Use of Wavelet Transform to Analyse Observed Precipitation at Bukoba, Tanzania to Recognize Impacts to Hydrological Response in Lake Victoria Basin

Abstract

This research exploited seventy years of daily precipitation data from Bukoba, Tanzania to understand precipitation variability and change important for water resource management. Morlet Wavelet Transform was applied to discriminate the distinct time-frequency rainfall variability in the 92 days long MAM ((March and May)) and OND (October to December) primary rainfall seasons for intraseasonal characteristics over the periods 1931-1960, 1961-1990 and 1971-2000. The time-frequency analysis yielded wavelets outlining the intraseasonal nature of sporadic wet and dry spells in each epoch. The characters of the spells designate changing distribution, intensity and frequency of occurrence across the three epochs. The profile of the erratic wet and dry spells speculates shift change and fading of high frequency, quasi biweekly and low frequency oscillations. The oscillations which act together across atmosphere, oceans and land surfaces through convective processes are likely to influence seasonal precipitation anomalies at intraseasonal scale. The variability of the observed daily precipitation is thus hypothesized to be linked to the fading oscillations in the later two epochs particularly during the main MAM season and thus the declining precipitation in the study domain. Global increased atmospheric carbon dioxide concentration could catalyze this process.