FLOOD ESTIMATION IN A RIVER BASIN DOWNSTREAM OF A MAJOR RESERVOIR IN COASTAL REGION

Abstract

In the present study, flood estimation of Dhadhar River basin downstream of Dev dam up to the Gulf of Khambhat has been carried out using unit hydrograph based conventional hydro-meteorological approach and HEC-HMS model. The results of hydro-meteorological approach have been compared with the results of HEC-HMS model. Dhadhar basin has been divided into two sub-basins. The first sub-basin is from downstream of Dev dam to Pingalwada gauging site. The second sub-basin is from Pingalwada to Gulf of Khambhat. The second sub-basin is ungauged. Unit hydrograph has been developed using Nash model at Pingalwada gauge site. For estimation of flood for the second sub-basin, dimensionless unit hydrograph for the region was developed and used. At Gulf of Khambhat a fresh water lake having a surface area of 2000 km2 (known as Kalpasar), is being developed by constructing the dam across the Gulf of Khambhat. In general, the basin has got flat slopes. The sub-basin 2, which is ungauged and ultimately meets Gulf of Khambhat has further flatter slopes. The regional formula developed by Central Water Commission for computation of time to peak, gives very less value of time to peak as compared to the observed time to peak. The application of dimensionless unit hydrograph requires time to peak for the ungauged basins. Hence, an improved formula was developed for computation of time to peak (tp) based on physiographic parameters. This formula is applicable for lower portion ungauged sub-basins of Dhadhar, Mahi and Sabarmati rivers. The combined flood peak at basin outlet due to basin probable maximum flood (PMF) by using 2-bells method and dam design flood release works out to be 8436 m3/s with time to peak as 80 hrs. The peak flood volumes corresponding to 3 hr to 12 hr durations vary from 91.1 MCM to 359.48 MCM. For evaluating the performance of HEC-HMS model, 1991 and 1993 events were used for calibration and 2010 event was used for validation. The Nash Sutcliffe efficiencies for 1991, 1993 and 2010 events using HEC-HMS model are 93.2%, 97.1% and 70.6% respectively. For calibration and validation phase, the percent difference in peak flows to be in the range of 1.29 % to 13.11 % and the percent difference in volume to be -4.78 % to 7.86 %. The combined flood peak at basin outlet due to PMP over basin and dam design flood release was also computed using HEC-HMS model. For comparison purpose, the entire basin below Dev dam was divided in two sub-basins as was done in conventional approach. The peak discharge at v Pingalwada works out to be 5910 cumecs using HEC-HMS model against 6507 cumecs by conventional approach. The time to peak works out to be 64 hrs in both the methods. The peak flood volumes for 4hr, 8hr and 12 hr durations are approximately 9.18% lesser as compared to conventional approach. At the outlet of basin, the peak discharge works out to be 7282 cumecs against 8436 cumecs by conventional approach. The time to peak works out to be 96 hrs against 80 hrs and the peak volume of 4hr, 8hr and 12 hr durations are approximately 14.4% lesser as compared to those given by the conventional approach. The results establish the validation of HEC-HMS model for flood computation. The model is found well capable of capturing the hydrologic response of the catchment for flood events and can be useful for computing peak discharges, time to peak and peak flood volumes.