Information Entropy Theory-Based Optimizing of Gauge Networks for Hydrological Modelling—A Case Study in the Loess Plateau, China

Runoff prediction based on error analysis is paramount in flood monitoring and early warning, as well as in flood management. Precipitations are one of the most important input parameters in developing hydrological models to simulate runoff, and they can be represented accurately with a low-cost and well-designed rain gauge network. This study aims to use information entropy to optimize the rain gauge network of the Chabagou Watershed (CW) located in the Loess Plateau, China, as this area is largely affected by extreme precipitation and flood events. There are 13 rain gauges in the CW, and the optimal network is examined by using the Soil and Water Assessment Tool (SWAT) model to simulate the runoff. The results are (i) the gauge network composed of 10 stations performs similarly to the network formed by 13 stations; (ii) the performance of the SWAT model with different rain gauge distribution is stable, but the performance increases with the increase of rain gauge number. In summary, the information entropy-based optimization strategy helps to improve the performance of the gauge network, eventually helping in developing reliable regional hydrological models to be used in flood management.