Variational graph encoders effectively combine graph convolutional networks with variational autoencoders, and have been widely employed for biomedical graph-structured data. Lam and colleagues developed a framework based on the variational graph encoder, NYAN, to facilitate the prediction of molecular properties in computer-assisted drug design. In NYAN, the low-dimensional latent variables derived from the variational graph autoencoder are leveraged as a universal molecular representation, yielding remarkable performance and versatility throughout the drug discovery process. In this study we assess the reusability of NYAN and investigate its applicability within the context of specific chemical toxicity prediction. The prediction accuracy—based on NYAN latent representations and other popular molecular feature representations—is benchmarked across a broad spectrum of toxicity datasets, and the adaptation of NYAN latent representation to other surrogate models is also explored. NYAN, equipped with common surrogate models, shows competitive or better performance in toxicity prediction compared with other state-of-the-art molecular property prediction methods. We also devise a multi-task learning strategy with feature enhancement and consensus inference by leveraging the low dimensionality and feature diversity of NYAN latent space, further boosting the multi-endpoint acute toxicity estimation. The analysis delves into the adaptability of the generic graph variational model, showcasing its aptitude for tailored tasks within the realm of drug discovery. Ruijiang Li et al. assess the reproducibility of a variational graph encoder-based framework and examines its reusability for chemical toxicity prediction. It explores how a generalist model can function as a specialist model with adaptation.

This study investigates the transport of air pollutants around the Yangtze River Delta with an aim to identify if there would be a relationship towards health effects during the COVID-19 lockdown period. It is well-known that due to lockdown, the number of socio-economical activities are reduced and hence there is an observable reduction in air pollution. We would like to investigate if this consequential reduction of air pollution would lead to improvement in health amongst its population. A number of integrated methodologies are utilized, including collection and correlation of statistical data and numerical modeling to correlate the mortality rates difference with and without COVID-19 lockdown. In particular air quality changes during the COVID-19 lockdown period are compared with similar periods of the previous years using Brute Force Method. It is found that in general there is significant reduction in air-pollution related mortality, like stroke, ischemic cardio diseases, obstructive pulmonary disease, lung cancer and acute lower respiratory infection are all reduced as a result of relative improvement in PM2.5 level during the lockdown period. Further investigation of the trajectories suggests that these PM2.5 originate from afar with multiple sources, and do not suggest COVID-19 are transported to the region via long-range transport. Our results demonstrate the need for more stringent policy measure to tackle air pollution as it has strong evidence that it increases mortality rate.

This study investigates the transport of air pollutants around the Yangtze River Delta with an aim to identify if there would be a relationship towards health effects during the COVID-19 lockdown period. It is well-known that due to lockdown, the number of socio-economical activities are reduced and hence there is an observable reduction in air pollution. We would like to investigate if this consequential reduction of air pollution would lead to improvement in health amongst its population. A number of integrated methodologies are utilized, including collection and correlation of statistical data and numerical modeling to correlate the mortality rates difference with and without COVID-19 lockdown. In particular air quality changes during the COVID-19 lockdown period are compared with similar periods of the previous years using Brute Force Method. It is found that in general there is significant reduction in air-pollution related mortality, like stroke, ischemic cardio diseases, obstructive pulmonary disease, lung cancer and acute lower respiratory infection are all reduced as a result of relative improvement in PM2.5 level during the lockdown period. Further investigation of the trajectories suggests that these PM2.5 originate from afar with multiple sources, and do not suggest COVID-19 are transported to the region via long-range transport. Our results demonstrate the need for more stringent policy measure to tackle air pollution as it has strong evidence that it increases mortality rate.

