a Regional Study of Seasonal Rainfall Conditions in the Sudan

Available from UMI in association with The British Library. This thesis is primarily a regional climatological study of some spatial and temporal rainfall characteristics, with a special emphasis being made on quantitative analysis of the wet season rains over the Sudan. The study area has been objectively classified into five main "rainfall regions" on the basis of wet season duration and timing at 60 observing stations. Also a regionalisation derived from two multivariate techniques is compared and critically discussed. The spatial distributions of mean wet and dry season rainfall over individual regions, are described and explained by reference to the main weather systems and topographic features using the method of trend surface analysis. Rainfall variability has been quantified by the coefficient of variation index and analysed in a similar way to that of rainfall distribution. The results have shown an inverse relationship between variability indices and mean rainfall amounts. In relation to the study of variability, an assessment of rainfall probability and reliability has been discussed with reference to certain critical rainfall amounts and by applying the concept of binomial frequency distribution for critical time periods. The spatial distribution of mean rain days tends to portray a similar pattern to that of the mean rainfall. This apparent relationship is confirmed by the highly significant correlation coefficients between the two parameters. The relation between mean rainfall and daily rainfall intensity is not always a direct one, since wetter areas are mostly associated with increased frequency of rain days rather than greater rainfall intensities. Daily rainfall analysis has also revealed that the probability distributions of wet and dry spells of various sequence lengths can be mathematically derived from a Markov chain model and these were found to fit the actual daily data closely. Finally, the contributions of various atmospheric factors to the rainfall amounts are determined by applying multiple regression analysis. The provision of partial regression equations also permits an estimate of rainfall with a reasonable degree of accuracy for any place in the Sudan.