Equatorial total ozone as a predictor of sea surface temperatures useful for East African seasonal rainfall prediction

Abstract

Monthly total ozone data for the period 1979 to 2001, obtained from Total Ozone Mapping Spectrophotometer (TOMS), for two equatorial African locations of Nairobi, Kenya and Mahe, Seychelles were correlated with monthly sea surface temperatures of the Indian Ocean, Pacific and Atlantic oceans. Total ozone was considered as the independent variable. Significant correlation coefficients (r) were obtained over Indian Ocean area, which were perceived to be useful for seasonal prediction of rainfall over greater horn of Africa. Analysis of lagged r values of SSTs versus total ozone over Nairobi and Mahe, respectively, were performed for all the months of the year. Ocean areas with significant lagged correlation between monthly total ozone values and monthly SSTs of Indian Ocean, Pacific Ocean and Atlantic oceans were obtained for both of the ozone stations. For both Nairobi and Mahe, Total ozone for each month of the year showed significant lagged correlation with areas within Indian, pacific and Atlantic oceans. The lagged r values for all months, with ozone as the independent variable, were generally persistent for a lag time of about eight months. This alluded to the heat memory of the oceans. The influence of African equatorial total ozone variability on Indian Ocean, whose SSTs are operationally used for East African seasonal rainfall prediction was investigated. The SSTs utilized for prediction of March –May (long) rains, namely, January and February, SSTs were averaged and correlated with December ozone. Similarly, the SSTs utilized for prediction of October –December (short) rains, namely, July, August and September, SSTs were averaged and correlated with June ozone. Three areas, which are normally operationally considered in prediction process of the seasonal East African rainfall, were found to have significant correlation with ozone. SST prediction models for some of the ocean areas were developed. These models exhibited satisfactory skills. This is an indication that the time evolution of the SSTs over these ocean areas can be predicted with a lead-time of several months, using total ozone, and hence increase the lead-time of prediction of seasonal rainfall over East Africa.