Open biomass burning (OBB) plays an important role in air pollution and climate change by releasing short-term but intensive amounts of particulate matter and gaseous air pollutants. During past years, policies with respect to prohibition on open straw burning have been issued in China in order to mitigate the air pollution problems and the effectiveness of these straw burning bans in different regions remains to be evaluated. In this study, open crop straw burning (OCSB) emissions during 2010–2018 were analyzed based on a commonly used emission inventory with high spatial and temporal resolution. High emissions concentrated over Northeast China (31.8% of national total PM2.5 emissions in 2018), East China (24.0%), and North China (16.6%). Simulations based on an integrated meteorology-air quality modeling system and an exposure-response function show that OCSB emissions could increase monthly PM2.5 concentration by as much as 10 μg/m3 during burning seasons in Northeast China and were associated with 4741 premature deaths in 2018. Spatial heterogeneities were observed with respect to the trends of OCSB emissions during 2010–2018. In East China, North China, and Central China, OCSB emissions showed a general declining trend since 2013 while an opposing increasing trend was observed in Northeast China with peak emissions in 2017. Comparing 2013 (before intensive implementation of straw burning bans) and 2018 (after), national total PM2.5 emissions from OCSB activities decreased by 46.9%, ranging from −14.1% to +70% depending on the specific regions. Northeast China is the only region that showed higher OCSB emissions in 2018 compared to 2013, probably associated with the relatively delayed implementation of the straw burning bans. Avoided number of premature deaths due to reduced OCSB emissions was estimated to be 4256 on a national scale, with most health benefits gained in East and Central China. Results from this study demonstrate the importance of OCSB contribution to PM2.5 concentrations and spatial heterogeneities exist in terms of the effectiveness of the straw burning bans in reducing OCSB emissions and gained health benefits.

Biogenic volatile organic compounds (BVOCs) play an important role in atmospheric chemistry due to their large quantities and high reactivity. In this study, the impacts of BVOC emissions on ozone formation were investigated based on MEGANv3.1 in the Yangtze River Delta (YRD) region, which has increasingly suffered from ozone pollution in recent years. The sensitivities of BVOC emissions to different drought stress configurations and the quality of emission factors were evaluated. Furthermore, BVOC contribution to ozone formation was simulated by integrated meteorology and air quality model system and the impacts of different BVOC emission scenarios on ozone concentration were discussed. Annual BVOC emissions estimated with the default drought stress configuration (i.e., base case) was 6.8 × 10 5 tons. The drought stress algorithm implemented in MEGANv3.1 could suppress BVOC emissions by 58% and this algorithm was sensitive to the choice of wilting point values. The BVOC contribution to the average of daily maximum 8 h ozone concentration without drought stress effect in July 2016 was 104% higher than that in the base scenario when the drought stress effect is activated. Using an alternative set of wilting point led to BVOC contribution being 48% higher than that in the base scenario. High contributions of BVOCs to simulated ozone concentration were found in northern Zhejiang, especially in Hangzhou and its surrounding areas.

The outbreak of coronavirus (COVID-19) has forced China to lockdown many cities and restrict transportation, industrial, and social activities. This provides a great opportunity to look at the impacts of pandemic quarantine on air quality and premature death due to exposure to air pollution. In this study, we applied the difference-in-differences (DID) model to quantify the casual impacts of COVID-19 lockdown on air quality at 278 cities across China. A widely used exposure-response function was further utilized to estimate the short-term health impacts associated with changes in PM2.5 due to lockdown. Results show that lockdown has caused drastic reduction in air pollution level in terms of all criteria pollutants except ozone. On average, concentrations of PM2.5, PM10, NO2, SO2 and CO are estimated to drop by 14.3 μg/m3, 22.2 μg/m3, 17.7 μg/m3, 2.9 μg/m3, and 0.18 mg/m3 as the result of lockdown. Cities with more confirmed cases of COVID-19 are related to stronger responses in air quality, despite that similar lockdown measures were implemented by the local governments. The improvement of air quality caused by COVID-19 lockdown in northern cities is found to be smaller than that of southern cities. Avoided premature death associated with PM2.5 exposures over the 278 cities was estimated to be 50.8 thousand. Our results re-emphasize the effectiveness of emission controls on air quality and associated health impacts. The high cost of lockdown, still high level of air pollution during lockdown and smaller effects in northern cities implies that source-specific mitigation policies are needed for continuous and sustainable reduction of air pollution.

