A Random Forest Approach to Estimate Daily Particulate Matter, Nitrogen Dioxide, and Ozone at Fine Spatial Resolution in Sweden

Long-term spatiotemporal variation of benzo[a]pyrene in Japan: Significant decrease in ambient concentrations, human exposure, and health risk

Although Benzo[a]pyrene (BaP) is considered carcinogenic to humans, the health effects of exposure to ambient levels have not been sufficiently investigated. This study estimated the long-term spatiotemporal variation of BaP in Japan over nearly two decades at a fine spatial resolution of 1 km. This study aimed to obtain an accurate spatiotemporal distribution of BaP that can be used in epidemiological studies on the health effects of ambient BaP exposure. The annual BaP concentrations were estimated using an ensemble machine learning approach using various predictors, including the concentrations and emission intensities of the criteria air pollutants, and meteorological, land use, and traffic-related variables. The model performance, evaluated by location-based cross-validation, exhibited satisfactory accuracy (R2 of 0.693). Densely populated areas showed higher BaP levels and greater temporal reduction, whereas BaP levels remained higher in some industrial areas. The population-weighted BaP in 2018 was 0.12 ng m-3, a decrease of approximately 70% from its 2000 value of 0.44 ng m-3, which was also reflected in the estimated excess number of lung cancer incidences. Accordingly, the proportion of BaP exposure below 0.12 ng m-3, which is the BaP concentration associated with an excess lifetime cancer risk of 10-5, reached 67% in 2018. Our estimates can be used in epidemiological studies to assess the health effects of BaP exposure at ambient concentrations.

High resolution mapping of nitrogen dioxide and particulate matter in Great Britain (2003-2021) with multi-stage data reconstruction and ensemble machine learning methods

In this contribution, we applied a multi-stage machine learning (ML) framework to map daily values of nitrogen dioxide (NO2) and particulate matter (PM10 and PM2.5) at a 1 km2 resolution over Great Britain for the period 2003-2021. The process combined ground monitoring observations, satellite-derived products, climate reanalyses and chemical transport model datasets, and traffic and land-use data. Each feature was harmonized to 1 km resolution and extracted at monitoring sites. Models used single and ensemble-based algorithms featuring random forests (RF), extreme gradient boosting (XGB), light gradient boosting machine (LGBM), as well as lasso and ridge regression. The various stages focused on augmenting PM2.5 using co-occurring PM10 values, gap-filling aerosol optical depth and columnar NO2 data obtained from satellite instruments, and finally the training of an ensemble model and the prediction of daily values across the whole geographical domain (2003-2021). Results show a good ensemble model performance, calculated through a ten-fold monitor-based cross-validation procedure, with an average R2 of 0.690 (range 0.611-0.792) for NO2, 0.704 (0.609-0.786) for PM10, and 0.802 (0.746-0.888) for PM2.5. Reconstructed pollution levels decreased markedly within the study period, with a stronger reduction in the latter eight years. The pollutants exhibited different spatial patterns, while NO2 rose in close proximity to high-traffic areas, PM demonstrated variation at a larger scale. The resulting 1 km2 spatially resolved daily datasets allow for linkage with health data across Great Britain over nearly two decades, thus contributing to extensive, extended, and detailed research on the long-and short-term health effects of air pollution.

Association between ambient air pollution a week prior to delivery and preterm birth using a nationwide study in Sweden

BACKGROUND

Air pollution exposure has been linked with increased risk of preterm birth, which is one of the leading causes of infant mortality. Limited studies have attempted to explore these associations in low-polluted areas. In this study, we aimed to assess the association between short-term exposure to ambient air pollution and preterm birth in Sweden.

METHOD

In this population-based study we included preterm births between 2014 and 2019 from the Swedish Pregnancy Register. We applied a spatiotemporal model to estimate daily levels of particulate matter <2.5 μm (PM2.5), PM < 10 μm (PM10), nitrogen dioxide (NO2), and ozone (O3) at the residential address of each participant. We applied a time-stratified case-crossover design with conditional logistic regression analysis to estimate odds ratios (OR) of preterm birth per 10 μg/m3 (PM10, NO2, O3) and 5 μg/m3 (PM2.5) increase in air pollution exposure at 0-6-day lag. Two-pollutant models were applied to evaluate the independent association of each exposure on preterm birth. We also stratified by maternal characteristics to identify potential effect modifiers.

