WATER AVAILABILITY ASSESSMENT USING HEC HMS MODEL FOR MADI RIVER BASIN OF NEPAL

Abstract

Among the seven main sub-basins of the Gandak river basin, the Madi sub-basin is one of them. This basin is occupied by glacier and snow covered area in the northernmost part. The basin principally belongs to the Higher Himalayan region, which is one of the major perennial tributary of Gandak river. A study of hydrological modelling along with flood frequency analysis and water availability of the river has been conducted in this study. This kind of study would be very useful in use of water resource and hydropower production. The HEC-HMS model was prepared in HEC-GeoHMS10.2 extensions in the Arc-GIS10.2 platforms and exported to use as an input in HEC-HMS. The HEC-HMS version 4.6 model was used to continuous rainfall-runoff modelling of the basin to carry out the performance evaluation of two Loss methods i.e. Green and Ampt and Soil Moisture Accounting Loss Method and three transform methods i.e. Clark UH Model, Snyder UH Model and SCS UH Model. The other parameters selected for the modelling were simple canopy for canopy, simple surface for surface, constant monthly for base flow, Muskingum-Cunge for routing and temperature index for snow melt consideration. The model was calibrated for the periods from 1990 to 2005 and validated for the periods from 2006 to 2012. The performance criteria of NSE, R2 and PBIAS was used to evaluate the performance of the model where the Clark UH transform model with Green and Ampt loss method was more satisfactory and acceptable as compared to other models. Overall performance of each model is found very good in both calibration and validation. Each model underestimates the flow during the calibration period and overestimates during the validation period. The performance of the model was found better while considering temperature index. The values of NSE, R2 and PBIAS of Model with considering temperature index were 0.796, 0.798, 3.9, for calibration along with 0.804, 0.86, 0.99 for validation respectively. The flood frequency analysis of the outlet station was performed using the different statistical distribution and computed using the best fit distribution. The 100, 500 and 1000 year return period flood of the station is 792.02 m3/s, 890.73m3/s, 934.08 m3/s, respectively. The flood frequency analysis was carried out to compare the probable flood in climate change scenarios where the 40s time series of CNRM CM5 model forecasts the maximum probable flood of 2004 m3/s in RCP 8.5 scenario. The future climate data of temperature and precipitation for RCP 4.5 and RCP 8.5 scenarios from ACCESS1 and CNRM-CM5 model after bias correction was used for generating the future flows employing the pre-calibrated HEC-HMS model. Only one station located at Siklesh with influence about 58.63% in the study area was taken for future precipitation analysis. This shows the highest precipitation expected in July and August of 2040s decade by RCP8.5 iii scenarios. The average annual basin precipitation is projected to be increased in all scenarios as well as intensity of precipitation increased in most of the cases. The peak runoff does not shift from August but increased in all scenarios except under ACCESS 1 RCP 4.5 of 40s. Excluding the probability of exceedance up to 25% all percentiles in future scenarios are predicted to be decreased than that on baseline average daily flow. There is a higher variability on monthly flows, which shows a higher decrease in flows on pre-monsoon and post monsoon months and a relatively increase in flows in monsoon which shows dry months get drier and wet months being even wetter. With limited data availability or data-scarce mountainous river basins of Nepal, HEC HMS model can be used effectively for scientific planning, development and management of water resources and to find out the impact of climate change in such basins.