基于多方法优选预报因子的天山西部山区融雪径流中长期水文预报

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。

Medium and Long-term Snowmelt Runoff Forecasting in Western Tianshan Mountain Based on Multi-method Optimized Factors

In order to realize the sustainable development and utilization of water resource in Kashi river basin where located at the west Tianshan mountains of glacier snowmelt area and better support the development of industry and agriculture in snowmelt watershed, it is necessary to research medium and long-term hydrological forecasting of snowmelt runoff. The correlation coefficient method, PCA and synthetic approach are selected to get the optimized predictors, respectively. And then BP neural network model and combination wavelet BP neural network model are used to forecast the runoff in study basin. The results show that the integrated approach for screening of the predictor set may get better forecast results; based on the predicted results under different scenarios, the results of combination wavelet BP neural network model is better than that of BP neural network model, and the forecast accuracy is much better than others. The research results can provide a reference for the study of snowmelt runoff simulation and flood forecast in this area.