Previous observational studies have demonstrated a correspondence between patterns of deforestation and those of cloudiness over the Amazon basin, pointing toward the potential importance of landcover heterogeneity on local climate. A numerical study using a mesoscale model over the region of Rondonia has also shown an enhancement of low-level cumulus clouds over partially cleared areas under conditions of weak synoptic forcing. In this study, the potential influence of deforestation on local climate is further tested by comparing observed patterns of shallow clouds and rainfall over regions of high forest cover variability in the Amazon basin. Shallow cumulus cloud maps are derived three times daily over the Amazon basin using a cloud recognition algorithm with GOES-8 infrared and visible image pairs. Satellite infrared rain estimates are derived using the Convective-Stratiform Technique (CST) calibrated by the TRMM Microwave Imager (TMI). The rain estimates consist of seasonally-averaged diurnal cycle of hourly rainfall. Both rainfall and cloud products have a 4-km resolution. Covariance functions between the distribution of cloud/rainfall and that of forest cover are calculated and analyzed at various scales to determine under which conditions a correlation exists between landcover variability and cloud cover or rainfall.