In the recent decade, the North China Plain (NCP) has been among the region’s most heavily polluted by PM2.5 in China. For the nonattainment cities in the NCP, joint pollution control with related cities is highly needed in addition to the emission controls in their own cities. However, as the basis of decision-making, the spatial characteristics of PM2.5 among these cities are still insufficiently revealed. In this work, the spatial characteristics among all nonattainment cities in the northern part of the North China Plain (NNCP) region were revealed based on data mining technologies including clustering, coefficient of divergence (COD), network correlation model, and terrain and meteorology analysis. The results indicate that PM2.5 pollution of cities with a distance of less than 180 km exhibits homogeneity in the NCP region. Especially, the sub-region, composed of Xinxiang, Hebi, Kaifeng, Zhengzhou, and Jiaozuo, was strongly homogeneous and a strong correlation exists among them. Compared with spring and summer, much stronger correlations of PM2.5 between cities were found in autumn and winter, indicating a strong need for joint prevention and control during these periods. All nonattainment cities in this region were divided into city-clusters, depending on the seasons and pollution levels to further helping to reduce their PM2.5 concentrations effectively. Air stagnation index (ASI) analysis indicates that the strong correlations between cities in autumn were more attributed to the transport impacts than those in winter, even though there were higher PM2.5 concentrations in winter. These results provided an insight into joint prevention and control of pollution in the NCP region.

Numerical air quality models (AQMs) have been applied more frequently over the past decade to address diverse scientific and regulatory issues associated with deteriorated air quality in China. Thorough evaluation of a model's ability to replicate monitored conditions (i.e., a model performance evaluation or MPE) helps to illuminate the robustness and reliability of the baseline modeling results and subsequent analyses. However, with numerous input data requirements, diverse model configurations, and the scientific evolution of the models themselves, no two AQM applications are the same and their performance results should be expected to differ. MPE procedures have been developed for Europe and North America, but there is currently no uniform set of MPE procedures and associated benchmarks for China. Here we present an extensive review of model performance for fine particulate matter (PM2.5) AQM applications to China and, from this context, propose a set of statistical benchmarks that can be used to objectively evaluate model performance for PM2.5 AQM applications in China. We compiled MPE results from 307 peer-reviewed articles published between 2006 and 2019, which applied five of the most frequently used AQMs in China. We analyze influences on the range of reported statistics from different model configurations, including modeling regions and seasons, spatial resolution of modeling grids, temporal resolution of the MPE, etc. Analysis using a random forest method shows that the choices of emission inventory, grid resolution, and aerosol- and gas-phase chemistry are the top three factors affecting model performance for PM2.5. We propose benchmarks for six frequently used evaluation metrics for AQM applications in China, including two tiers – “goals” and “criteria” – where goals represent the best model performance that a model is currently expected to achieve and criteria represent the model performance that the majority of studies can meet. Our results formed a benchmark framework for the modeling performance of PM2.5 and its chemical species in China. For instance, in order to meet the goal and criteria, the normalized mean bias (NMB) for total PM2.5 should be within 10 % and 20 %, while the normalized mean error (NME) should be within 35 % and 45 %, respectively. The goal and criteria values of correlation coefficients for evaluating hourly and daily PM2.5 are 0.70 and 0.60, respectively; corresponding values are higher when the index of agreement (IOA) is used (0.80 for goal and 0.70 for criteria). Results from this study will support the ever-growing modeling community in China by providing a more objective assessment and context for how well their results compare with previous studies and to better demonstrate the credibility and robustness of their AQM applications prior to subsequent regulatory assessments.

