Land surface conditions of the Niamey region: ecological and hydrological implications

Abstract

A land surface map of the region of Niamey (2–3°E, 13–14°N), Niger, derived from a classification of SPOT multispectral sattelite data with 20 × 20-m resolution, is described. Sixteen distinct classes were defined based on landform, vegetation soil crust type and land use. The classification system was designed to emphasize the role of these factors in hydrological and ecological processes. The primary purpose was to provide information for integrating local-scale measurements of land-atmosphere interactions in HAPEX-Sahel (Hydrological and Atmospheric Pilot Experiment) up to a region 1° × 1° square, and to assist hydrologists in interpreting the spatial distribution of soil moisture. A large number of ground descriptions were integrated with high spatial resolution multispectral satellite reflectance data. Since most of the hydrological and atmospheric observations in HAPEX-Sahel were carried out in only three land cover types (tiger bush on the plateaux, fallow savanna and millet fields in the sand valley or plains), the 16 classes were merged successively degree of nine then six so that three final maps were produced, each one reflecting a different degree of aggregation. Some difficulties could not be resolved. One stemmed from the continuum linking land left fallow and cultivation. This uncertainty can be a problem since bush-fallow and cropped fields are hydrologically and ecologically quite distinct. Another major difficulty was the detection of sparse vegetation cover using the satellite data. Each surface condition class was assigned a range of runoff capability parameters derived from numerous rainfall simulation tests conducted in the Sahelian zone. Consequently, runoff production could be assessed for each 20 × 20-m pixel. However, due to the endoreic nature of the watersheds in the region, the use of the map to predict runoff production must be restricted to small areas ranging from the pixel size to catchments of 1–10 km².