ENVIRONMENTAL MONITORING AND CROP FORECASTING IN THE SAHEL THROUGH THE USE OF NOAA NDVI DATA - A CASE-STUDY - NIGER 1986-89

Several investigations have shown that NOAA NDVI data accumulated duri ng a rainy season can be related to total rainfall or final primary pr oductivity in the Sahel. However, serious problems can arise when look ing for quantitative relations to monitor and forecast crop yield from NDVI values. Geographical variability can affect such relations, whil e the use of data taken from a whole season is impractical for forecas ting. The present paper proposes a complete methodology of NDVI data p rocessing which only utilizes NOAA AVHRR scenes from the first part of successive rainy seasons. A series of basic corrections are first app lied to the original data to obtain reliable NDVI maximum value compos ites at the middle of the rainy seasons considered. Next, the variabil ity in land resources is accounted for by means of a standardization p rocess which normalizes the mean NDVI levels of some areas on the rele vant multi-temporal averages and standard deviations. In this way, goo d estimates of the actual condition of vegetation can be obtained in r elation to the local seasonal trend. The methodology was applied to th e Sahelian sub-departments of Niger with data from four years (1986-19 89). The most interesting result achieved concerns the estimation of f inal grain (millet and sorghum) yield for the sub-departments by the e nd of July with a mean error of about 0.08 T ha(-1). This timely evalu ation could be of great utility in the context of an efficient drought early warning system.