Accurate forecast of PM2.5 pollution is highly needed for the timely prevention of haze pollution in many cities suffered from frequent haze pollution. In this work, an online recurrent extreme learning machine (OR-ELM) technique with online data update was used in the forecast of PM2.5 pollution for the first time, and a hybrid model (OR-ELM-AR) by combining autoregressive (AR) model was proposed to enhance its forecast ability to capture the variations of hourly PM2.5 concentration. Evaluation of forecast performances in terms of pollution levels, forecast times, spatial distributions were conducted over the Yangtze River Delta (YRD) region, China. Results indicated that the OR-ELM-AR model could quickly respond to short-term changes and had better forecast performance. Therefore, the OR-ELM-AR model is a promising tool for air pollution forecast of supporting the government to take urgent actions to reduce the frequency and severity of haze pollution in cities or regions.

Although many efforts have been done to reduce PM2.5 pollution in recent years, there are still many nonattainment cities in East China for national standards. How to collaborate with surrounding cities to reduce PM2.5 has been a critical issue for all nonattainment cities. In this work, a framework of developing a practical scheme system of joint prevention and control of PM2.5 pollution based on data mining technologies was presented. Moreover, one specific scheme system for Yangtze River Delta region was developed and evaluated by five indicators. The results indicate that the scheme system was divided into three grades depending on pollution levels except for summer with two grades. The heavier the PM2.5 pollution was, the higher the grade was recommended to be implemented along with low grades. Moreover, more city-clusters were involved in each grade in winter than any other season. For Grade one in winter, the city-cluster with Xuzhou and Suqian was the most urgent to control emission jointly; the city-cluster with Suzhou and Shanghai was linked to the biggest impact on health, but for integrated indicator, the city-cluster with Nanjing and Changzhou ranked first. These results not only enable an enriched understanding of the significance of PM2.5 control in city-clusters, but also accelerate PM2.5 reduction for all nonattainment cities by collaboration with their most correlated cities in this region.

Abstract. Photochemical grid models (PGMs) are being applied more frequently to address diverse scientific and regulatory compliance associated with deteriorated air quality in China for the past decade. Solid evaluation of model performances guarantees the robustness and reliability of the baseline modelling results, so subsequent applications are built on top of it; thus, model performance evaluation (MPE) is a critical step of any PGM applications. MPE procedures and associated benchmarks have been proposed for PGM applications in the United States and Europe. However, with numerous input data needed, diverse model configurations, and evolution of the model itself, no two PGM applications are exactly the same. Therefore, those MPE benchmarks proposed based on studies outside China may not be appropriate for evaluation of the increasing number of PGM applications in China. Here we follow an established approach as published in previous literatures, to recommend statistical benchmarks for evaluation of simulated particulate matter (PM) concentrations in China. A total of 128 peer-reviewed articles published between 2006 and mid-2019 that applied one of four most frequently used PGMs in China are compiled to summarize operational model performance results. Quantile distributions of common statistical metrics are presented for total PM2.5 and speciated components. Influences of different model configurations, including modelling regions and seasons, spatial resolution of modelling grids, temporal resolution of MPE, etc., on the range of reported statistics are discussed. Benchmarks for four frequently used evaluation metrics are provided for two tiers – “goals” and “criteria”, where “goals” represent the best model performance that a model is currently expected to achieve and “criteria” represent the model performance that the majority (i.e. two thirds) of studies can meet. Our proposed benchmarks are further compared with those developed for United States and Europe. Additional recommendations for MPE practices are also given. Results from this study shall help the ever-growing modelling community in China to have a better objective assessment of how well their simulation results are compared with previous studies and to better demonstrate the credibility and robustness of their PGM applications prior to subsequent regulatory assessments.

