Modelling streamflow using the SWAT model and multi-site calibration utilizing SUFI-2 of SWAT-CUP model for high altitude catchments, NW Himalaya's

Abstract

Modelling streamflow is important in developing effective long-term management, soil conservation planning, and utilization of water resource management strategies. The current study aims to establish a stable hydrological model simulating streamflow with minimal uncertainty amongst the calibration parameters. The SWAT 2012 model is utilized to simulate the Lidder catchment streamflow located in the Southeastern part of the Kashmir valley for the (2007–2015) periods. SWAT- CUP’s algorithm called Sequential Uncertainty Fitting 2 (SUFI-2) is employed in the study for multi-site model calibration and validation for monthly time steps. The calibration results for monthly simulation for the (2009–2012) period displayed an excellent model performance for flow rates with R2 of 0.89, 0.85, and 0.89, and NSE values obtained are 0.73, 0.72, and 0.83. However, the validation results for monthly simulation for the (2013–2015) period also displayed good model performance for flow rates with R2 values obtained as 0.86, 0.81, and 0.80, and NSE values of 0.76, 0.52, and 0.56. Following calibration, the combined impact of each used parameter is ranked by the SWAT-CUP global sensitivity function. A total of twenty parameters are optimized in the calibration. From the study, we find that the most sensitive parameters were Snowfall temperature (SMTMP), Temperature lapse rate (TLAPS), melt factor for snow (SFTMP), and initial SCS curve number value (CN2). Therefore, it is concluded that while developing a reliable hydrologic model, complete knowledge about the hydrological processes occurring within the river basins and awareness of an appropriate and meaningful set of parameters is important.