Journal of Kenya Meteorological Society
The Atlantic-Indian Ocean Dipole and Its Influence on East African Seasonal Rainfall
William Nyakwada*, Laban A Ogallo**, Raphael E. Okoola**
*Kenya Meteorological Department, **Department of Meteorology, University of Nairobi, Kenya
Kenya Meteorological Department,
P.O. Box 30259, Nairobi, Kenya.
(Manuscript received 12 March 2008, in final form 13 May 2009)
This study has used principal component analysis, composite analysis and correlation analysis to establish the sea surface temperature modes that could represent the combined influence of the Atlantic and Indian Oceans on the seasonal rainfall over East Africa. The results from principal component analysis indicated that sixteen, sixteen, fifteen, and fourteen modes, accounting for about 93%, 94%, 93%, and 93% of the total seasonal sea surface temperature variance, were significant for the December-February, March-May, JuneAugust and September-December periods, respectively. Most of the first four modes represented sea surface temperature variability associated with the individual oceans such as basin wide warming/cooling associated with El Niño/ Southern Oscillation, inter-hemispheric SST variability over the Atlantic Ocean, and Indian Ocean Dipole. The decadal and inter-decadal variability were observed with the time coefficients associated with the modes. The results from correlation analysis indicated that the mode representing Atlantic-Indian Ocean Dipole together with the associated gradient has significant relationships with rainfall for March-May and September- December periods. The gradient mode accounted for the highest rainfall variance with September-December rainfall. The use of the gradient mode improved the values of correlation compared to those observed with the sea surface temperatures of the centres used to develop the mode indicating the ability of the gradient modes to improve relationships with rainfall. The results from composite analysis indicated that the gradient associated with the mode delineated the March-May and September-December rainfall associated with its opposite phases. The opposite phases of the mode were associated with opposite patterns of seasonal rainfall and wind currents confirming that the observed relationships are realistic. These results have documented a mode together with the associated gradient that can be used to represent the combined influence of the Indian and Atlantic oceans on the rainfall over the region, and improve the monitoring and prediction of seasonal rainfall over the region. However, more studies need to be done to understand further the dynamics of this mode and its association with rainfall over the region.