633 
850 900 
lS data 
850 900 
for leaf 
principal plane, 
hese simulation 
term of r.m.s.e. 
not higher than 
l better. Beside, 
lendicular plane 
any geometric 
eflec tances. The 
September. The 
nodel on ASAS 
es according to 
aits, for the two 
reproduce, with 
il effects should 
Wav«l*ngth 
Figure 5 r.m.s.e of limited sampling inversion 
results as a function of wavelength 
- Cotton — Bare soil 
Figure 6 SPOT reflectances on the 8 vs reflectances on Figure 7 Modeled versus measured reflectances over the 
the 7 for the 3 sites sites, x September 7o September 8 6 
6. Conclusion 
In this study, we used three different data sets to test the possibility of retrieving surface bidirectional reflectance factor 
from remotely sensed measurements. The first step was to use these bidirectional reflectance measurements to retrieve 
parameters of several bidirectional reflectance models. The inversion procedure, aimed at minimizing the distance 
between measured and simulated reflectances, appeared to be very efficient, and generally, there was no problem to make 
the simulated reflectances look like the measured reflectances. The retrieval of surface or canopy characteristics, though 
mere delicate, seemed also possible. 
The input data set provided to the inversion was then limited to a single plane, to test the ability of the model to predict 
reflectances in a different direction. This test showed that, if this operation seemed feasible when the sampled plane is the 
solar principal plane, the retrieval of realistic parameters seemed difficult or impossible, with measurements off principal 
plane. The last part of this study is concerned with the comparison between simulated and measured reflectances, using 
two differents instruments, SPOT and ASAS, and appears most promising for this kind of application, especially when 
considering inter calibration problems. 
Acknowledgments 
This work is a contribution to an EOS interdisciplinary Investigation (NAGW-2425). It was supported by the United 
States Department of Agriculture, French “Ministère de la Recherche et de l'Espace”, “Progamme National de 
Télédétection Spatiale” and “Centre National d'Etudes Spatiales”. We also acknowledge the helpful contributions and 
comments of S. Moulin (LERTS-CNES/CNRS, TOULOUSE, FRANCE), K. Thome and D. Gellman (University of 
Arizona Optical Sciences Center, TUCSON.AZ. USA).