Regional Low Flow Analysis: The Case Of Upper Omo Gibe River Basin, Ethiopia

Abstract

Assessment of low flow of a river in magnitude as well as in frequency is crucial for planning and design of water resource projects such as planning of water supplies, water quality management issuing, hydropower project, irrigation project and the impact prolonged drought on aquatic ecosystems. The objectives of this study were to quantify the characteristics of low flows in rivers of upper Omo gibe, and to estimate the magnitude, frequency, flow regionalization, and to fit best fit statistical distribution. Twenty hydrometric stations in the upper Omo gibe which have more than 16 years of complete data were selected for the current low flow study. L-moment based approach and geographical proximity location of the station were applied for regional frequency analysis of annual minimum 7-day low flows and four separate homogeneous regions were identified. The most frequently used distributions in the analysis of hydrologic extreme variables are: Generalized Extreme Value (GEV), Lognormal (LN), Generalized Pareto distribution (GP) and generalized logistic (GLO). For selection of best-fit distributions L-MRD, XLSTAT Statistical computer software, EASY FIT Statistical computer software and Matlab were employed. XLSTAT Statistical computer software were used to select methods of parameters estimation for at-site low flow Frequency Analysis and Matlab software were selected for parameter estimation depend on RMSE. Using the goodness-of- fit tests (Chi test and Kolmogorov-Smirnov test), for most of the stations, the selected probability distributions were GEV and GP distributions. Method of Moments (MOM) was selected for all distribution used in the study for estimation of parameter. The result of (Zdist) ) indicated that Generalized Extreme Value and Generalized Pareto distributions are most appropriate probability distribution for Region-1, Region-2 and Region-3, Region-4 respectively. Finally, the growth curves developed using the estimated at site and regional quantiles for all stations and identified regions.