Meron Teferi Taye

Exceptional pattern of extreme rainfall variability at urban centre of Addis Ababa, Ethiopia

This paper examines the long-term historical changes in frequency and amplitude of rainfall extremes in two very close stations of Addis Ababa, the Addis Ababa Observatory (AAO) and Addis Ababa Bole (AAB) records including some surrounding measurements using the second half of 20th century data. The method used to examine trends and oscillatory patterns is the quantile perturbation method, especially designed for extreme conditions. The result shows that the Addis Ababa Observatory (AAO) shows an exceptionally increasing trend in its extreme events while the rest of the stations show an oscillating pattern. Exceptional behaviour at AAO is also noticed from unique global climatic indices. Both Pacific Decadal Oscillation and the Southern Oscillation Index are opposed to other stations. A possible reason for the increasing trend is linked with the growing urbanization that has been occurring in Addis Ababa city in the past decades. Nevertheless the variability of AAO rainfall extremes is within the uncertainty range that indicates natural variability. Although the change may not be severe at present, it is important to note the increasing trend to avoid serious consequences in the future.

Decadal Analysis of River Flow Extremes Using Quantile-Based Approaches

Next to the traditional analysis of trends in time series of hydro-climatological variables, analysis of decadal oscillations in these variables is of particular importance for the risk assessment of hydro-climatological disasters and risk-based decision-making. Conventional parametric and nonparametric tests, however, need implementing a set of background assumptions related to serial structure and statistical distribution of data. They neither focus on the extreme events and their probability of occurrence. In order to get rid of these limitations, we suggest a modified version of the Sen Method (SM), combined with the Quantile Perturbation Method (QPM) for examining temporal variation of extreme hydrological events. The developed method is tested for decadal analysis of monthly and annual river flows at 10 hydrometric stations in the Qazvin plain in Iran. The results show oscillatory patterns in extreme river flow quantiles, with a positive anomaly during the 1990s and a negative one during the 2000s. It is also shown that the concurrent use of the two methods allows to set a complete picture on the temporal changes in high and low extremes in historical river flow observations in different seasons.