Photochemical grid models (PGMs) are being applied more frequently to address diverse scientific and regulatory compliance associated with deteriorated air quality in China for the past decade. Solid evaluation of model performances guarantees the robustness and reliability of the baseline modelling results, so subsequent applications are built on top of it; thus, model performance evaluation (MPE) is a critical step of any PGM applications. MPE procedures and associated benchmarks have been proposed for PGM applications in the United States and Europe. However, with numerous input data needed, diverse model configurations, and evolution of the model itself, no two PGM applications are exactly the same. Therefore, those MPE benchmarks proposed based on studies outside China may not be appropriate for evaluation of the increasing number of PGM applications in China. Here we follow an established approach as published in previous literatures, to recommend statistical benchmarks for evaluation of simulated particulate matter (PM) concentrations in China. A total of 128 peer-reviewed articles published between 2006 and mid-2019 that applied one of four most frequently used PGMs in China are compiled to summarize operational model performance results. Quantile distributions of common statistical metrics are presented for total PM2.5 and speciated components. Influences of different model configurations, including modelling regions and seasons, spatial resolution of modelling grids, temporal resolution of MPE, etc., on the range of reported statistics are discussed. Benchmarks for four frequently used evaluation metrics are provided for two tiers – “goals” and “criteria”, where “goals” represent the best model performance that a model is currently expected to achieve and “criteria” represent the model performance that the majority (i.e. two thirds) of studies can meet. Our proposed benchmarks are further compared with those developed for United States and Europe. Additional recommendations for MPE practices are also given. Results from this study shall help the ever-growing modelling community in China to have a better objective assessment of how well their simulation results are compared with previous studies and to better demonstrate the credibility and robustness of their PGM applications prior to subsequent regulatory assessments.

The outbreak of COVID-19 has spreaded rapidly across the world. To control the rapid dispersion of the virus, China has imposed national lockdown policies to practise social distancing. This has led to reduced human activities and hence primary air pollutant emissions, which caused improvement of air quality as a side-product. To investigate the air quality changes during the COVID-19 lockdown over the YRD Region, we apply the WRF-CAMx modelling system together with monitoring data to investigate the impact of human activity pattern changes on air quality. Results show that human activities were lowered significantly during the period: industrial operations, VKT, constructions in operation, etc. were significantly reduced, leading to lowered SO2, NOx, PM2.5 and VOCs emissions by approximately 16-26%, 29-47%, 27-46% and 37-57% during the Level I and Level II response periods respectively. These emission reduction has played a significant role in the improvement of air quality. Concentrations of PM2.5, NO2 and SO2 decreased by 31.8%, 45.1% and 20.4% during the Level I period; and 33.2%, 27.2% and 7.6% during the Level II period compared with 2019. However, ozone did not show any reduction and increased greatly. Our results also show that even during the lockdown, with primary emissions reduction of 15%-61%, the daily average PM2.5 concentrations range between 15 and 79 μg m-3, which shows that background and residual pollutions are still high. Source apportionment results indicate that the residual pollution of PM2.5 comes from industry (32.2-61.1%), mobile (3.9-8.1%), dust (2.6-7.7%), residential sources (2.1-28.5%) in YRD and 14.0-28.6% contribution from long-range transport coming from northern China. This indicates that in spite of the extreme reductions in primary emissions, it cannot fully tackle the current air pollution. Re-organisation of the energy and industrial strategy together with trans-regional joint-control for a full long-term air pollution plan need to be further taken into account.

