Flood Change Detection and Attribution Using Simulation Approach in Data-Scarce Watersheds: A Case of Wabi Shebele River Basin, Ethiopia

摘要

Flood events vary with sub-regions, sites and time and show complex characteristics. This study investigated temporal variabilities in flood discharges and relationships with principal driving factors in data scarce Wabi Shebele River Basin. The preliminary analysis using exploratory data analysis (EDA) on annual and seasonal maximum discharge reveals that there are cycles of extreme flows at five- and ten-year intervals respectively throughout the basin. The statistical verification using the Mann-Kendall test and Quantile perturbation method indicates a significant trend in flood magnitude and frequency entire the basin in the early 21st century. For longest period (1980-2010) annual maximum stream flow shows significant positive trend (p-value < 0.05) in middle catchments and negative trend (p-value < 0.05) in eastern catchments. The years: 1986-1995, 2006-2010 are the years in which positive significant anomalies occurred in all seasons, while the years: 1980-1985, 1996-2005 are the occurrence years of significant negative anomalies. Rainfall from climate drivers; DA, BE, VS and fraction of sand from environmental background drivers; fraction of forest and population density from external factors were identified as the powerful driving factors of flood variabilities in the Wabi Shebele River Basin.