Sahelian ecosystems, especially those installed on sandy soils react very quickly and intensely to small climatic changes. In this semi-arid area, wind erosion which occurs mainly on cultivated millet fields and is threatening the sustainable use of land resources. Furthermore, the rapid population growth and the severe droughts of the twentieth century have had important impacts on ecosystems. To better manage this fragile environment, it is important first to understand how it works. This thesis was included in the framework of the AMMA project (African Monsoon Multidisciplinary Analyses) and was in the heart of a CORUS 2 (Coopération pour la Recherche Universitaire et Scientifique) program which aimed to study wind erosion of cultivated sandy soils in the region of Niamey. The strategy followed in this study was based on detailed experimental measurements aiming to precisely quantify and understand current environment dynamics and more precisely, wind erosion dynamics. The results of experimental measurements were used to explain the changes we observed on small catchments of two permanent lakes near Niamey: Bangou Bi and Bangou Kirey (13.51 ° N - 2, 21 ° E). The sediment cores collected in these lakes were then analyzed in light of their catchments changes during the last century. In this study, we highlighted the major role of crop residues cover in traditionally cultivated field against wind erosion. Even at extremely low cover rate, they provide efficient protection for soils. However, below a threshold of 2%, their efficiency dramatically decrease and wind erosion caused large amount of soil losses (in average 130 t per year) and is actively involved in soil crusting. It is likely that this mechanism (wind erosion due very low crop residue cover and soil crusting) occurred by the middle of eighties due to the combined effects of generalized soil cultivation and sever drought of 1985. These processes explained the large degradation of the lakes watersheds we observed from surveys and diachronic mapping between 1975 and 2009. With sediment core analyses, we proposed a model of sedimentation in the Niamey lakes. This model suggests increasing erosion since the 1980’s. From the study of sediments core collected in lakes, we proposed a model explaining sedimentation that suggests an intensification of erosion from the mid 80s.
