The Comparison of Historical Time Series of Gridded Monthly Precipitation Data from Different Sources CRU, GPCC, and HSM_SIEREM Over Africa for the Period (1940- 1999)

Abstract

Nowadays, many precipitation gridded datasets are used for climatic research in Africa, a way for scientists to complement their observed data for several studies as river-runoff relationship, impact of climate change or hydrological modelling. These gridded datasets supposedly based on the same gauging stations delivered by any National Meteorological Institution seem to differ greatly according to different causes: the number of stations used, the quality check of data, the length of the time series, the interpolation method and lack of data for several periods and countries. In our study, we compared a number of available precipitation gridded dataset which are GPCC, CRU and HSM-SIEREM at a resolution of 0.5 x 0.5 degrees, during period between 1940 and 1999, and over Africa, then, we checked their correlation with ground based data chosen randomly to conclude the best gridded data product which appears to be closest to the reality. The results shows HSM-SIEREM presents best correlation coefficient with 0.9761, followed by GPCC with an insignificant difference by a correlation coefficient 0.9734 and a little behind CRU with a coefficient of correlation 0.9389.