COMPARTIVE EVALUTION OF HEC-HMS AND SRM MODEL IN SIMULATING RAILFALL RUNOFF BEHAVIOR OF BUDHIGANDAKI CATCHMENT

Abstract

The estimation of stream flow from ungauged basin is very important for the planning of water resource projects in the basin. Most of the basins in the Himalayan region are ungauged due to highly rugged mountainous terrain where it is very difficult to observe hydro meteorological data and it is not possible to do the measurement at every small catchment and at every location of the stream however great potential for development of hydroelectric projects exists in these catchments. There is no way of estimating the stream flow at the desired location except some mathematical analysis or modeling of the stream flow which is very crucial for the planning of projects at a particular location to assess the availability of water. This study mainly focus on the comparative analysis of two models which are supposed to be applicable for the Himalayan region due to their ability to simulate both rainfall and snowmelt runoff. The two models HEC-HMS developed by US Army corps of Engineers and SRM model developed by Martinec et al., (1975) have been applied for the study of Budhigandaki catchment in Nepal. Various hydro-meteorological data required to these models have been collected from the Department of Hydrology and Metrology, Nepal. Arc GIS software has been used extensively during the study to create database related to topographical features required by the models and to create HEC-HMS project from the SRTM-DEM. The MODIS Aqua (MYD10A2) and Terra (MOD10A2) 8 day maximum snow cover extent product have been used to derive the snow cover area (SCA) in the basin at 8 day time step and this 8 day time step snow cover area have been interpolated into daily time step as required by the SRM model. Both the models were calibrated using the observed data for the year 2003 and validated using data for the year 2004 to 2008. The various parameters required by the models were initially estimated by using the historical data, previous similar studies and topographical features of the basin. Finally, the parameters were tuned by using manual and auto calibration technique. The models were evaluated by using three well known evaluation criteria NSE, R2 and PEV (Percentage error in volume). The values of NSE, R2 and PEV were found better in SRM model than in HEC-HMS in both calibration and validation periods. In calibration period, the values of NSE, R2 and PEV were found to be 0.944, 0.944 and -0.3% for SRM model and 0.778, 0.785 and -1.83% for HEC-HMS model respectively. Similarly, in validation period the average values of NSE, R2 and PEV were found to be 0.811, iv 0.816 and 2.01% for SRM model and 0.773, 0.777 and 3.5% for HEC-HMS model respectively. Sensitivity analysis of the models revealed that constant loss rate is very sensitive parameters for HEC-HMS model whereas runoff coefficients and degree day factors are most sensitive parameters for SRM model. The degree day factor which is a common parameter to both the models showed the higher sensitivity in SRM model than in HECHMS. In conclusion, the performance of the models revealed that both the models could be applied to the catchment to carryout hydrologic modeling, however, in comparison between two models SRM model has given betters simulation results than HEC-HMS model and also SRM model is simpler than HEC-HMS and requires less data. So SRM model is recommended for the study area.