Hydrological modeling and uncertainty analysis for a snow-covered mountainous river basin.

Uncertainty analysis is crucial before hydrological simulation to address model vulnerability in snow-covered mountainous river basins, ensuring effective soil conservation and water resource management planning. The present study is conducted on Sutlej River Basin, which is located in Western Himalayas and originated from Manasarovar Lake at Tibetans platue in China. The Soil and Water Assessment Tool (SWAT) model is calibrated; the parameter sensitivity and uncertainty of the SWAT were quantified using four optimization methods such as (i) sequential uncertainty fitting (SUFI-2), (ii) particle swarm optimization (PSO), (iii) generalized likelihood uncertainty estimation (GLUE), and (iv) parameter solution (ParaSol). The model is performed with calibration period (1982–2000) and validation period (2001–2013), using monthly observed streamflow at the Bhakra gauging station. The statistical performance criteria such as Nash–Sutcliffe efficiency coefficient (NSE) and coefficient of determination (R2), values showing all four models, performed very well in the model calibration and validation. During the calibration period, SUFI-2 and ParaSol algorithms show higher performance with NSE and R2 values of 0.82 and 0.83. In the validation period, SUFI-2 shows high accuracy with NSE and R2 values of 0.77 and 0.78, while ParaSol also performed well with NSE and R2 values of 0.71 and 0.76. These results highlighted the superior performance of SUFI-2 and ParaSol compared to the other two (PSO, GLUE) techniques. Further, the values of p-factor and r-factor reveal that ParaSol method performed poor during both the calibration period (p-factor: 0.25, r-factor: 0.32) and the validation period (p-factor: 0.26, r-factor: 0.38). These results suggest that while ParaSol was effective in optimizing parameter sets, it fails in providing accurate estimates of uncertainty during both model calibration and validation, whereas other three methods perform very well in the model uncertainty analysis. This study fills a significant research gap by offering guidance on adjusting sensitive parameters, reducing uncertainty in streamflow simulation, and addressing the challenges specific to hydrological modeling in mountainous river basins, which is different from the previous studies conducted in plain river basins. [ABSTRACT FROM AUTHOR]