地球信息科学学报 ›› 2021, Vol. 23 ›› Issue (2): 284-296.doi: 10.12082/dqxxkx.2021.200273

• 疫情建模与仿真 • 上一篇    下一篇

基于多元数据的中国地理空间疫情风险评估探索——以2020年1月1日至4月11日COVID-19疫情数据为例

方云皓1(), 顾康康1,2,*()   

  1. 1.安徽建筑大学建筑与规划学院,合肥 230022
    2.安徽省城镇化建设协同创新中心,合肥 230022
  • 收稿日期:2020-05-31 修回日期:2020-08-22 出版日期:2021-02-25 发布日期:2021-04-25
  • 通讯作者: 顾康康 E-mail:1198321182@qq.com;kangkanggu@163.com
  • 作者简介:方云皓(1995— ),男,安徽合肥人,硕士生,主要从事地理信息系统方面研究。E-mail: 1198321182@qq.com
  • 基金资助:
    安徽省自然科学基金面上项目(2008085ME160)

Exploration on Geospatial Risk Assessment in China based on Multiple Data: A Case Study of COVID-19 Data from January 1 to April 11, 2020

FANG Yunhao1(), GU Kangkang1,2,*()   

  1. 1. School of Architecture and Planning, Anhui Jianzhu University, Hefei 230022, China
    2. Research Center of Urbanization Development in Anhui Province, Anhui Jianzhu University, Hefei 230022, China
  • Received:2020-05-31 Revised:2020-08-22 Online:2021-02-25 Published:2021-04-25
  • Contact: GU Kangkang E-mail:1198321182@qq.com;kangkanggu@163.com
  • Supported by:
    Natural Science Foundation of Anhui Province(2008085ME160)

摘要:

新冠肺炎(COVID-19)在空间上具有一定的传播风险,对城市的安全健康构成了威胁,防止疫情传播成为紧迫的任务。在2020年1月1日至4月11日,COVID-19疫情经历了发生、迅速发展和趋于稳定的发展过程,利用初期的COVID-19数据进行宏观层面的疫情风险评估,为防疫控制措施提供一定的参考。因此本研究基于行政区划、定点医院、疫情小区以及道路交通等多元数据,在宏观层面提出并构建全国地理空间疫情风险性评估,对疫情风险分布探讨的同时进行评估效果验证,并根据构建指标揭示影响风险的因素及其机理,主要结论: ① 地理空间风险评估具有有效的可行性。② 地理空间疫情风险分布全局Moran's I指数为0.758,具有显著的空间集聚特征。同时,不同的省区市之间的局部LISA值呈现空间差异性,其中高—高聚类省区市占比全国25.81%,风险程度较高,主要分布在湖北、河南、湖南、江西、安徽、浙江、江苏、上海,低—低聚类省区市占比全国9.68%,风险程度较低,主要分布在青海、西藏、新疆。③ 地理空间疫情风险分布与地理区位、道路交通、医疗卫生、疫情现状指标均存在一定的相关性。根据统计学的Pearson相关性分析,其相关指标R 2存在差异,在数值上由高到低依次为疫情现状、地理区位、道路交通、医疗卫生,在属性上其相关因子存在正负2种效应,地理空间疫情风险与武汉市地理距离、定点医院密度以及居民-医院地理距离呈现显著的负相关,其R 2分别为0.813、0.545、0.436,与铁路网密度、公路网密度以及疫情小区密度呈现显著的正相关,其R 2分别为0.751、0.792、0.825。④ 地理空间疫情风险的构成因素错综复杂,其受到多种因子的共同作用,根据空间分层异质性分析,不同因子之间均存在交互作用,其中居民—医院地理距离与公路网密度、铁路网密度交互作用较强,q值分别为0.9842、0.9837。本研究在宏观层面为城市管理中重大疫情的空间资源分配以及区域空间的联防联控策略提供了相应的依据。

关键词: COVID-19, 多元数据, 地理空间疫情风险, 空间自相关, Pearson相关性, 空间分层异质性, 影响因素, 疫情防控

Abstract:

COVID-19 has a spatial transmission risk and poses a threat to the safety and health of the city. Preventing the spread of COVID-19 is therefore the urgent need for society now. The COVID-19 experienced a process that occurred, developed rapidly and stabilized from January 1 to April 11, 2020. Using the initial COVID-19 data for macro-level epidemic risk assessment can provide a certain reference for epidemic prevention and control measures. In this study, by using multiple data, including administrative division data, designated hospital data, epidemic community data, and road traffic data, we proposed a macro-level geospatial risk assessment and validated its effectiveness in China. Based on this, this study also analyzed the construction factors of geospatial risk assessment in order to explore its distribution rules. The following four conclusions showed that: ① Geospatial risk assessment of Covid-19 in China is effective to some extent. ② The spatial distribution of global Moran's I index of geospatial risk in China was 0.758, which had significant spatial agglomeration characteristics. At the same time, the LISA index in different provinces showed spatial differences. Some regions, including Hubei, Henan, Hunan, Jiangxi, Anhui, Zhejiang, Jiangsu, and Shanghai, were identified as high-high clusters and accounted for 25.81% of the provinces in China. The geospatial risk of these provinces was higher. Regions like Qinghai, Tibet, and Xinjiang, with a low degree of geospatial risk, accounted for 9.68% of the provinces in the country. ③ Some indicators, including geographic location indicators, road traffic indicators, medical and health indicators, and epidemic status indicators, were related to the distribution of geospatial risk. According to the statistical Pearson correlation analysis, there were differences in the correlation index R 2. In terms of numerical values, the epidemic status indicators, geographic location indicators, road traffic indicators, and medical and health indicators were ranked from highest to lowest. Different secondary factors were composed of four indicators, and they exhibited two effects of positive and negative correlation. Specifically, the factor of geographic distance from Wuhan, the designated hospital density factor, and the resident-hospital geographical distance factor showed significant negative correlations, with R 2 of 0.813, 0.545, and 0.436, respectively. However, the remaining factors showed significant positive correlations, including railway network density factor, road network density factor, and epidemic community density factor, and their R 2 were 0.751, 0.792, and 0.825, respectively. ④ The components of geospatial risk were intricate and complicated by multiple factors. According to the spatial stratified heterogeneity analysis, we found that there were interactions between different factors. Among them, the resident-hospital geographic distance factor interacted strongly with the railway network density factor and the road network density factor, and their q values were 0.9842 and 0.9837, respectively. This study not only explored the spatial resource allocation of major epidemics in urban management, but also provided a basis for regional spatial prevention and control strategies.

Key words: COVID-19, multivariate data, geospatial epidemic risk, spatial autocorrelation, pearson correlation, spatial stratified heterogeneity, influencing factors, epidemic prevention and control