Diouf, A.Lambin, E. F.2016-04-192016-04-192001Journal of Arid Environments 48(2): 129-1480140-1963http://hdl.handle.net/10919/65960Metadata only recordDryland degradation rarely translates into linear, declining trends in vegetation cover due to interannual climatic variability. Appropriate indicators of land-cover modifications need to be defined for semi-arid regions. Our hypothesis is that degradation can be measured by: (1) a decrease in the resilience of vegetation to droughts; (2) a decrease in rain-use efficiency; and (3) a modification of floristic composition. The objective of this paper is to test the relationships between a remotely sensed indicator of vegetation, rainfall data and field measurements of biomass and floristic composition. The study was based on field measurements of vegetation conditions covering a period of 10 years, in the semi-arid region of the Ferlo in Senegal. Our results indicate that land-cover modifications in the Ferlo are best measured by changes in rain-use efficiency. No consistent trend in the relative abundance of grass species was visible at the scale of the decade, even on the two sites affected by degradation. Just after a drought, a given increase in rainfall results in less biomass production than is the case for normal years. ---Academic Presstext/plainen-USIn CopyrightDesertificationDroughtSemiarid zonesSoil degradationSoilRemote sensingLand cover changesLand degradationSahelLand useRainBotanical compositionMonitoringTemporal variationRain-use efficiencyEcosystemAbstractCopyright 2001 Academic Press