ABSTRACT
Kenya, Ethiopia, and Sudan ............................................................. 353 14.5 Conclusions ................................................................................................... 355 References .............................................................................................................. 356
This chapter presents a novel interpolation approach that combines long-term mean satellite observations, station data, and topographic –elds to produce grids of climate normals and trends. The approach was developed by the Climate Hazard Group (CHG) at the University of California, Santa Barbara (UCSB), to support food security analyses for the U.S. Agency for International Development’s (USAID) Famine Early Warning Systems Network (FEWS NET). The resulting FEWS NET Climatology (FCLIM) combines moving window regressions (MWRs) with geostatistical interpolation (kriging). Satellite and topographic –elds often exhibit strong local correlations with in situ measurements of air temperature and rainfall. The FCLIM method uses these relationships to develop accurate and unbiased temperature and rainfall maps. The geostatistical estimation process provides standard error –elds that take into account the density and spatial distribution of the point observations. These error –elds are especially important when evaluating climate trends. Numerous climate change analyses present trend evaluations without assessing spatial uncertainty. In many of these studies, the number of recent observations can be very low, potentially invalidating the results. This study presents analyses for the Sahelian and eastern African rainfall and air temperatures. The results indicate signi–cant rainfall declines in Sudan, Ethiopia, and Kenya. Every country exhibits signi–cant increases in average air temperatures, with Sudan warming the most. This chapter concludes with a short discussion of how these results are being used to guide climate change adaptation, with a case study focused on Ethiopia.
