新时代自然灾害态势感知的实践与方法探索  

The practice and method of natural disasters situational awareness in the new era

作　　者：程昌秀[1,2] 裴韬[3] 刘瑜[4] 杜云艳 沈石 江净超 CHENG Changxiu;PEI Tao;LIU Yu;DU Yunyan;SHEN Shi;JIANG Jinchao(Key Laboratory of Environmental Change and Natural Disaster,Beijing Normal University,Beijing 100875,China;State Key Laboratory of Earth Surface Processes and Resource Ecology,Beijing Normal University,Beijing 100875,China;State Key Laboratory of Resources and Environment Information System,Institute of Geographic Sciences and Natural Resources Research,CAS,Beijing 100101,China;Institute of Remote Sensing and Geographic Information System,Peking University,Beijing 100871,China;School of Automation,Hangzhou Dianzi University,Hangzhou 310018,China)

机构地区：[1]北京师范大学环境演变与自然灾害教育部重点实验室,北京100875 [2]北京师范大学地表过程与生态环境国家重点实验室,北京100875 [3]中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室,北京100101 [4]北京大学遥感与地理信息系统研究所,北京100871 [5]杭州电子科技大学自动化学院,杭州310018

出　　处：《地理学报》2023年第3期548-557,共10页Acta Geographica Sinica

基　　金：国家重点研发计划(2019YFA0606901)。

摘　　要：随着人类世的到来,“常态化”的极端天气与自然灾害成为人类生存面临的首要问题。大数据智能化的技术驱动、大科学(计划)的学科基础、全球化治理与应对的需求牵引,共同构成了新时代自然灾害研究的主旋律。论文提出自然灾害态势感知的概念,并结合新时代背景提出洞悉“态”、预测“势”两个不同层次的感知。在洞悉“态”方面,论文辨析了传统灾害观测与大数据态势感知的区别和联系;并以社交媒体、手机信令、视频监控等为例,梳理了大数据在台风、洪涝、地震、极端高温等灾害态势理解中的研究与实践。在预测“势”方面,论文总结了系列大数据观测—机器学习—机理模型整合的方法,并以城市洪涝事件演进模拟为例进行说明;在区域或全球尺度,提出应在大科学(计划)统一定义的未来情境下开展跨领域综合态势的感知与预测,并服务于区域的可持续发展;特别是利用人地耦合模型,感知灾害对社会经济的级联影响和远程效应,用以消除和解决非传统安全的威胁。最后,建议国家成立相关研究机构推进灾害大数据的共享和应用,开展灾害灾情标准知识库、训练库的建设;建议进一步推进机理模型、大数据、机器学习的整合与应用,推进自然灾害人地系统的耦合,提升中国自然灾害的社会治理与决策支持能力。Natural disasters and normalized extreme weather have emerged as the main threats to human life with the onset of the Anthropocene. The technological drive of big data and intelligence, the disciplinary basis of big science(planning) and the demand pull of globalized governance and response together constitute the main theme of natural disaster research in the new era. The paper puts forward the concept of natural disaster situational awareness and its two levels: situation understanding and situation predicting. In situation understanding, the paper identifies the differences and connections between traditional disaster observation and big data situational awareness;and systematically compares the research and practice of big data in disaster situational understanding of typhoons, floods, earthquakes, and extreme heat,using social media, cell phone signaling, and video surveillance as examples. In situation predicting, the paper summarizes a series of big data observation, machine learning and mechanical model integration methods and validates them with the evolution simulation of urban flooding events as an example. At the regional or global scale, it proposes that crossdomain integrated situational awareness and prediction should be carried out for some future scenarios defined by the big science(program) and serve the sustainable development of the region;especially, to use the human-earth coupling model to sense the cascading impact and remote effect of disasters on socio-economy, which can be used to eliminate and address nontraditional security threats. Finally, it is recommended that the state should establish relevant research institutions to promote the sharing and application of disaster big data, and carry out the construction of disaster standard knowledge base and training library. It is recommended to further promote the integration and application of mechanism models, big data and machine learning, and promote the coupling of natural disaster human-terrestrial systems to enhance the social governa

关 键 词：自然灾害 态势感知 大数据 机器学习 人地系统耦合 

分 类 号：X43[环境科学与工程—灾害防治]