Ambient air pollutants concentration prediction during the COVID-19: A method based on transfer learning.

Research on the correlation analysis between COVID-19 and air pollution has attracted increasing attention since the COVID-19 pandemic. While many relevant issues have been widely studied, research into ambient air pollutant concentration prediction (APCP) during COVID-19 is still in its infancy. Most of the existing study on APCP is based on machine learning methods, which are not suitable for APCP during COVID-19 due to the different distribution of historical observations before and after the pandemic. Therefore, to fulfill the predictive task based on the historical observations with a different distribution, this paper proposes an improved transfer learning model combined with machine learning for APCP during COVID-19. Specifically, this paper employs the Gaussian mixture method and an optimization algorithm to obtain a new source domain similar to the target domain for further transfer learning. Then, several commonly used machine learning models are trained in the new source domain, and these well-trained models are transferred to the target domain to obtain APCP results. Based on the real-world dataset, the experimental results suggest that, by using the improved machine learning methods based on transfer learning, our method can achieve the prediction with significantly high accuracy. In terms of managerial insights, the effects of influential factors are analyzed according to the relationship between these influential factors and prediction results, while their importance is ranked through their average marginal contribution and partial dependence plots.

Associations of Ambient Air Pollutants and Meteorological Factors With COVID-19 Transmission in 31 Chinese Provinces: A Time Series Study.

Evidence regarding the effects of environmental factors on COVID-19 transmission is mixed. We aimed to explore the associations of air pollutants and meteorological factors with COVID-19 confirmed cases during the outbreak period throughout China. The number of COVID-19 confirmed cases, air pollutant concentrations, and meteorological factors in China from January 25 to February 29, 2020, (36 days) were extracted from authoritative electronic databases. The associations were estimated for a single-day lag as well as moving averages lag using generalized additive mixed models. Region-specific analyses and meta-analysis were conducted in 5 selected regions from the north to south of China with diverse air pollution levels and weather conditions and sufficient sample size. Nonlinear concentration-response analyses were performed. An increase of each interquartile range in PM2.5, PM10, SO2, NO2, O3, and CO at lag4 corresponded to 1.40 (1.37-1.43), 1.35 (1.32-1.37), 1.01 (1.00-1.02), 1.08 (1.07-1.10), 1.28 (1.27-1.29), and 1.26 (1.24-1.28) ORs of daily new cases, respectively. For 1°C, 1%, and 1 m/s increase in temperature, relative humidity, and wind velocity, the ORs were 0.97 (0.97-0.98), 0.96 (0.96-0.97), and 0.94 (0.92-0.95), respectively. The estimates of PM2.5, PM10, NO2, and all meteorological factors remained significantly after meta-analysis for the five selected regions. The concentration-response relationships showed that higher concentrations of air pollutants and lower meteorological factors were associated with daily new cases increasing. Higher air pollutant concentrations and lower temperature, relative humidity and wind velocity may favor COVID-19 transmission. Controlling ambient air pollution, especially for PM2.5, PM10, NO2, may be an important component of reducing risk of COVID-19 infection. In addition, as winter months are arriving in China, the meteorological factors may play a negative role in prevention. Therefore, it is significant to implement the public health control measures persistently in case another possible pandemic.

Temporal air quality (NO2, O3, and PM10) changes in urban and rural stations in Catalonia during COVID-19 lockdown: an association with human mobility and satellite data.

In this study, changes in air quality by NO2, O3, and PM10 in Barcelona metropolitan area and other parts of Catalonia during the COVID-19 lockdown with respect to pre-lockdown and to previous years (2018 and 2019) were evaluated. Selected air monitoring stations included 3 urban (Gràcia, Vall d'Hebron, and Granollers), 1 control site (Fabra Observatory), 1 semi-urban (Manlleu), and 3 rural (Begur, Bellver de Cerdanya, and Juneda). NO2 lockdown levels showed a diminution, which in relative terms was maximum in two rural stations (Bellver de Cerdanya, - 63% and Begur, - 61%), presumably due to lower emissions from the ceasing hotel and ski resort activities during eastern holidays. In absolute terms and from an epidemiologic perspective, decrease in NO2, also reinforced by the high amount of rainfall registered in April 2020, was more relevant in the urban stations around Barcelona. O3 levels increased in the transited urban stations (Gràcia, + 42%, and Granollers, + 64%) due to the lower titration effect by NOx. PM10 lockdown levels decreased, mostly in Gràcia, Vall d'Hebron, and Granollers (- 35, - 39%, and - 39%, respectively) due to traffic depletion (- 90% in Barcelona's transport). Correlation among mobility index in Barcelona (- 100% in retail and recreation) and contamination was positive for NO2 and PM10 and negative for O3 (P < 0.001). Satellite images evidenced two hotspots of NO2 in Spain (Madrid and Barcelona) in April 2018 and 2019 that disappeared in 2020. Overall, the benefits of lockdown on air quality in Catalonia were evidenced with NO2, O3 and PM10 levels below WHOAQG values in most of stations opposed to the excess registered in previous years.