基于机器学习降雨动态时空特征识别山丘区小流域洪水预报方法研究  

Research on flood forecasting method in mountainous small watersheds based on machine learning for identifying rainfall dynamic spatiotemporal features

作　　者：刘媛媛[1,2] 刘业森[1,2] 刘洋[3] 刘正风 杨伟韬[5] 胡文才 LIU Yuanyuan;LIU Yesen;LIU Yang;LIU Zhengfeng;YANG Weitao;HU Wencai(State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin,China Institute of Water Resources and Hydropower Research,Beijing 100038,China;Research Center on Flood and Drought Disaster Reduction of the Ministry of Water Resources,Beijing 100038,China;MWR General Institute of Water Resources and Hydropower Planning and Design,Beijing 100120,China;Fujian Water Conservancy and Hydropower Survey and Design Institute,Fuzhou 350001,China;Guangxi Zhuang Autonomous Region Water Conservancy and Electric Power Survey and Design Institute Co.,Nanning 530023,China;The Yi-Shu-Si River Basin Administration,Xuzhou 221018,China)

机构地区：[1]中国水利水电科学研究院流域水循环模拟与调控国家重点实验室,北京100038 [2]水利部防洪抗旱减灾工程技术研究中心,北京100038 [3]水利部水利水电规划设计总院,北京100120 [4]福建省水利水电勘测设计研究院有限公司,福建福州350001 [5]广西壮族自治区水利电力勘测设计研究院有限责任公司,广西南宁530023 [6]淮河水利委员会沂沭泗水利管理局水文局(信息中心),江苏徐州221018

出　　处：《水利学报》2024年第9期1009-1019,共11页Journal of Hydraulic Engineering

基　　金：国家自然科学基金重大基金项目(52394235);沂沭河流域超标准洪水防控能力提升措施建议(减JZ0145B042024)

摘　　要：山丘区洪水产汇流速度快,破坏力强,预报难度大。如何进一步提高山丘区洪水预报的准确性和预见期,是当前亟待解决的主要问题。针对该问题,本文基于机器学习技术,创新性地提出了一种洪水预报的新方法。该方法通过识别与当前降雨动态时空特征最相似的历史降雨洪水过程,“借古喻今”进行洪水预报。结果表明,在人为影响小、流域面积在600 km^(2)左右的山丘区小流域,该方法预报洪峰流量平均误差为8.33%,洪量平均误差为14.27%,峰现时间平均误差1 h,均达到了洪水预报精度要求。区别于传统的洪水预报方法,该方法从整场降雨发展趋势的角度上预报山洪,更有针对性,为山丘区小流域洪水预报提供了新思路,为“三道防线”数据深度挖掘,防洪“四预”智能化水平提升提供有力技术支撑。The mountainous region experiences fast-flowing and highly destructive floods,posing challenges for accurate and timely forecasting.Enhancing the accuracy and lead time of flood prediction in mountainous areas is a pressing issue.Addressing this concern,this paper proposes an innovative flood forecasting method based on machine learning technology.The approach identifies historical rainfall-flood events with the most similarity to the current dynamic spatiotemporal features of rainfall,employing a“learn from the past to predict the present”strategy.The results indicate that,in small watersheds with minimal human influence and a basin area of approximately 600 km^(2)in mountainous regions,the method not only predicts the overall trend of rainfall but also forecasts the associated mountainous flood processes under this rainfall trend.The average errors for peak flow,flood volume,and peak time are 8.33%,14.27%,and 1 hour,respectively,meeting the accuracy requirements for flood forecasting.Distinguished from traditional flood forecasting methods,this approach predicts mountainous floods from the perspective of the overall rainfall trend,providing a targeted strategy for flood forecasting in small watersheds in hilly areas.

关 键 词：人工智能 流形学习 降雨时空特征 山丘区小流域洪水预报 

分 类 号：P338[天文地球—水文科学]