'orth latitude). 
m Index (NDVI) 
Is of NDVI have 
sflectances (TOA 
[well, 1990) have 
68] jim) and near 
infrared channels, 
weekly time step 
75° N to 55° S in 
ilution data and a 
(Kidwell, 1990). 
ellite change and 
channels and can 
targets have been 
t the top of the 
ing by gases and 
sorption occurs in 
anel (Fig. 2). Due 
i flat profile along 
vapor and ozone 
lore important in 
ility of aerosols is 
sume that aerosol 
nizontal visibility 
g the transect in 
22Aerosol optical thickness 
At present time, we did not have any data set of aerosol optical thickness that we can use at global scale. We 
must rely on a sensitivity study with some guidance from a few measurement campaigns. Two cases are 
tested. In the first one, we assume an aerosol optical thickness (Tp) constant in time and space, and set it to 
0.05 at 550 nm (This case is referenced as Case 1). In the second one, we assume that the aerosol optical 
thickness varies as a function of the latitude (Case 2). Tp at 550 nm is set to 0.05 at polar latitudes and 0.2 at 
the equator (Fig. 3). The choice of a low optical depth agrees with the GVI compositing method which is 
supposed to select clearest data (Faizoun and Dedieu, 1993). For Case 2, we set aerosol optical thickness for 
tropical regions which roughly agrees with in-situ measurements for clear days (e.g. Faizoun et al„ 1994). 
Figure 3: Variation of Rayleigh and aerosol optical thicknesses for AVHRRIVIS channel along the transect 
for Case 1 (left) and Case 2 (right). Solid lines and dash lines refer to Rayleigh and aerosol optical 
thicknesses respectively. 
23 Water Vapor Content 
Water vapor climatology (Oort, 1983), has been vertically integrated and used to correct TOA reflectances 
from atmospheric effects. Climatology results from compilation of data over a long time period (1958 to 
1973). Initial spatial resolution (2.5° x 5°) and temporal resolution (month) of the data have been resampled to 
the GVI resolution. Figure 4 exhibits the latitudinal variation of water vapor content. One can notice a gap in 
the data around 10° of north latitude that illustrates the difficulty to get global representative measurements, 
even on a mean of fifteen years of data. We considered that the mean value of water content given by the 
climatology can be used with relatively good confidence at global scale. The water vapor content is 
characterized by a daily variability involved by atmospheric meteorological circulation, but a seasonnal trend 
exists and roughly reproduces from year to year. 
2.4 Ozone Content 
Globally, the variability of ozone content is weak except at high latitudes where seasonnal trend exists. 
Therefore, correction of ozone absorption can be made using London et al.'s climatology (1976) which has 
been compiled from measurements made between 1957 and 1967. We used mean monthly values for 10 
degrees zonal bands, resampled to GVI resolution (Fig. 5). 
Figure 4: Variation of water wapor content along the 
transect (July) 
Figure 5: Variation of ozone content along the 
transect (July) 
35