RESULTS

28,216 (4.5%) preterm births were included. An increase in O3 exposure was associated with increased odds of preterm birth [OR = 1.06 per 10 μg/m3 (95% CI, 1.02; 1.10]. PM2.5 and PM10 were not significantly associated with preterm birth, and NO2 displayed a negative nonlinear association with preterm birth. We did not observe any notable effect modification, but we found suggestive larger associations between O3 and preterm birth when stratifying by male sex, spontaneous delivery, and spring season.

CONCLUSIONS

Increased O3 exposure one week before delivery was associated with an increased risk of preterm birth in Sweden, a country with levels of air pollution below the current World Health Organization air quality guidelines. Increases in O3 levels with climate change make these findings especially concerning.

Transformer-based ozone multivariate prediction considering interpretable and priori knowledge: A case study of Beijing, China

Ozone pollution is the focus of current environmental governance in China and high-quality prediction of ozone concentration is the prerequisite to effective policymaking. The studied ozone pollution time series exhibits distinct seasonality and secular trends and is associated with various factors. This study developed an interpretable hybrid model by combining STL decomposition and the Transformer (STL-Transformer) with the prior information of ozone time series and global multi-source information as prediction basis. The STL decomposition decomposes ozone time series into trend, seasonal, and remainder components. Then, the three components, along with other air quality and meteorological data, are integrated into the input sequence of the Transformer. The experiment results show that the STL-Transformer outperforms the other five state-of-the-art models, including the standard Transformer. Specially, the univariate forecasting for ozone relies on mimicking the patterns and trends that have occurred in the past. In contrast, multivariate forecasting can effectively capture complex relationships and dependencies involving multiple variables. The method successfully grasps the prior and global multi-source information and simultaneously improves the interpretability of ozone prediction with high precision. This study provides new insights for air pollution forecasting and has reliable theoretical value and practical significance for environmental governance.

Predicting particulate matter, nitrogen dioxide, and ozone across Great Britain with high spatiotemporal resolution based on random forest models

In Great Britain, limited studies have employed machine learning methods to predict air pollution especially ozone (O3) with high spatiotemporal resolution. This study aimed to address this gap by developing random forest models for four key pollutants (fine and inhalable particulate matter [PM2.5 and PM10], nitrogen dioxide [NO2] and O3) by integrating multiple-source predictors at a daily level and 1-km resolution. The out-of-bag R2 (root mean squared error, RMSE) between predictions from models and measurements from monitoring stations in 2006-2013 was 0.85 (3.63 μg/m3) for PM2.5, 0.77 (6.00 μg/m3) for PM10, 0.85 (9.71 μg/m3) for NO2, and 0.85 (9.39 μg/m3) for maximum daily 8-h average (MDA8) O3 at daily level, and the predicting accuracy was higher at monthly and annual level. The high-resolution predictions captured characterized spatiotemporal patterns of the four pollutants. Higher concentrations of PM2.5, PM10, and NO2 were distributed in densely populated southern regions of Great Britain while O3 showed an inverse spatial pattern in general, which could not be fully depicted by monitoring stations. Therefore, predictions produced in this study could improve exposure assessment with less exposure misclassification and flexible exposure windows for future epidemiological studies to investigate the impact of air pollution across Great Britain.

