Flood Frequency Analysis; Case of Awash-Awash Sub Basin, Awash River Basin, Ethiopia.

Abstract

This Flood Frequency Analysis aimed at analyzing flood magnitude and its probability of occurrence with fairly accurate not only aimed at preventing catastrophes but also at avoiding excessive costs in case of overestimating the flood magnitude or excessive damage while under estimating flood potential. The HEC-SSP software flood frequency analysis, MoM & L Moment parameters estimation and KS and X2 distribution fitting test statistics were used to achieve the study objectives. The stream flow data record length was varying from 18 to 44 years (1970-2014). The extreme theory for annual maxima has been applied and the best fitted distribution generated by KS model test statistics values range from 0.074 to 0.115 which is considered as heavy tail distribution and also these KS values are within the highly acceptable range since KS test statistics value threshold is less than or equal to 0.5 dimensionless value. The MoM statistics for CV range from 0.299 to 0.627 and the Cs value range from 0.519 to 2.007 while the L moment method generated ratio for L-Skew range from 0.011 to 0.366, the L-Kurtosis range from -0.014 to 0.300 which is the negative sign indicates the flat thin tail distribution and the L-CV value range from 0.184 to 0.319 which is the sample data are moderately variable. The shape parameters for all station data analyzed with GPA and Log pearson type III range from 0 to 2.15 which means the distributions have finite upper bound. Finally, the software model has generated the quantile flood magnitude with very small (0.001 to 0.584%) percent difference among the observed and computed values of stream flow records for the return period of 2, 5, 10, 20, 50, 100, 200 and 500 years. At the inlet of the sub basin the model generated quantile flood magnitude of 513.8m3/sec is identified as lower bound and 1208.2m3/sec valued as upper bound while at the outlet of the sub basin 366.3m3/sec and 659.8m3/sec computed as the lower and upper bound respectively. The slope of a flood frequency curve (FFC) graphically represents the standard deviation of the flood frequency distribution. The higher the slope, the greater the standard deviation in flood discharges