Sensitivity analysis of KW-GIUH rainfall runoff model with respect to infiltration methods and roughness coefficients

Flood forecasting in a sound way leading to correct results has been a challenge for all researchers and engineers for many decades, which is the basic reason for developing many different types of mathematical rainfall runoff models. Correct estimation of infiltration during a storm is essential to a justified and rational modeling of runoff at watershed scale. There are many experimental or mathematical models for simulating infiltration and deriving net rainfall with pros and cons. In the present research, instead of evaluating infiltration simulation methods solely, four more widely used methods; namely, ɸ index, Horton, Green-Ampt and SCS methods were selected to find out their effects on the performance of a kinematic wave based geomorphological model called KW-GIUH. Furthermore, a sensitivity analysis with respect to different infiltration methods and different overland and channel roughness coefficient were performed. Horton and ɸ method led to a better performance of the model in terms of observed and simulated hydrographs in the study watershed. In this regard, Nash-Sutcliffe efficiency indices for Horton and ɸ index methods were obtained as 94.9 and 90.3, respectively, while it was 77.7 for Green–Ampt and 76.3 for SCS infiltration method. According to results of sensitivity analysis, KW-GIUH performance has the most and least sensitivity when using Green–Ampt and ɸ index as the infiltration method, respectively. Although studying in a steep watershed with an overland slope of about 17% and a small area of about 38 km2, changing the overland roughness coefficient has more effect on the model performance comparing with the change of channel roughness coefficient. Simulated flood peaks changed about 64 % due to changing the overland roughness coefficient while this value amounted to 25% for channel one. Shortly, it is concluded that KW-GIUH is highly sensitive to infiltration simulation method and overland roughness coefficient.