Multi-objective synchronous calibration and Pareto optimality of runoff and sediment parameters in an arid and semi-arid watershed.

Calibration methodologies must extract as much information as possible from available data, but it is not well understood in investigating the multi-objective synchronous calibration strategy by using multiple sources of information and by exploiting the data in better ways. The non-dominated sorting genetic algorithm II (NSGA-II) is introduced to study the calibration performance of runoff and sediment parameters under nine targeted scenarios, which considers the best choice to obtain high-cost performance results for decision makers through multi-objective optimization and the calculation of Pareto-optimal front with a high precision. (i) SWAT model has good adaptability in the runoff simulation of the Yanhe River watershed. Both the calibration results of NSGA-II and sequential uncertainty fitting approach-version 2 (SUFI-2) can meet the accuracy requirements of runoff simulation. Particularly, the NSGA-II based on multiple objective functions not only has strong applicability but also can better constrain the parameter process, making the calibrated model more in line with the physical conditions of the watershed. (ii) The two-site synchronous calibration of runoff or sediment can make full use of data information of different sites, reduce the impact of spatial heterogeneity on model parameters, and improve the calibration efficiency of the model. The single-site synchronous calibration of runoff and sediment based on NSGA-II not only has high calibration efficiency but also can avoid the tedious steps of calibrating runoff and sediment separately. (iii) The two-site synchronous calibration of runoff and sediment based on NSGA-II combines the advantages of the above synchronous calibration strategies, which can get Pareto-optimal front and represent the best trade-offs among different objectives, and its applicability is stronger than the traditional single-site or single-element calibration strategy. This study provides new and competing ways to evaluate hydrological models and their performance, and the multiple criteria approach for watershed modeling is one of the focuses in future research extensions.