过程——一种地理时空动态分析的新视角  

Process:A New View of Geographical Spatiotemporal Dynamic Analysis

作　　者：薛存金 苏奋振[3] 何亚文[4] XUE Cunjin;SU Fenzhen;HE Yawen(Key Laboratory of Digital Earth Science,Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China;International Research Center of Big Data for Sustainable Development Goals,Beijing 100094,China;State Key Laboratory of Resource and Environmental Information System,Institute of Geographical Science and Natural Resources Research,Chinese Academy of Sciences,Beijing 100101,China;College of Oceanography and Space Informatics,China University of Petroleum,Qingdao 266580,China)

机构地区：[1]中国科学院空天信息创新研究院数字地球重点实验室,北京100094 [2]可持续发展大数据国际研究中心,北京100094 [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]中国石油大学(华东)海洋与空间信息学院,山东青岛266580

出　　处：《地球科学进展》2022年第1期65-79,共15页Advances in Earth Science

基　　金：中国科学院A类先导专项子课题“全球海洋异常变化过程遥感监测系统”(编号:XDA19060103);国家自然科学基金项目“面向过程的海洋异常变化时空聚类挖掘模型”(编号:41671401)资助。

摘　　要：地理世界中存在一类具有产生、发展和消亡的地理现象/对象,综合对地观测技术和多源信息获取技术的发展提升了获取这种动态地理现象的能力。现行的地理时空分析方法以点、线、面、体为基本单元,以数据获取尺度为分析尺度,割裂了地理现象的时间连续性,限制了地理时空动态的分析能力。把产生、发展和消亡的动态演变抽象为地理过程,从演变过程的尺度,提出一种新的地理时空分析方法。首先,提出"地理过程—演变序列—时刻状态"的分解抽象和逐级包含的地理时空过程语义,并基于"节点—边"的图思想建立地理时空过程图表达方法和存储模型,实现地理对象(节点)和对象演变行为(边)一体化表达和存储;其次,以地理过程为基本单元,设计"地理状态对象提取—演变序列追踪—过程对象重构"的过程对象提取方法,并基于节点的出度(该节点引起其他节点变化的边的个数)和入度(其他节点引起本节点变化的边的个数)实现过程对象演变行为的识别;再次,以地理过程为分析尺度,拓展时空邻域为过程邻域、时空相似性为过程相似性,设计面向过程的地理时空挖掘方法,开展地理对象及其演变行为的时空模式挖掘;最后,以1950—2019年月尺度的太平洋海洋表面温度异常变化过程对象为例,挖掘了海洋表面温度异常变化结构及其行为,并分析了与ENSO类型之间的关联模式,验证了所提方法的可行性和适用性。There exists a sort of dynamic geographic phenomena with a property from production through development to dissipation in a real world,and the integrated Earth observation technology and the crowd sources technology promote the capabilities of obtaining these dynamics. The traditional spatiotemporal methods take a point,a line,a polygon or a voxel as an analyzing unit,and a scale of data acquisition as the analyzing scale,which splits a continuity of temporal evolutions,and limits their dynamic analysis. This paper abstracts the property from production through development to dissipation into a geographical process,takes it as a scale of analysis,and proposes a novel approach of geographical dynamic analysis. Firstly,a process semantics with a hierarchical abstraction of "geographical process-evolution sequence-instantaneous state" is proposed. And a graph-based model with a "node-edge" is used to represent and store geographical objects and their evolution behaviors. Secondly,a geographical process is used as a unit to design a process-oriented object extracting method with an "extracting of instantaneous state-tracking of evolution sequence-reconstructing of geographical process",and on the basis of in-degree and out-degree of a graph,four evolution behaviors are identified. Then,series of new concepts about spatiotemporal mining are redefined on the basis of a scale of process,and some process-oriented mining methods are designed,e.g. spatiotemporal clustering,association rule mining. Finally,a real dataset of monthly Sea Surface Temperature Anomaly(SSTA) during the period of January. 1950 to December. 2019 is to explore their evolving structures in Pacific Ocean,and the association patterns between the evolving structures of SSTA and the types of ENSO are addressed. Results demonstrate the effectiveness and the advantages of process-oriented dynamic analyzing method.

关 键 词：时空分析 地理动态 演变行为 过程挖掘 海洋环境异常变化 

分 类 号：P90[天文地球—自然地理学]