Understanding and revealing the intrinsic impacts of the COVID-19 lockdown on air quality and public health in North China using machine learning

Impact of COVID-19 restrictions liberalization on air quality: a case study of Chongqing, Southwest China

To mitigate the societal impact of the COVID-19 pandemic, China implemented long-term restrictive measures. The sudden liberalization at the end of 2022 disrupted residents' daily routines, making it scientifically intriguing to explore its effect on air quality. Taking Chongqing City in Southwest China as an example, we examined the impact of restriction liberalization on air quality, identified potential sources of pollutants, simulated the effects of abrupt anthropogenic control relaxation using a Random Forest Model, and applied an optimized model to predict the post-liberalization pollutant concentrations. The results showed increases in PM2.5 (72.3%), PM10 (67.7%), and NO2 (21.9%) concentrations, while O3 concentration decreased by 20.5%. Although potential pollution source areas contracted, pollution levels intensified with northeastern Sichuan, interior Chongqing, and northern Guizhou being major contributors to pollutant emissions. Anthropogenic emissions accounted for 26.7 ~ 33% changes in PM2.5 and PM10 concentrations while meteorological conditions contributed to 40.2 ~ 43.3% variations observed during the period. The optimized model demonstrated a correlation between predicted and observed values with R2 ranging from 0.70 to 0.89, enabling accurate prediction of post-liberalization pollutant concentrations. This study can enhance our understanding regarding the impact of sudden social lockdown relaxation events on air quality while providing support for urban air pollution prevention.

Quantifying the pollution changes and meteorological dependence of airborne trace elements coupling source apportionment and machine learning

Airborne trace elements (TEs) present in atmospheric fine particulate matter (PM2.5) exert notable threats to human health and ecosystems. To explore the impact of meteorological conditions on shaping the pollution characteristics of TEs and the associated health risks, we quantified the variations in pollution characteristics and health risks of TEs due to meteorological impacts using weather normalization and health risk assessment models, and analyzed the source-specific contributions and potential sources of primary TEs affecting health risks using source apportionment approaches at four sites in Shandong Province from September to December 2021. Our results indicated that TEs experience dual effects from meteorological conditions, with a tendency towards higher TE concentrations and related health risks during polluted period, while the opposite occurred during clean period. The total non-carcinogenic and carcinogenic risks of TEs during polluted period increased approximately by factors of 0.53-1.74 and 0.44-1.92, respectively. Selenium (Se), manganese (Mn), and lead (Pb) were found to be the most meteorologically influenced TEs, while chromium (Cr) and manganese (Mn) were identified as the dominant TEs posing health risks. Enhanced emissions of multiple sources for Cr and Mn were found during polluted period. Depending on specific wind speeds, industrialized and urbanized centers, as well as nearby road dusts, could be key sources for TEs. This study suggested that attentions should be paid to not only the TEs from primary emissions but also the meteorology impact on TEs especially during pollution episodes to reduce health risks in the future.

Anthropogenic-driven changes in concentrations and sources of winter volatile organic compounds in an urban environment in the Yangtze River Delta of China between 2013 and 2021

Volatile organic compounds (VOCs) serve as crucial precursors to surface ozone and secondary organic aerosols (SOA). In response to severe air pollution challenges, China has implemented key air quality control policies from 2013 to 2021. Despite these efforts, a comprehensive understanding of the chemical composition and sources of urban atmospheric VOCs and their responses to emission reduction measures remains limited. Our study focuses on analyzing VOCs composition and concentrations during the winters of 2013 and 2021 through online field observations in urban Nanjing, a typical city in the Yangtze River Delta region of China. Using a machine learning approach, we found a notable reduction in total VOCs concentration from 52.4 ± 30.4 ppb to 33.9 ± 21.6 ppb between the two years, with dominant contributions (approximately 94.3 %) associated with anthropogenic emission control. Furthermore, alkanes emerged as the major contributors (48.6 %) to such anthropogenic-driven decline. The total SOA formation potential decreased by approximately 27.4 %, with aromatics identified as the major contributing species. Positive matrix factorization analysis identified six sources. In 2013, prominent contributors were solid fuel combustion (43.6 %), vehicle emission (16.7 %), and paint and solvent use (12.8 %). By 2021, major sources shifted to solid fuel combustion (31.9 %), liquefied petroleum gas and natural gas (26.8 %), and vehicle emission (25.5 %). Solid fuel combustion emerged as the primary driver for total VOCs reduction. The lifetime carcinogenic risk in 2021 decreased by 72.6 % relative to 2013, emphasizing the need to address liquefied petroleum gas and natural gas source, and vehicle emissions for improved human health. Our findings contribute critical insights for policymakers working on effective air quality management.

