Abstract
In many hydrological studies, the main limiting factor in model performance is the low meteorological data quality. In some cases, no meteorological records even exist. Installing weather stations becomes a necessity in these areas when water resource management becomes an issue. The objective of this study is to propose a new experimental and exploratory method for determining the optimal density of a weather station network when being used for long-term hydrological modeling. Data from the Canadian Regional Climate Model at 15-km resolution (CRCM15) were used to create a virtual network of stations with long and complete series of meteorological data over the Toulnustouc River basin in central Québec, Canada. The weather stations to be fed to HSAMI, Hydro-Québec's lumped rainfall–runoff hydrological model, were selected in order to minimize the number of stations while maintaining the best hydrological performance possible using a multi-objective optimization algorithm. It was shown that the number of stations making up the network on the Toulnustouc River basin should be at least two but not higher than four. If the stations are positioned optimally, there is little to no gain to be made with a denser network. The optimization algorithm clearly identified that combinations of two or three stations can result in better hydrological performance than if a high-density network was fed to the model. Thus, the major conclusion of this study is that if weather stations are positioned at optimal locations, a very few number of them are required to model runoff with as good as or better performance than when a high-density network is used.
Optimizing climate model selection for hydrological modeling: A case study in the Maumee River basin using the SWAT