Comprehensive evaluation on carrying capacity of water resources based on the artificial neural network — A case study in Guizhou Province

This paper uses BP network to evaluate the water resource carrying capacity in karst area and compares the evaluating results with that calculated by gray relational projection method. Evaluating results of the two methods are similar overall. However, the water resource carrying capacity in Anshun and Tongren calculate d by the two ways is quite different. The result of BP network shows that the water resource carrying capacity in Tongren is the greatest, while Anshun is very small, only higher than Guiyang. But the result of gray relational projection shows that the water resource carrying capacity in Tongren is much lower than that in Anshun. To compare with former research results, it is proves that the method of BP network is more reasonable. By means of Kohonen network, the analyzed results show t hat the amount of water resources and economic level is the dominant factors impacting the water resource carrying capacity. In addition, the population has a greater influence to the water resource carrying capacity in Zunyi. The influence of karst conditions to the water resource carrying capacity in Tongren and Bijie is higher than other regions.