基于信息熵的雨量站网优化研究  

Research on Rainfall Network Optimization Based on Information Entropy

作　　者：刘秀林 李雨 陈金凤 LIU Xiulin;LI Yu;CHEN Jinfeng(Bureau of Hydrology of Changjiang Water Resources Commission,Wuhan 430010,China)

机构地区：[1]长江水利委员会水文局,湖北武汉430010

出　　处：《水文》2021年第2期26-31,共6页Journal of China Hydrology

基　　金：长江水利委员会长江科学院开放研究基金项目(CKWV2016368/KY);国家重点研发计划课题“江河湖库水文要素在线监测装备技术集成与示范应用”(2017YFC0405706)。

摘　　要：基于信息熵理论,研究金沙江下游流域站点信息熵对降雨信息量和时空变异性特征的反映程度,发现信息熵能够在降雨量级和时空分布上充分表达降雨信息,确定了熵理论运用于本流域雨量站网规划的合理性;采用与组内互信息量最小为原则对三个实验流域站点进行重要性排序,以组内站点信息量和与组外互信息量总和最大为目标函数进行末位淘汰优选站网,结果显示站点优先级顺序一定程度上表达了其本身的信息量和在流域内的空间信息,优化后的站网比原站网数量减少15%~25%。Based on information entropy theory,the reflection of information entropy on rainfall information and spatial-temporal variability characteristics in the lower reaches of Jinsha River was studied.It is found that information entropy could adequately express rainfall information on rainfall magnitude and spatial-temporal distribution,which determines the rationality of applying entropy theory to the planning of rainfall network in the lower reaches of Jinsha River.Using the principle of minimizing intra-group mutual information,the importance of stations in three experimental basins was ranked.The final selection of the station network was based on the maximum sum of intra-group and extra-group mutual information.The results show that the priority order of the stations expresses their own information quantity and spatial information in the basin to some extent.Compared with the original station network,the optimized station network reduces by 15%-25%.

关 键 词：信息熵 降雨信息代表性 互信息 站网优化 

分 类 号：TV125[水利工程—水文学及水资源]