function of wavelength. Aircraft data 
- Hapke -- Verstraete Dering .. SAIL 
Figure 4.c Same as figure 4.a except for leaf Figure 4.d Same as figure 4.b except for leaf 
transmittance transmittance 
we conducted here consisted in retrieving model parameters with measurements acquired only in the principal plane, 
simulating measurements in different planes (perpendicular and SPOT acquisition plane) and comparing these simulation 
with original measurements, available over cotton and bare soil. The comparison has been evaluated in term of r.m.s.e. 
and is shown on figure 5. It can be seen on this graph that the level of uncertainty given by this method is not higher than 
the one we got when using the complete data set for the inversion. In fact, for bare soil, the result is even better. Beside, 
the retrieved parameters in both cases (partial and complete sampling) show very good agreement. 
On the other hand, another attempt was made in reverse order, that is retrieving parameters using perpendicular plane 
measurements and simulating principal plane measurements, but the results appeared to be discouraging. 
5.3 Comparisons 
The retrieved parameters allow to simulate the bidirectional reflectance of the selected surfaces in any geometric 
configuration. This can reveal most helpful to emphasize the bidirectional effects observable on SPOT reflectances. The 
table 3 gives the geometric configurations corresponding to the two images avalaible on the 7 and 8 of September. The 
figure 6 shows the difference between the surface reflectances acquired on these two days. 
Table 3 geometric configurations for the SPOT acquisitions on September 7 and 8 
Solar zenith 
View zenith 
Relative azimuth 
September 7 
30 
25 
124 
September 8 
31.4 
9.7 
43 
We then computed simulated reflectances, using parameters retrieved from the inversion of Rahman model on ASAS 
measurements, simulating ASAS reflectances in SPOT configuration and combining these reflectances according to 
SPOT spectral response. The results of these computation is compared with original SPOT measurements, for the two 
dates and the three sites, on the figure 7. The agreement is very good, and allow to assume that if we can reproduce, with 
a simple bidirectional reflectance model, the variation observed in the SPOT reflectances, the directional effects should 
be the main source of variation. 
632