Inventaire par télédétection des agro-écosystèmes de bas-fonds dans le centre du Bénin

Abstract

In sub-Saharan Africa, inland valleys are types of agro-ecosystems suitable for rice cultivation and anti seasonal crops. These important resources have not been developed enough due to their extreme diversity and to the fact that their potentialities are remain poorly identified by scientists and the actors of development. The present research aimed at estimating the interest of an information system on inland valleys, using the Dassa-Zoumé area in the centre of Benin as the study area. The methodology adopted for the inventory of the inland valleys is based on the calculation of the new channel NDVI (Normalized Difference Vegetation Index) from the Landsat 7-ETM+ image of 2000 with a spatial resolution of 28.5 m and on the realization of the slope map generated from a DEM (Digital Elevation Model) of the area. A total of 181 inland valleys were identified with a total surface of 1,124 hectares. About 60 % of the valleys identified have an area comprised between 0 and 5 ha while less than 5% vary from 20 to 25 ha. The detection and inventory approach used in this study, and based essentially on the exploitation of a satellite image, shows a level of reliability that should be extended in the search of lowlands for agricultural perspectives. It facilitates savings in time in comparison to other approaches that would necessitate the surveying of all the inland valleys.