The spatial patterns of correlations between the convective boundary layer (CBL) temperature and both vegetation and upper tropospheric temperature over Africa have been detected by a singular value decomposition (SVD) analysis of monthly data for the period 1981--1994. The data used in the present study consist of the National Oceanic and Atmospheric Administration (NOAA)-derived normalized difference vegetation index (NDVI) and the 850 hPa/200 hPa potential temperatures from the National Centers for Environmental Prediction (NCEP) objective reanalysis. In general, greater-than-normal vegetation over the Sahel is related to lower-than-normal CBL temperature aloft, whereas less vegetation corresponds to higher CBL temperature aloft. This vegetation-CBL correlation was extracted as the first SVD mode that has centers of action concentrated over the Sahel. The correlation is strongest during the late dry season (especially during February and April) and is weaker during the rainy season. Typical correlation is observed during April when marked cooling of the CBL occurs in concurrence with positive vegetation anomalies. These anomalies are likely to be produced by an occasional premonsoon rain resulting from a cloud band that propagates from the midlatitudes. During the rainy season the CBL temperature is more strongly associated with upper tropospheric temperatures, which are likely modulated by large-scale circulation, than with the Sahelian vegetation. ¿ 2000 American Geophysical Union