Frequent wildfires emit large amounts of black carbon (BC) into the atmosphere in the semiarid regions of the African continent. This atmospheric BC efficiently absorbs shortwave radiation and thus modifies the climate system on a regional scale. Therefore, it is essential to understand how geographical distribution patterns of BC emissions are controlled by climate and vegetation in these regions. We applied a principal component analysis (PCA) to the correlations between dry season BC emissions observed by satellite and climate variables during the vegetation growing and dry seasons, and to correlations between BC and the leaf area index during the growing season, as independent values. We analysed the burned fraction (BF) in the same way, but its factor loadings did not differ significantly from those of BC in sign or magnitude. During the growing season, the response pattern of vegetation productivity (an index of wildfire fuel loading) to climate variables explained 57.5% of the regional variability in BC emissions. This vegetation productivity was more closely correlated with the geographical distribution patterns of BC emission than climate variables such as temperature during the dry season. The response pattern of vegetation productivity to climate during the vegetation growing season was roughly determined by vegetation parameters such as biome type and tree cover, which are heterogeneously distributed in Africa. Therefore, regional BC emission patterns would differ even if climate change occurred uniformly throughout semi-arid Africa. © The authors 2013.