COVID-19-Induced Lockdowns Indicate the Short-Term Control Effect of Air Pollutant Emission in 174 Cities in China

• Published in: Environmental science & technology Vol. 55; no. 7; pp. 4094 - 4102
• Main Authors: Lu, Dawei, Zhang, Jingwei, Xue, Chaoyang, Zuo, Peijie, Chen, Zigu, Zhang, Luyao, Ling, Weibo, Liu, Qian, Jiang, Guibin
• Format: Journal Article
• Language: English
• Published: United States American Chemical Society 06.04.2021
• Subjects: Air Pollutants - analysis; Air pollution; Air Pollution - analysis; Air Pollution - prevention & control; Air pollution control; Air pollution effects; Anthropogenic Impacts on the Atmosphere; China; Cities; Communicable Disease Control; Coronaviruses; COVID-19; Emergency preparedness; Emergency response; Emissions; Emissions control; Humans; Learning algorithms; Machine learning; Megacities; Nitrogen dioxide; Outdoor air quality; Pandemics; Particulate matter; Particulate Matter - analysis; Pollutants; Pollution control; Pollution effects; Pollution levels; Prediction models; Public health; SARS-CoV-2; Sulfur dioxide; Urban planning; Weather forecasting; China; Cities; COVID-19; air pollution; public health emergency; machine learning; emission reduction
• ISSN: 0013-936X; 1520-5851
• DOI: 10.1021/acs.est.0c07170

The contradiction between the regional imbalance and an one-size-fits-all policy is one of the biggest challenges in current air pollution control in China. With the recent implementation of first-level public health emergency response (FLPHER) in response to the COVID-19 pandemic in China (a total of 77 041 confirmed cases by February 22, 2020), human activities were extremely decreased nationwide and almost all economic activities were suspended. Here, we show that this scenario represents an unprecedented “base period” to probe the short-term emission control effect of air pollution at a city level. We quantify the FLPHER-induced changes of NO2, SO2, PM2.5, and PM10 levels in 174 cities in China. A machine learning prediction model for air pollution is established by coupling a generalized additive model, random effects meta-analysis, and weather research and forecasting model with chemistry analysis. The short-term control effect under the current energy structure in each city is estimated by comparing the predicted and observed results during the FLPHER period. We found that the short-term emission control effect ranges within 53.0%–98.3% for all cities, and southern cities show a significantly stronger effect than northern cities (P < 0.01). Compared with megacities, small–medium cities show a similar control effect on NO2 and SO2 but a larger effect on PM2.5 and PM10.