大气PM 2.5 和O 3 是中国城市大气中最受关注的大气污染物,它们之间存在着复杂的影响关系.本文将O 3,max 作为光化学活性指标,把CO作为一次排放源的示踪剂,对2017年4—10月期间北京、上海和广州的臭氧与二次气溶胶的协同增长关系进行深入分析.研究结果表明,不同光化学活性条件下,3个城市在协同增长时段PM 2.5 质量浓度增长量的均值与估算二次气溶胶质量浓度的变化趋势基本一致;北京和广州在春季、秋季出现协同增长的小时数较多,且北京在各光化学活性水平下的协同增长时段中PM 2.5 浓度增长量均最大;出现PM 2.5 和O 3 协同增长现象时段中上海市风速最大,广州市气温最高,北京市相对湿度范围最大;随着光化学活性的增强,上海和广州两市二次气溶胶的生成量呈倍数增加,其生成量范围分别为13.6～29.2μg·m -3 和9.1～28.7μg·m -3 ,北京市二次气溶胶的生成量则变化不大(25.0～34.0μg·m -3 ),但各光化学活性水平下的北京二次气溶胶生成量均高于上海和广州.这些研究结果表明北京、上海和广州3个城市的O 3 污染对二次气溶胶生成协同增长作用明显,而且北京尤其值得关注.

In this study, we analyzed several pollution episodes that occurred in the autumn and winter of 2018-2019 using multiple methods including the hierarchical clustering analysis, backward trajectory, and potential source contribution analysis based on monitored air quality and meteorological data. Bengbu, being a representative city to the north of the Yangtze River Delta (YRD) region and located in a heavily polluted area during these two pollution processes, is the focus of this work. The results indicated that the northern part of the YRD region is affected because of unfavorable meteorological conditions such as weak ground pressure, high humidity, low temperature, low wind speeds, and regional transport. The regional pollution processes over the YRD region in the autumn and winter seasons exhibit characteristics of wide influence and long duration with mainly two types of pollution:regional transport and intra-regional accumulation. During the two selected pollution episodes, the average PM2.5 concentration in the northern YRD region reached 131.6 μg·m-3 and 115.4 μg·m-3, respectively. The former type had a shorter duration but exhibited rapid accumulation of pollutants in a short period of time with greater pollution intensity, wider pollution range, and deeper pollution intrusion. Qualitative and quantitative analysis of the potential sources of PM2.5 based on PSCF and CWT showed that the PM2.5 concentration during EP1 was due to transport from cities such as Linyi, Xuzhou, Suqian, and Lianyungang to the pollution trajectory. The CWT value generally exceeded 80 with the highest value near 200. In contrast, EP2 was affected by the neighboring cities such as Suqian, Suzhou, and Xuzhou, and the CWT value was over 60 with the highest approaching 160, indicating that the interaction among cities in the study area is significant. This study shows that cross-regional air pollution control strategies are particularly important for alleviating the pollution situation in the northern part of the YRD region.

The outbreak of COVID-19 in China has led to massive lockdowns in order to reduce the spread of the epidemic and control human-to-human transmission. Subsequent reductions in various anthropogenic activities have led to improved air quality during the lockdown. In this study, we apply a widely used exposure-response function to estimate the short-term health impacts associated with PM2.5 changes over the Yangtze River Delta (YRD) region due to COVID-19 lockdown. Concentrations of PM2.5 during lockdown period reduced by 22.9% to 54.0% compared to pre-lockdown level. Estimated PM2.5-related daily premature mortality during lockdown period is 895 (95% confidential interval: 637-1,081), which is 43.3% lower than pre-lockdown period and 46.5% lower compared with averages of 2017-2019. According to our calculation, total number of avoided premature death aassociated with PM2.5 reduction during the lockdown is estimated to be 42.4 thousand over the YRD region, with Shanghai, Wenzhou, Suzhou (Jiangsu province), Nanjing, and Nantong being the top five cities with largest health benefits. Avoided premature mortality is mostly contributed by reduced death associated with stroke (16.9 thousand, accounting for 40.0%), ischemic heart disease (14.0 thousand, 33.2%), and chronic obstructive pulmonary disease (7.6 thousand, 18.0%). Our calculations do not support or advocate any idea that pandemics produce a positive note to community health. We simply present health benefits from air pollution improvement due to large emission reductions from lowered human and industrial activities. Our results show that continuous efforts to improve air quality are essential to protect public health, especially over city-clusters with dense population.