Response of PM2.5 chemical composition to the emission reduction and meteorological variation during the COVID-19 lockdown

PM2.5 is a main atmospheric pollutant with various sources and complex chemical compositions, which are influenced by various factors, such as anthropogenic emissions (AE) and meteorological conditions (MC). MC have a significant impacts on variations in atmospheric pollutant; therefore, emission reduction policies and ambient air quality are non-linearly correlated, which hinders the accurate assessment of the effectiveness of control measures. In this study, we conducted online observations of PM2.5 and its chemical composition in Hohhot, China, from December 1, 2019, to February 29, 2020, to investigate how the chemical compositions of PM2.5 respond to the variations in AE and MC. Moreover, the random forest (RF) model was used to quantify the contributions of AE and MC to PM2.5 and its chemical composition during severe hazes and the COVID-19 pandemic lockdown period. During the clean period, MC reduced PM2.5 concentrations by 124%, while MC incresed PM2.5 concentrations by 49% during severe pollution episode. Inorganic aerosols (SO42-, NO3-, and NH4+) showed the strongest response to MC. MC significantly contributed to PM2.5 (36%), SO42- (32%), NO3- (29%), NH4+ (28%), OC (22%), and SOC (17%) levels during pollution episodes. From the pre-lockdown to lockdown period, AE (MC) contributed 52% (48%), 81% (19%), 48% (52%), 68% (32%), 59% (41%), and 288% (-188%) to the PM2.5, SO42-, NO3-, NH4+, OC, and SOC reductions, respectively. The variations in MC (especially the increase in relative humidity) rapidly generated meteorologically sensitive species (SO42-, NO3-, and NH4+), which led to severe winter pollution. This study provides a reference for assessing the net benefits of emission reduction measures for PM2.5 and its chemical compositions.

AI-based prediction of the improvement in air quality induced by emergency measures

Several cities in the developing world, of which the capital city of India, New Delhi, is an example, often experience air quality in which pollutant levels go way above the levels considered hazardous for human health. To bring down the air quality to within permissible limits quickly, the measures typically taken involve shutting down certain high-polluting activities for some time to enable the air quality to recover temporarily. This paper presents a first-ever model based on artificial neural networks to forecast the extent of reduction in air quality parameters that can be achieved and the time period within which a change can be experienced when the source of the emissions is cut off temporarily. The model is based on the extensive data on the extent of reduction in air quality parameters that occurred during the lockdown that was imposed during the COVID-19 pandemic. The non-linear autoregressive exogenous network-based model chosen for the purpose employs the hour since stopping of emissions, relative humidity, wind speed, wind direction, and ambient temperature as input parameters to predict the rate of change of PM2.5 with respect to the concentration at the start of the stopping of the emissions. Air quality data from a key monitoring station in New Delhi was used to develop the model. The model predicted the rate of drop in PM2.5 with an R and MSE of 0.0044 and 0.9736, respectively, while training and 0.0095 and 0.9583 while testing. The model was then tested with data from 19 other stations in New Delhi, and accuracy of the model was found to be exceptionally accurate, with the correlation between the measured and the predicted PM2.5 levels ranging from 0.74 to 0.94 and the MSE ranging from 0.0110 to 1.0746. Thus, the model can be employed to determine the number of hours of temporary stoppage of emissions required for the PM2.5 concentration to reach safe levels. The methodology of development of the model can be extrapolated to construct models tailored for use in other parts of the world as well.

Development of surface observation-based two-step emissions adjustment and its application on CO, NOx, and SO2 emissions in China and South Korea

It is challenging to estimate local emission conditions of a downwind area solely based on concentrations in the downwind area. This is because air pollutants that have a long residence time in the atmosphere can be transported over long distances and influence air quality in downwind areas. In this study, a Two-step Emissions Adjustment (TEA) approach was developed to adjust downwind emissions of target air pollutants with surface observations, considering their long-range transported emission impacts from upwind areas calculated from air quality simulations. Using the TEA approach, CO, NOx, and SO2 emissions were adjusted in China and South Korea between 2016 and 2021 based on existing bottom-up emissions inventories. Simulations with the adjusted emissions showed that the 6-year average normalized mean biases of the monthly mean concentrations of CO, NOx, and SO2 improved to 0.3 %, -2 %, and 2 %, respectively, in China, and to 5 %, 7 %, and 4 %, respectively, in South Korea. When analyzing the emission trends, it was estimated that the annual emissions of CO, NOx, and SO2 in China decreased at a rate of 7.2 %, 4.5 %, and 10.6 % per year, respectively. The decrease rate of emissions for each of these pollutants was similar to that of ambient concentrations. When considering upwind emission impacts in the emissions adjustment, CO emissions increased by 1.3 %/year in South Korea, despite CO concentrations in the country decreasing during the study period. During the study period, NOx and SO2 emissions in South Korea decreased by 3.9 % and 0.5 %/year, respectively. Moreover, the TEA approach can account for drastic short-term emission changes (e.g., social distancing due to COVID-19). Therefore, the TEA approach can be used to adjust emissions and improve reproducibility of concentrations of air pollutants suitable for health studies for areas where upwind emission impacts are significant.

