人口活动与美国新冠疫情时空分布关联关系研究
Study on the Correlation between Population Activities and Spatiotemporal Distribution of COVID-19 in the United States

新冠疫情(以下简称“疫情”)严重影响了人类生命安全，地理学者开展了对疫情时空分布、影响因素等方面的研究，人口活动对疫情时空传播具有显著的影响，但相关的研究较少。本文以疫情高发地美国为例，探究人口活动(州内日常活动、州际人口流动)与疫情时空分布(时空扩散、时空聚集)的关联关系。针对时空扩散：根据暴露指数构建人口流动网络，对迁入指数和实时基本再生数进行相关分析、对疫情演化序列和网络节点序列进行序列相似度对比，对环境数据和疫情演化结果进行相关分析；针对时空聚集：对人口流动网络并进行社区划分，采用空间自相关判断病例分布是否产生聚集，利用时空扫描得到疫情的时空聚类簇，将网络社区划分结果和时空聚类簇进行对比。结果表明：从实时基本再生数来看，美国疫情是趋于缓和的；美国疫情时空扩散与人口流动和人口日常活动存在关联关系，餐饮购物活动和人口流入是主要影响因素；美国疫情时空聚集与人口流动存在关联关系，人口流动网络中社区内部的密切人口往来塑造了疫情时空聚集特征。人口活动与疫情时空分布存在关联关系，为了遏制疫情应当严格控制人口活动。
The COVID-19 epidemic has seriously affected the safety of human lives. Geographers have carried out studies on the spatiotemporal distribution and influencing factors of COVID-19, etc. Population activities have a significant impact on the spatiotemporal spread of COVID-19, but there are few relevant studies. This paper takes the United States as an example to explore the relationship between population activities (daily activities within the state, interstate population movement) and the spatiotemporal distribution of the epidemic (spatiotemporal diffusion, spatiotemporal aggregation). For the spatiotemporal diffusion: the population flow network was constructed according to the exposure index, the correlation analysis was conducted between the migration index and the real-time basic regeneration number, the sequence similarity comparison was made between the epidemic evolution sequence and the network node sequence, and the correlation analysis was conducted between the environmental data and the epidemic evolution result. Spatial and temporal clustering: the population flow network was divided into communities, spatial autocorrelation was used to judge whether the distribution of cases clustered, and spatiotemporal clustering cluster of the epidemic was obtained by spatiotemporal scanning, and the results of network community division were compared with the spatiotemporal clustering cluster. The results showed that the epidemic in the United States was moderating from the perspective of real-time basic regenerative number. The spatiotemporal spread of the epidemic in the United States was correlated with population mobility and daily activities, with catering and shopping activities and population inflow as the main influencing factors. The spatiotemporal clustering of epidemics in the United States is correlated with population migration, and the close population exchanges within communities in the network of population migration shape the characteristics of spatiotemporal clustering of epidemics. There is a correlation between population activities and the temporal and spatial distribution