RAINFALL-RUNOFF MODELING FOR EXTREME FLASH FLOODS IN WADI SAMAIL, OMAN

Abstract

Flash flood forecasting is indispensable to construct efficient warning and mitigation of the increasing threat of flash floods in wadi systems. Flash floods modeling in arid region is hindered by the lack of appropriate hydrological models and data availability. Rainfall-Runoff-Inundation (RRI) model is applied in Wadi Samail as a case study for the most extreme flash flood events in Oman (1,000 year return period). Detailed sensitivity analysis were conducted for RRI model parameters indicating that the channel and hillslope roughness coefficients are the most significant parameters followed by the soil characteristics parameters such as soil depth, porosity and hydraulic conductivity. The most severe tropical cyclone in the recent history in Oman, Gonu-2007 and Phet-2010 were used to calibrate and validate the hydrological model. The results and the statistical analysis indicating that RRI model could efficiently simulate the extreme flash flood events in the arid wadi system. Where for model calibration, Nash-Sutcliffe efficiency (NSE) and Percent bias (PBIAS) indices equal to 0.93 and -14.3, respectively, and for model validation NSE and PBIAS equal to 0.86 and -12.0, respectively. Further improvements in RRI model is recommended to include the transmission losses and groundwater processes for better representation of the wadi system.