Trends and Frequencies of Extreme Rainfall Events in the Urban Catchments of Dar es Salaam, Tanzania

Abstract

Understanding the characteristics of extreme rainfall events is necessary for proper planning and management of urban flooding impacts. In this paper, daily rainfall data from four key weather stations for the period 1967-2017 were used to investigate temporal variability in annual, seasonal, and extreme rainfall in the urban catchments of Dar es Salaam, Tanzania. The MannKendall trend analysis and Sen’s slope estimator were used to quantify the magnitudes and significance of long-term trends in rainfall. The frequencies of extreme rainfall events were modelled using the Generalized Pareto model. Results of trend analysis provided evidence of a decrease in total annual rainfall, with the highest decrement being 6.59mm per year. The statistical significance of the decrease in total annual rainfall was inconclusive. Observations of increase in both annual and seasonal maximum rainfall were also made; with the highest increments being 1.01mm and 0.79mm per event, for annual and seasonal maximum rainfall, respectively. The statistical significance of the increase in annual maximum rainfall was certain at 3 out of 4 stations. Frequencies of extreme rainfall events investigated using the R6 threshold provided reasonable results based on actual experience in the study area. Results indicated that most of the pluvial and fluvial flooding are from rainfall events with a 2 to 10-year return period. This is indicative of issues with the drainage systems in the area; either in their designed capacity or the reduction of their water carrying capacity due to anthropogenic factors.

Keywords: rainfall trends, frequency analysis, extreme rainfall, generalized pareto, catchment.