Rainfall – runoff modelling using SWAT for Ong river basin
DOI:
https://doi.org/10.59797/q1hq5131Keywords:
SWAT, SWAT-CUP, Rainfall-runoff modeling, Sensitivity analysis, Uncertainty analysisAbstract
The Ong basin, covering major portions of western Odisha in India, has been repeatedly facing threats of hydro-meteorological calamities such as floods, droughts and cyclones in the recent times. In order to understand the different hydrological processes occurring in Ong basin, which covers about 5,128 km2area with a very diverse hydrological variability, hydrological analysis of the basin has been carried out utilizing Soil and Water Assessment Tool (SWAT). Digital Elevation Model (DEM), soil map, land use (LU) / land cover (LC) map, climatic data, streamflow data etc. have been used. A total of 52 hydrologic response units (HRUs) were created in the 11 subbasins by applying 5% threshold value of both land use and slope and 10% in soil classes. The calibration process was carried out using SWAT-Calibration and Uncertainty Procedures (SWAT-CUP) tool with SUFI-2 algorithm. Observed monthly stream-flow data at Salebhata (only existing G&D site) of the Ong basin for a period of 16 years (1983 to 1998) were given as input through auto calibration tool which gives Nash-Sutcliffe Efficiency (NSE) of 0.81, R2 value 0.856, PBIAS of 16.55%, RSR value of 0.43. The validation was also carried out by using the monthly stream-flow data of 13 years (1999 to 2011) which gives NSE of 0.85, R2 value 0.859, PBIAS of 10.07%, RSR value of 0.39 indicating a decent model performance. Sensitivity analysis was also performed using LH-OAT technique, and out of 10 calibrated parameters 3 parameters viz., effective hydraulic conductivity in main channel alluvium, initial SCS runoff curve number II, and available water capacity factor were found to be highly sensitive. A close observation of these sensitive parameters revealed that the flow characteristics of this area were affected by both surface water and groundwater flow properties. In uncertainty analysis, the P-factor was 0.36 and Rfactor was 0.27 for calibration and validation, so calibration and validation can be considered satisfactory for this study.