Considering flood scaling property in multi-objective calibration of the SWAT model: a case study in Zijinguan watershed, Northern China

Adopting new technologies to calibrate hydrological models to produce good simulations for water resources or flood risk management is one of the important research topics in hydrology field. Incorporating hydrological signatures and concepts in hydrological model calibration becomes prevalent in recent years. A good realistic simulation of hydrological processes within the catchment is more significant than only achieving good performance at watershed outlet with single-statistical-objective function. This research adopts the concept of flood scaling property, which describes the statistical relationship between flood peak, its contributing areas and additional catchment attributes, to behave as one of the constraints under multi-objective model calibration framework. Several designed calibration scenarios are tested by employing the soil and water assessment tool hydrological model. In comparison with single-objective calibration scenarios, multi-objective calibration method is recommended to obtain both good simulation at catchment outlet and sub-basins reaches. The multi-objective calibration method assists in improving the model performance in terms of flow duration curve and also in reducing the bias of long-term runoff ratio. The proposed model calibration approach reflects the hydrological interaction of flood peaks across sub-basins and takes into account the catchment antecedent wetness and climatic condition for each flood event.