ABSTRACT
Multispectral satellite data, when properly calibrated and standardized, can be used synergistically for a quantitative analysis of land surface change. Relationships among multispectral satellite data (visible reflectance, surface temperature and polarization difference of microwave emission at 37 GHz frequency) have been used to develop hypotheses concerning the relative sensitivity of these data to varied land surface characteristics, which need to be validated by field observations. Radiative transfer models have also been developed to understand these multispectral data. Inter-annual variations of visible reflectance and polarization difference for the period 1982 to 1986 over the Sahel and the Sudan zones of Africa show a lagged response with respect to the rainfall deficit during recovery from drought, which needs to understood in terms of biophysical parameters. Changes in the NOAA satellite orbit (that is, the time of observation) have introduced significant inter-annual variation of infrared surface temperature and have also affected the reflectances to some extent (particularly over arid and semi-arid regions) because of changes in the shadows cast by the vegetation. Some changes in the observations have also occurred because of differences in the response characteristics of the AVHRR sensor on board different NOAA satellites. These sensor/satellite related changes of visible and near-infrared reflectances and surface temperature have to be removed before these data can be used quantitatively to study land surface change.
