyll content and water content. The output of the PROSPECT 
model can be used directly as input into the SAIL model. 
2.3 Crop Parameter Estimation 
2.3.1 LAI. Clevers (1988, 1989) has described a simplified, semi-empirical, reflectance model for estimating 
LAI of a green canopy. First, a WDVI (Weighted Difference Vegetation Index) is ascertained as a weighted difference 
between the measured NIR (r J and red reflectances (r r ) in order to correct for soil background: 
WDVI = r L . - C . r„ (1) 
with C as the ratio between NIR and red reflectance of the underlying soil. 
Then, this WDVI is used for estimating LAI according to the inverse of an exponential function: 
LAI = -1/« . ln(l - WDVI/WDVI.), (2) 
with a and WDVI, as two empirical parameters. Bouman et al. (1992) arrived at the same formulation of the 
relationship between LAI and WDVI through a similar line of reasoning. They empirically found consistent parameters 
for various years, locations, cultivars and growing conditions for some main agricultural crops.