Forecasting local climate for policy analysis : a pilot application for Ethiopia

This paper describes an approach to forecasting future climate at the local level using historical weather station and satellite data and future projections of climate data from global climate models (GCMs) that is easily understandable by policymakers and planners. It describes an approach to synthesize the myriad climate projections, often with conflicting messages, into an easily-interpreted set of graphical displays that summarizes the basic implications of the ensemble of available climate models. The method described in the paper can be applied to publicly-available data for any country and for any number of climate models. It does not depend on geographic scale and can be applied at the subnational, national, or regional level. The paper illustrates the results for future climate for Ethiopia using future climate scenarios projects by 8 global climate models. The graphical displays of nine possible future climate regimes (average temperature, precipitation and their seasonal distribution) for each grid-cell about 50km X 50 km). It also provides the probability associated with each of the nine-climate regimes.