Is model-estimated PM2.5 exposure equivalent to station-observed in mortality risk assessment? A literature review and meta-analysis

Model-estimated air pollution exposure assessments have been extensively employed in the evaluation of health risks associated with air pollution. However, few studies synthetically evaluate the reliability of model-estimated PM2.5 products in health risk assessment by comparing them with ground-based monitoring station air quality data. In response to this gap, we undertook a meticulously structured systematic review and meta-analysis. Our objective was to aggregate existing comparative studies to ascertain the disparity in mortality effect estimates derived from model-estimated ambient PM2.5 exposure versus those based on monitoring station-observed PM2.5 exposure. We conducted searches across multiple databases, namely PubMed, Scopus, and Web of Science, using predefined keywords. Ultimately, ten studies were included in the review. Of these, seven investigated long-term annual exposure, while the remaining three studies focused on short-term daily PM2.5 exposure. Despite variances in the estimated Exposure-Response (E-R) associations, most studies revealed positive associations between ambient PM2.5 exposure and all-cause and cardiovascular mortality, irrespective of the exposure being estimated through models or observed at monitoring stations. Our meta-analysis revealed that all-cause mortality risk associated with model-estimated PM2.5 exposure was in line with that derived from station-observed sources. The pooled Relative Risk (RR) was 1.083 (95% CI: 1.047, 1.119) for model-estimated exposure, and 1.089 (95% CI: 1.054, 1.125) for station-observed sources (p = 0.795). In conclusion, most model-estimated air pollution products have demonstrated consistency in estimating mortality risk compared to data from monitoring stations. However, only a limited number of studies have undertaken such comparative analyses, underscoring the necessity for more comprehensive investigations to validate the reliability of these model-estimated exposure in mortality risk assessment.

Relation of prenatal and postnatal PM2.5 exposure with cognitive and motor function among preschool-aged children

The literature informing susceptible periods of exposure on children's neurodevelopment is limited. We evaluated the impacts of pre- and postnatal fine particulate matter (PM2.5) exposure on children's cognitive and motor function among 1303 mother-child pairs in the Spanish INMA (Environment and Childhood) Study. Random forest models with temporal back extrapolation were used to estimate daily residential PM2.5 exposures that we averaged across 1-week lags during the prenatal period and 4-week lags during the postnatal period. The McCarthy Scales of Children's Abilities (MSCA) were administered around 5 years to assess general cognitive index (GCI) and several subscales (verbal, perceptual-performance, memory, fine motor, gross motor). We applied distributed lag nonlinear models within the Bayesian hierarchical framework to explore periods of susceptibility to PM2.5 on each MSCA outcome. Effect estimates were calculated per 5 μg/m3 increase in PM2.5 and aggregated across adjacent statistically significant lags using cumulative β (βcum) and 95% Credible Intervals (95%CrI). We evaluated interactions between PM2.5 with fetal growth and child sex. We did not observe associations of PM2.5 exposure with lower GCI scores. We found a period of susceptibility to PM2.5 on fine motor scores in gestational weeks 1-9 (βcum = -2.55, 95%CrI = -3.53,-1.56) and on gross motor scores in weeks 7-17 (βcum = -2.27,95%CrI = -3.43,-1.11) though the individual lags for the latter were only borderline statistically significant. Exposure in gestational week 17 was weakly associated with verbal scores (βcum = -0.17, 95%CrI = -0.26,-0.09). In the postnatal period (from age 0.5-1.2 years), we observed a window of susceptibility to PM2.5 on lower perceptual-performance (β = -2.42, 95%CrI = -3.37,-1.46). Unexpected protective associations were observed for several outcomes with exposures in the later postnatal period. We observed no evidence of differences in susceptible periods by fetal growth or child sex. Preschool-aged children's motor function may be particularly susceptible to PM2.5 exposures experienced in utero whereas the first year of life was identified as a period of susceptibility to PM2.5 for children's perceptual-performance.

Estimating morning and evening commute period O3 concentration in Taiwan using a fine spatial-temporal resolution ensemble mixed spatial model with Geo-AI technology

