Parameter calibration of VIC model based on intelligent algorithm

Abstract

The parameters of the hydrological model are important parts of the model, and parameter calibration plays a key role in the output of the model. To analyze the influence of different calibration methods on the simulation results of the VIC hydrological model, the parameter calibration methods are coupled with the VIC model combining the characteristics of the Qinhuai River Basin. Three automatic optimization methods, including the genetic algorithm, simulated annealing algorithm and SCE-UA(shuffle complex evolution algorithm) algorithm, are used to automatically calibrate the 6 parameters of the VIC model. Three evaluation methods of the Nash efficiency coefficient (NSE), mean square error (MSE) and percentage deviation (PBIAS) are used to optimize the parameters. The model’s simulated runoff results are compared and analyzed. The results show that the SCE-UA algorithm with the Nash efficiency coefficient as the objective function can make the VIC model parameter calibration achieve ideal results. Different objective functions have different effects on the peak runoff during the model simulation verification period. By optimizing the parameters with the Nash efficiency coefficient and mean square error as objective functions, the simulated peak runoff is more similar to the measured data. The results can provide a reference for other models parameters automatic calibration, and are of great practical significance for rapid emergency response of hydrological simulation and forecasting.