Clevers, Jan 
  
LAI = -1/o X In(1 - WDVI/WDVL,), (2) 
with o, as a parameter describing the rate with which the function of equation (2) runs to its asymptotic value, and 
WDVL. as the asymptotic limiting value for WDVI. 
Bouman et al. (1992) validated the CLAIR model for various crops. They used field-based reflectance measurements 
obtained in The Netherlands over different experimental fields during a large number of years. For instance, for winter 
wheat they found for o a value of 0.400 and for WDVI, a value of 47.0 if the WDVI is based on NIR and green 
reflectance. They found a single regression line that was not significantly different for cereals like wheat, barley and 
oats during the whole growing season. Clevers and Van Leeuwen (1996) showed that such field-based parameter 
estimates could well be applied to airborne measurements. 
Clevers (1991) found a similar regression line for barley at the vegetative stage, whereby the reflectances now were 
derived from calibrated aerial photographs. However, for the barley crop at the generative stage the regression line was 
significantly different. This was also found by others, e.g. Ahlrichs and Bauer (1983). Clevers (1991) obtained for œ a 
value of 0.252 at the vegetative stage and for WDVI. a value of 68.6 if the WDVI is based on NIR and red reflectance. 
At the generative stage he obtained for o a value of 0.530 and for WDVL. a value of 57.9. 
In the current study the parameter estimates obtained for cereals in The Netherlands will be translated to cereals in the 
South of France. Moreover, the results for field-based and airborne data will be translated to spaceborne data. 
22 Calibrating Crop Growth Model 
The mechanistic crop growth model ROTASK v 1.5 is a dynamic simulation model for the growth of a range of arable 
crops. It has a continuously running water and soil organic matter balance, and crop growth is driven by daily weather 
data (Jongschaap, 1996). Management practices like sowing, fertilization and irrigation can be executed. Potential, 
water limited and nitrogen limited growth rates are calculated on a daily basis and eventual growth rates are given 
according to the law of minimums. The model was calibrated for winter wheat and for the soil conditions of the test site 
near Avignon, using 1996, 1997 and 1998 field data (Jongschaap, 2000). 
During simulation, two methods of introducing the SPOT-LAI values were applied. 
(1) A reset of the LAI value was performed at the date of each SPOT image in the growing season. In between SPOT 
images, the model simulated the LAI development. 
(2) The LAI values were simulated until the date of the first SPOT image. Then, the interpolated LAI values between 
SPOT images were forced on the model. Finally, after the date of the last SPOT image, the LAI values were 
simulated again until the end of the growing season. 
The results were plotted against observed field values. 
3 ALPILLES FIELD EXPERIMENT 
The main emphasis within the Alpilles field experiment was on the temporal dimension of modelling the involved 
processes (both canopy functioning and SVAT). Three crops were selected for the study: wheat, sunflower and alfalfa. 
Initially, calibration and validation fields were defined for each crop type in order to calibrate and validate, respectively, 
the developed models. However, because the main year studied (1997) was a very dry year, on some fields crop failure 
occurred and there were insufficient calibration and validation fields for a sound statistical analysis. Therefore, it was 
decided to calibrate the models developed on other data than the 1997 Alpilles data, and to use these data only for 
validation of the models and algorithms. 
Remote sensing data acquisition over the test site comprised both satellite and airborne data. Satellite data were 
obtained from SPOT, Landsat-TM, NOAA-AVHRR, ERS and Radarsat. In addition, an extensive airborne data 
acquisition scheme wat set up, including optical sensors (POLDER, CIMEL, DAIS, MAIS), thermal sensors (Heinman, 
Inframetrics, DAIS, MAIS) and microwave sensors (ERASME, RENE, IROE). For a more extensive description see 
Baret et al. (1999). 
3.1 Field Measurements 
During the Alpilles experiment, canopy measurements were carried out over the crop cycle of the wheat fields with time 
intervals varying between 8-10 days (calibration and validations fields: 101, 120) and 18-20 days (remote sensing 
fields: 208, 210, 300). At each date, 6 samples of 0.50 m by 3 rows (between 0.21 and 0.26 m?) were used to determine 
  
274 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000.