Indian Ocean sea surface temperature and Eritrean highlands rainfall Mehari Tesfazgi Mebrhatu * , Sue Walker Department of Soil, Crop and Climate Sciences (Agrometeorology), University of the Free State, P.O. Box 339, Bloemfontein 9300, South Africa Abstract Given an improved understanding of Eritrean climate, numerous benefits could be expected in many related activities: better management of agriculture and water resources stemming from more reliable seasonal predictions. In this study the Indian Ocean sea surface temperature was identified out of 11 predictors to be the most influential predictor for the July and August rainfall in the highlands of Eritrea. A statistical model was developed for peak rainy months (July–August, JA) of the study area. The model jack- knife skill test gave the correlation of 0.89 and 0.85 for Asmara and Mendefera stations, which is very high for rainfall prediction. Thus, validation of the model shows that the model can reproduce the measured monthly sum for JA rainfall totals with confidence. Ó 2004 Published by Elsevier Ltd. Keywords: Eritrean highlands; Indian Ocean SSTs; Jack-knife cross-validation; Statistical model 1. Introduction Eritrea lies between latitude 12°40 0 –18°02 0 N and longitudes 36°30 0 –43°20 0 E. Most parts of Eritrea receive rainfall from the south-western monsoon winds during the spring and summer months (April–October) (FAO, 1994). The rainfall is mainly convective. ‘‘Short rains’’ fall in April/May and the ‘‘main rains’’ in July and August. Seasonal forecasting has good prospects for early warning of low rainfall totals to help prepare for, and mitigate the effect of, famine, which so often results in Eritrea. The need for providing accurate forecasts for a coming rainfall season is becoming more and more necessary. Farmers could make better management deci- sions if they had a better assessment of the forthcoming season. Interannual variability of rainfall in East Africa re- sults from complex interactions of forced and free atmospheric variations (Ogallo, 1988; Mutai and Ward, 2000). There have been several recent studies examining connections between observed rainfall and a number of large-scale climate signals (Montecinos et al., 2000; Wang et al., 2000; Clark et al., 2003). Studies have also looked for prediction links based on correlation with raw station data (Nicholls, 1981) or area averages (Mak- arau and Jury, 1997). Promising seasonal forecast skill for the Oct–Nov–Dec ‘‘short’’ rains using multiple regression techniques have been found for East Africa and predictors based on eigenvectors of global sea sur- face temperatures (Mutai et al., 1998). Besides multiple regression, different techniques can also be used for sea- sonal forecasting models, for example, quadratic discri- minant analysis, (Mason, 1998) canonical correlation analysis, (Landman and Mason, 1999) or neural net- works (Hastenrath et al., 1995). An atmospheric general circulation model suggests that the Indian Ocean sea surface temperature (SST) ex- erts a greater influence over the East Africa short rains than the Pacific, (Goddard and Graham, 1999) specially the western Indian Ocean (Cadet and Diehl, 1984). A warmer tropical Indian Ocean is frequently associated with wet conditions over eastern Africa (Mason, 1995). This link between the Indian Ocean SST and the climate 1474-7065/$ - see front matter Ó 2004 Published by Elsevier Ltd. doi:10.1016/j.pce.2004.09.004 * Corresponding author. Tel.: +27 51 401 2222; fax: +27 51 401 2212. E-mail address: mebrhamt@sci.uovs.ac.za (M.T. Mebrhatu). www.elsevier.com/locate/pce Physics and Chemistry of the Earth 29 (2004) 1203–1207