Physics and Chemistry of the Earth vol:50-52 pages:24-33
A study was carried out to investigate variability in long term hydro-climatic extremes in the Lake Victoria basin, East Africa. The study aimed at determining whether the long term historical changes in frequency and magnitude of hydro-climatic extremes are statistically significant, to give more light on the differentiation of climate variability from climate change. Long term extremes for 22 rainfall and 10 river flow gauge stations were examined. The hydro-climatic extremes were aggregated at levels
from daily, decadal, to monthly scales defined for two wet seasons in the area, the long rainy season extending from March to May (MAM) and the short rainy season extending from October to December OND), and time slices of 10 years using a sliding window approach. An empirical statistical technique based on Quantile Perturbation Method (QPM) was used. Quantile perturbations that represent empirical
changes for precipitation and river flow extremes were derived. Significant decreasing trends in precipitation
were observed in the 1930s, 1970s and 1980s, while significant increasing trends were common in the 1960s, late 1980s, and 1990s to the most recent years (2000–2006). In general, significant trends were dominant in the OND compared to MAM season for precipitation and river flow extremes. Results indicated further that there are differences in geographic location of significant trends in the hydro-climatic variables investigated implying that impacts are not spatially coherent. Areas with significant trends appeared to be concentrated in the North to North eastern parts compared to those in the southern parts
of the basin.