Measured rainfall data from 33 ground-based rainfall stations were compared with rainfall estimates from CHIRPS (Climate Hazards Infrared Precipitation with Stations) across a rainfall gradient in central Namibia. There was close agreement between the two datasets across the interior of the country from the escarpment eastwards. However west of the escarpment the two datasets diverged. In this zone all CHIRPS estimates were higher than measured values and the seasonal variability of CHIRPS estimates declined towards the coast whereas measured rainfall variability rose. Quality assessments of CHIRPS in the literature have suggested there is a tendency for the model to overestimate the frequency of rainfall events, and to record low rainfall rather than zero rainfall in low rainfall areas. These effects may be exacerbated in Namibia by the prevalence of coastal fog. Increasing the number of reliable ground-based stations across the coastal zone may go some way to addressing the discrepancy in Namibia between CHIRPS estimates and ground measurements. Keywords: CHIRPS, Namibia, rainfall, satellite data, weather station data.