Geostatistical analysis of rainfall in the West African Sahel

Abstract

In this paper we attempt to resolve the controversy around interpretations of the changing rainfall patterns in the West African Sahel during the twentieth century. Some authors maintain that the data show clear evidence of drought starting in the 1970s and continuing through the 1980s (Dai et al., 2004), whilst other authors claim that the drought is an artefact of the decreasing number of rain gauges over the period in which the data were collected (Chapel and Agnew, 2004). It is somewhat surprising that none of the published studies on either side of this debate have used spatial methods, such as geostatistics, to model and interpret rainfall trends and to address data artefacts such as the decrease in the number of rain gauges. We report the results of a geostatistical study of rainfall patterns in the Sahel. We distinguish two sub-areas – the dry Sahel and the wet Sahel – and use block kriging with an anisotropic power semi-variogram to accommodate the latitudinal trend in rainfall. Our findings support the view that the period 1970-1989 was the most severe drought in the Sahel during the twentieth century and is not an artefact of the changing number of rain gauges and their locations.