The resolution of modern Global and Regional Climate Models is sufficient for many purposes. Nevertheless, the downscaling of model output is necessary when a higher spatial resolution is needed for smaller scale applications. This article introduces intermediate results of an approach for climate regionalization using the free and open source GIS software SAGA, applied on historical and future REMO regional climate model scenarios for the Okavango River Basin of southwestern Africa. The approach makes use of an SRTM digital elevation model and regionalizes temperature via a linear elevation correction/polynomial fit and precipitation with a geographically weighted regression. The results are then compared to the WorldClim interpolated station dataset. For four locations of interest without historical climate observations, time series are generated. Keywords: downscaling, elevation correction, geographically weighted regression, regional climate.