Drivers and impacts of decreasing concentrations of atmospheric volatile organic compounds (VOCs) in Beijing during 2016-2020

China has implemented various policies and measures for controlling air pollutants. However, our knowledge of the long-term trends in ambient volatile organic compounds (VOCs) after the implementation of these action plans in China remains limited. To address this, we conducted a five-year analysis (2016-2020) of VOC compositions and concentrations in Beijing. The annual VOC concentration decreased from 44.0 ± 28.8 to 26.2 ± 16.4 ppbv, with alkanes being the most prevalent group. The annual average concentrations of alkenes, alkynes, and aromatics have experienced a significant decrease of over 50 %. Seasonal variations indicated higher VOC concentrations in winter and autumn, with more significant reductions observed in winter and autumn. The impact of meteorological conditions caused variations in VOC reductions during the Chinese Spring Festival. Satellite-based measurements of formaldehyde (HCHO) columns confirmed the reduction of VOC emissions during the Coronavirus (COVID-19) lockdown. The normalized annual average VOC concentration decreased by 2.9ppbv yr-1 from 2016 to 2020, and emission reduction contributed to 58.8 % of VOC reduction from 2016 to 2020 after meteorological normalization, indicating the effectiveness of implemented control measures. Based on receptor model, vehicle emissions and industrial sources were identified as the largest contributors to VOC concentrations. Vehicle emissions, liquefied petroleum gas/natural gas (LPG/NG) use, and coal combustion were major drivers of VOC reduction. Potential source region analysis revealed that air masses transported from northwestern and southern regions significantly contributed to VOC concentrations in Beijing. The range of source regions shrunk in both northwestern and southern regions with the reduction in VOC concentrations. The annual variations of ozone formation potential indicated a significant decrease in VOC reactivities through emission control. These results could provide insights into future emission control and coordinated efforts to improve PM2.5 and ozone levels in China.

The application of machine learning to air pollution research: A bibliometric analysis

Machine learning (ML) is an advanced computer algorithm that simulates the human learning process to solve problems. With an explosion of monitoring data and the increasing demand for fast and accurate prediction, ML models have been rapidly developed and applied in air pollution research. In order to explore the status of ML applications in air pollution research, a bibliometric analysis was made based on 2962 articles published from 1990 to 2021. The number of publications increased sharply after 2017, comprising approximately 75% of the total. Institutions in China and United States contributed half of all publications with most research being conducted by individual groups rather than global collaborations. Cluster analysis revealed four main research topics for the application of ML: chemical characterization of pollutants, short-term forecasting, detection improvement and optimizing emission control. The rapid development of ML algorithms has increased the capability to explore the chemical characteristics of multiple pollutants, analyze chemical reactions and their driving factors, and simulate scenarios. Combined with multi-field data, ML models are a powerful tool for analyzing atmospheric chemical processes and evaluating the management of air quality and deserve greater attention in future.

Exploring the driving factors of haze events in Beijing during Chinese New Year holidays in 2020 and 2021 under the influence of COVID-19 pandemic

Unexpected outbreak of the 2019 novel coronavirus (COVID-19) has profoundly altered the way of human life and production activity, which posed visible impacts on PM_2.5 and its chemical species. The abruptly emergency reduction in human activities provided an opportunity to explore the synergetic impacts of multi-factors on shaping PM_2.5 pollution. Here, we conducted two comprehensive observation measurements of PM_2.5 and its chemical species from 1 January to 16 February in Beijing 2020 and the same lunar date in 2021, to investigate temporal variations and reveal the driving factors of haze before and after Chinese New Year (CNY). Results show that mean PM_2.5 concentrations during the whole observation were 63.83 and 66.86 μg/m³ in 2020 and 2021, respectively. Higher secondary inorganic species were observed after CNY, and K⁺, Cl^- showed three prominent peaks which associated closely with fireworks burnings from suburb Beijing and surroundings, verifying that they could be used as two representative tracers of fireworks. Further, we explored the impacts of meteorological conditions, regional transportation as well as chemical reactions on PM_2.5. We found that unfavorable meteorological conditions accounted for 11.0 % and 16.9 % of PM_2.5 during CNY holidays in 2020 and 2021, respectively. Regional transport from southwest and southeast (south) played an important role on PM_2.5 during the two observation periods. Higher ratio of NO₃^-/SO₄^2- were observed under high O_X and low RH conditions, suggesting the major pathway of NO₃^- and SO₄^2- formation could be photochemical process and aqueous-phase reaction. Additionally, nocturnal chemistry facilitated the formation of secondary components of both inorganic and organic. This study promotes understandings of PM_2.5 pollution in winter under the influence of COVID-19 pandemic and provides a well reference for haze and PM_2.5 control in future.