Development of A Universal Calibration Platform for Watershed Models Using Global Optimization

摘要

Calibration is an essential part of watershed models, and a universal calibration platform based on advanced genetic algorithms is needed. In this study, a universal platform was constructed for different watershed models by transferring the configuration files of models and incorporating the Non-dominated Sorted Genetic Algorithm-II (NSGA-II). It was tested in two real cases studies by using two commonly used models, including the Hydrological Simulation Program-FORTRAN (HSPF) and the Storm Water Management Model (SWMM). For HSPF, the results showed that the goodness-of-fit indicators, in terms of NSE and R-2, were 0.82, 0.83 and 0.66, 0.67 during the calibration and validation period, respectively. For SWMM, NSE ranged from 0.854 to 0.920 and R-2 ranged from 0.737 to 0.912. The results indicated that this universal platform provided good model calibrations for both two models and it could be extended to other watershed models and other catchments as an effective and robust method for model calibration.