Third generation rainfall climatologies: satellite rainfall and topography provide a basis for smart interpolation

This short paper discusses a technique for producing ‘improved rainfall estimates’ (IRE) for Africa. The central objective of IRE is to make reasonable use of the high resolution mean fields, moderate density gauge data, and timely satellite rainfall data typically available at the national/regional level in Africa. In general, traditional rainfall interpolation schemes can be classified into two coarse categories. In the beginning, raw station observations were interpolated by combining the values of surrounding stations. Later, 2nd generation techniques interpolated anomalies from long terms means, thereby reducing the bias and error of the resulting estimates. For air temperature, the systematic relationship between elevation and station observations has been used as a basis for 'smart' interpolation, enhancing estimation accuracies even further. The relationship between precipitation and topography is, alas, substantially more complicated. This talks demonstrates how a combination of temporally-static 0.1 degree satellite-enhanced precipitation mean fields and coarse (2.5 degree) time-varying precipitation fields can be combined to provide a reasonable basis for the 'smart' interpolation of rainfall in areas of complex terrain and limited in situ gauge data. Accuracy assessments are carried out for Ethiopia and Kenya.