论文部分内容阅读
为实现天山西部山区喀什河流域冰川融雪区域的水资源可持续开发利用,更好地支撑所在区域工农业生产发展,有必要开展融雪径流中长期水文预报研究。基于相关系数法、主成分分析法及两种方法相结合的综合方法优选预报因子,采用BP神经网络模型和组合小波BP神经网络模型预报径流。结果表明,采用综合方法筛选出的预报因子集合可以得到更好的预报结果;组合小波BP神经网络模型在3个不同方案中的预测效果均优于BP神经网络模型的预测结果,其预报精度更高。研究成果可为该区域融雪径流模拟研究及洪水预报提供参考。
In order to realize the sustainable development and utilization of water resources in glaciers and snowmelts in the Kashi River basin in the western Tianshan Mountains and to better support the development of agriculture and industry in the region, it is necessary to carry out long-term hydrological forecast studies of snowmelt runoff. Based on the correlation coefficient method, the principal component analysis method and the combination of the two methods, the forecasting factors were optimized. The BP neural network model and the combined wavelet BP neural network model were used to forecast the runoff. The results show that the forecasting result can be obtained by using the comprehensive set of forecasting factors. The forecast results of the combined wavelet BP neural network model are better than those of the BP neural network model in three different scenarios, and the forecasting accuracy is more high. The research results can provide reference for simulating snowmelt runoff and flood forecasting in this area.