Abstract
Calibration is an important step for the applicability of hydrological models as different parameter sets could produce similar results, calling for the use of appropriate performance criteria differentiating different parameter sets. This study focuses on the evaluation and comparison of SWAT-based hydrological modeling using both classical Nash Sutcliffe (NS) and the so called model selection criteria (MSC). Twelve SWAT models of the Sirwan River Basin in Iran are built based on different combinations of observed data, number of parameters and strategies used for calibration. The models are then evaluated against NS and eight MSC including AIC, AICc, AICu, CAIC, SIC, SICc, HIC and HICc, to rank the models and discriminate the most promising calibration setting from the aforementioned twelve candidate settings. Results show outperformance of MSC in terms of the robustness of the ranking outcomes both in calibration and validation phases as well as their discriminating power.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have