Geographically varying relationships between population flows from Wuhan and COVID-19 cases in Chinese cities

• Main Authors: Gang Xu, Wenwu Wang, Dandan Lu, Binbin Lu, Kun Qin, Limin Jiao
• Format: Article
• Language: English
• Published: Taylor & Francis Group 2021-09-01
• Series: Geo-spatial Information Science
• Subjects: covid-19; population flow; scaling law; geographically weighted regression (gwr); spatial heterogeneity
• Online Access: http://dx.doi.org/10.1080/10095020.2021.1977093

The COVID-19 epidemic widely spread across China from Wuhan, Hubei Province, because of huge migration before 2020 Chinese New Year. Previous studies demonstrated that population outflows from Wuhan determined COVID-19 cases in other cities but neglected spatial heterogeneities of their relationships. Here, we use Geographically Weighted Regression (GWR) model to investigate the spatially varying influences of outflows from Wuhan. Overall, the GWR model increases explanatory ability of outflows from Wuhan by 20%, with the adjusted R2 increasing from ~0.6 of Ordinary Least Squares (OLS) models to ~0.8 of GWR models. The coefficient between logarithmic of outflows from Wuhan and COVID-19 cases in other cities is generally less than 1. The sub-linear scaling relationship indicates the increasing returns of outflows was restrained, proving the epidemic was efficiently controlled outside Hubei at the beginning without obvious local transmissions. Coefficients in GWR models vary in cities. Not only cities around Wuhan but also cities having close connections with Wuhan experienced higher coefficients, showing a higher vulnerability of these cities. The secondary or multi-level transmission networks deserve to be further explored to fully uncover influences of migrations on the COVID-19 pandemic.