Time series analysis of COVID-19 stringency measures on the spatiotemporal dynamics of air pollution

Abstract

Introduction: Execution of COVID-19 lockdown measures caused variations in air pollution worldwide. This paper investigates the impact of COVID-19 stringency measures on the spatio-temporal dynamics of air pollution in Mumbai, India, using a comprehensive two-and-a-half-year pandemic period dataset. Materials and methods: We classified the pandemic period into 7 phases and 21 sub-phases based on the severity of the Oxford COVID-19 Government Response Tracker (OxCGRT) Stringency Index (SI). Optimized Hotspot analysis (OHS) and Ordinary Least Square Regression models explored the spatio-temporal fluctuations and the effect of stringency measures on air quality. Results: The R2 value varied; with the best model R2 of 0.61 for Particulate Matters (PM10) and Nitrogen dioxide (NO2) and lowest of 0.23 for Sulfur dioxide (SO2). A 10-point increase in SI caused a 3-7% reduction in air pollutants. Substantial reduction in average PM10, PM2.5, NO2, and Carbon monoxide (CO) was observed throughout the COVID-19 phases. Meteorology and SI collectively caused maximum reduction of 82.6%, 72.7%, 53.8%, 52.2%, 49.1%, 28.4% for NO2, PM2.5, PM10, NH3, CO, and SO2 respectively, during complete or extreme lockdown phases. Except SO2, seasonality significantly influenced the pollutant concentrations. Winter was the worst period while monsoon was the best. OHS identified central Mumbai wards as hotspots and areas close to the national park as coldspots. Conclusion: PM10, NO2 and CO were more affected by SI measures than NH3 and SO2. For a rapid emergency response to high PM10, implementation of SI, very high (≥ 80 score) and above is advised. Findings of this study have significant public health policy implications, especially among global south nations.

Keywords

• COVID-19; Air pollution; Stringency index; Regression models; Hotspot analysis