城市场次降雨情景库构建及预报降雨匹配技术  

Construction of urban rainfall scenario database and matching technology for predicting rainfall

作　　者：杨子昕 王佳[4,5] 刘家宏 王浩[4] 梅超 李峰平[1,2,3] YANG Zixin;WANG Jia;LIU Jiahong;WANG Hao;MEI Chao;LI Fengping(Key Laboratory of Groundwater Resources and Environment(Jilin University),Ministry of Education,Changchun 130021,Jilin,China;Jilin Provincial Key Laboratory of Water Resources and Water Environment,Jilin University,Changchun 130021,Jilin,China;College of New Energy and Environment,Jilin University,Changchun 130021,Jilin,China;State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin,China Institute of Water Resources and Hydropower Research,Beijing 100038,China;Key Laboratory of River Basin Digital Twinning of Ministry of Water Resources,Beijing 100038,China)

机构地区：[1]地下水资源与环境教育部重点实验室,吉林长春130021 [2]吉林省水资源与水环境重点实验室,吉林长春130021 [3]吉林大学新能源与环境学院,吉林长春130021 [4]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 [5]水利部数字孪生流域重点实验室,北京100038

出　　处：《水利水电技术（中英文）》2024年第10期13-24,共12页Water Resources and Hydropower Engineering

基　　金：国家重点研发计划项目(2022YFC3090600);国家自然科学基金项目(52192671);重点实验室项目(SKL2022TS11)。

摘　　要：【目的】城市降雨具有高度的不确定性和时空变异性,现阶段洪涝风险预测、应急响应预案的时效性和有效性有待加强。为提高“预报降雨—灾害风险识别—应急响应预案”的时效性和有效性【方法】提出了一种基于自组织特征映射网络的数据密集型城市场次降雨情景库构建方法和基于动态时间弯曲法的预报降雨与情景库场次降雨匹配技术。以北京城市副中心通县站1980—2015年(36年)的小时降雨监测数据作为情景库数据基础,并以同地区2023年降雨数据作为“预报降雨”进行匹配方法验证。【结果】结果显示:构建的城市场次降雨情景库可全面、有效的反映研究区域降雨特征。“预报降雨”匹配效果良好,平均纳什系数达到0.73,降雨总量平均相对误差为0.17,降雨量峰值平均相对误差为0.09。【结论】该方法可快速实现实时预报降雨与情景库中场次降雨的匹配,通过情景库中预置洪涝风险及应急预案知识库,实现城市洪涝风险的有效识别和及时响应。[Objective]Urban rainfall exhibits high uncertainty and spatiotemporal variability,and the timeliness and effectiveness of flood risk prediction and emergency response plans need to be enhanced.To improve the timeliness and effectiveness of“rainfall forecasting-disaster” risk identification-emergency response plan,[Methods]a data-intensive urban sub-hourly rainfall scenario library construction method based on Self-Organizing Maps and a forecast rainfall matching technique with dynamic time warping method for scenario rainfall library were proposed.Hourly rainfall monitoring data from Tongxian Station in Beijing Urban Sub-center from 1980 to 2015(36 years)were used as the basis for the scenario library data,and rainfall data for the same area in 2023 were used to validate the matching method for forecast rainfall.[Results]The result showed that the constructed urban sub-hourly rainfall scenario library could comprehensively and effectively reflect the rainfall characteristics in the study area.The matching effect of“forecast rainfall”was good,with an average Nash coefficient of 0.73,an average relative error of rainfall total amount of 0.17,and an average relative error of rainfall peak value of 0.09.[Conclusion]This method can quickly achieve realtime matching of forecast rainfall with scenario rainfall in the library.With pre-setting flood risk and emergency plan knowledge in the scenario library,effective identification of urban flood risk and timely response can be achieved.

关 键 词：雨型 动态时间弯曲 城市洪涝 预报降雨 风险评估 洪水 降水 气候变化 

分 类 号：TV125[水利工程—水文学及水资源]