For many African countries agriculture will remain an important engine for economic development for decades to come, and the livelihood of rural populations and the welfare of entire countries critically depend on agricultural productivity. At the same time, agricultural practices have enormous consequences for natural systems and threaten the natural resource endowment in many regions. Given the importance of agriculture for sustainable development in Africa it is paramount to develop baselines of land use to monitor and assess the natural and economic impacts of environmental change. This paper critically examines estimates of cropping patterns and cropland extent for Africa produced using various sources of remotely sensed data and compares them with non-spatial statistical inventories of cropland at the continental, regional and local scales. The analysis reveals substantial discrepancies across alternative sources of information about land cover in both the extent and location of croplands, and pinpoints shortcomings in currently available inventories of land cover and land use data derived from remote sensing. These inconsistencies have important implications for downstream analyses that use land use data and they highlight the need to strengthen technological and statistical capacity in the regions to provide the basis for