Atkinson SF, Kala AK, Tiwari C. Cascading effects of COVID-19 on population mobility and air quality: An exploration including place characteristics using geovisualization.
GEOSPATIAL HEALTH 2022;
17. [PMID:
35179013 DOI:
10.4081/gh.2022.1056]
[Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 11/19/2021] [Accepted: 01/22/2022] [Indexed: 06/14/2023]
Abstract
This study hypothesizes that public health responses to coronavirus disease 2019 (COVID-19), including a mandated restriction of activity (commonly called a 'lockdown') resulted in reduced transportation activities and changes in air quality in Texas, USA. This presented a natural experiment where population mobility and air quality before and after the lockdown could be compared. Changes in mobility were measured by SafeGraph mobility data (from opt-in smart phone applications that transmit location data) and air quality changes were based on NO2 concentrations measured by the European Space Agency's Sentinel-5 Precursor satellite (from the TROPOspheric Monitoring Instrument). The changes in population mobility and NO2 concentration between mid-March 2020 (lockdown initiated) and the end of 2020, as compared to the same time window in 2019, were the basis of exploring the lockdown hypothesis. Additionally, numerous socio-economic (place based) indicators were hypothesized to follow public health vulnerability assumptions based on COVID- 19 incidence patterns. This hypothesis was subjected to geovisualization techniques in order to find potential patterns and insights into the complex combinations of these place-based data. Our results suggest that simultaneously visualizing COVID-19, mobility, air quality and socio-economic data yields insights in underlying spatial processes related to public health policy decisions. The hypothesis that the lockdown resulted in reduced mobility and NO2 concentrations was found partially correct - this trend was observed in highly urbanized areas, but not in less populated areas. Data related public health vulnerability assumptions (e.g. a region's age, poverty, education, etc.) were agreed with in part, but disagreed with in part.
Collapse