黄河干流缺水决策树模型研究

以黄河干流上的贵德、兰州、河口镇、龙门、三门峡、花园口等6个主要水文站为节点,通过径流量频率分析,分别研究了各节点缺水状态。把各节点频率为50%的来水量作为缺水的基准,将缺水状态分为5个等级,以天然径流系列数据为样本,以贵德站为例建立了节点缺水决策树模型。决策树对观测样本的分类、预测直观明了,通过观测节点月径流量,可实现节点缺水状态的滚动预报。

Study on Decision-Making Tree Model of Water Deficiency of the Main Yellow River

The paper respectively studies the water deficiency state of each nodal point by taking the six main hydrometric stations of Guide,Lanzhou,Hekouzhen,Longmen,Sanmenxia and Huayuankou situated on the main Yellow River as nodal points and through runoff frequency analysis.It takes 50% coming water of each nodal point frequency as the standard of deficiency,divides water deficiency state into 5 grades,takes natural flow series data as a sample and establishes a decision-making tree model of nodal point water deficiency by taking Guide Station as an example.The decision-making tree is visual and clear for the classification of an observed sample and can achieve rolling forecast of water deficiency state of nodal points through nodal point monthly runoff observation.