Elevated levels of ground-level ozone (O3) can have harmful effects on health. While previous studies have focused mainly on daily averages and daytime patterns, it's crucial to consider the effects of air pollution during daily commutes, as this can significantly contribute to overall exposure. This study is also the first to employ an ensemble mixed spatial model (EMSM) that integrates multiple machine learning algorithms and predictor variables selected using Shapley Additive exExplanations (SHAP) values to predict spatial-temporal fluctuations in O3 concentrations across the entire island of Taiwan. We utilized geospatial-artificial intelligence (Geo-AI), incorporating kriging, land use regression (LUR), machine learning (random forest (RF), categorical boosting (CatBoost), gradient boosting (GBM), extreme gradient boosting (XGBoost), and light gradient boosting (LightGBM)), and ensemble learning techniques to develop ensemble mixed spatial models (EMSMs) for morning and evening commute periods. The EMSMs were used to estimate long-term spatiotemporal variations of O3 levels, accounting for in-situ measurements, meteorological factors, geospatial predictors, and social and seasonal influences over a 26-year period. Compared to conventional LUR-based approaches, the EMSMs improved performance by 58% for both commute periods, with high explanatory power and an adjusted R2 of 0.91. Internal and external validation procedures and verification of O3 concentrations at the upper percentile ranges (in 1%, 5%, 10%, 15%, 20%, and 25%) and other conditions (including rain, no rain, weekday, weekend, festival, and no festival) have demonstrated that the models are stable and free from overfitting issues. Estimation maps were generated to examine changes in O3 levels before and during the implementation of COVID-19 restrictions. These findings provide accurate variations of O3 levels in commute period with high spatiotemporal resolution of daily and 50m * 50m grid, which can support control pollution efforts and aid in epidemiological studies.

A comparative analysis of linear regression, neural networks and random forest regression for predicting air ozone employing soft sensor models

The proposed methodology presents a comprehensive analysis of soft sensor modeling techniques for air ozone prediction. We compare the performance of three different modeling techniques: LR (linear regression), NN (neural networks), and RFR (random forest regression). Additionally, we evaluate the impact of different variable sets on prediction performance. Our findings indicate that neural network models, particularly the RNN (recurrent neural networks), outperform the other modeling techniques in terms of prediction accuracy. The proposed methodology evaluates the impact of different variable sets on prediction performance, finding that variable set E demonstrates exceptional performance and achieves the highest average prediction accuracy among various software sensor models. In comparing variable set E and A, B, C, D, it is observed that the inclusion of an additional input feature, PM10, in the latter sets does not improve overall performance, potentially due to multicollinearity between PM10 and PM2.5 variables. The proposed methodology provides valuable insights into soft sensor modeling for air ozone prediction.Among the 72 sensors, sensor NNR[Y]C outperforms all other evaluated sensors, demonstrating exceptional predictive performance with an impressive R2 of 0.8902, low RMSE of 24.91, and remarkable MAE of 19.16. With a prediction accuracy of 81.44%, sensor NNR[Y]C is reliable and suitable for various technological applications.

Short-Term Ambient Air Pollution Exposure and Risk of Out-of-Hospital Cardiac Arrest in Sweden: A Nationwide Case-Crossover Study

Background Air pollution is one of the main risk factors for cardiovascular disease globally, but its association with out-of-hospital cardiac arrest at low air pollution levels is unclear. This nationwide study in Sweden aims to investigate if air pollution is associated with a higher risk of out-of-hospital cardiac arrest in an area with relatively low air pollution levels. Methods and Results This study was a nationwide time-stratified case-crossover study investigating the association between short-term air pollution exposures and out-of-hospital cardiac arrest using data from the SRCR (Swedish Registry for Cardiopulmonary Resuscitation) between 2009 and 2019. Daily air pollution levels were estimated in 1×1-km grids for all of Sweden using a satellite-based machine learning model. The association between daily air pollutant levels and out-of-hospital cardiac arrest was quantified using conditional logistic regression adjusted for daily air temperature. Particulate matter <2.5 μm exposure was associated with a higher risk of out-of-hospital cardiac arrest among a total of 29 604 cases. In a multipollutant model, the association was most pronounced for intermediate daily lags, with an increased relative risk of 6.2% (95% CI, 1.0-11.8) per 10 μg/m3 increase of particulate matter <2.5 μm 4 days before the event. A similar pattern of association was observed for particulate matter <10 μm. No clear association was observed for O3 and NO2. Conclusions Short-term exposure to air pollution was associated with higher risk of out-of-hospital cardiac arrest. The findings add to the evidence of an adverse effect of particulate matter on out-of-hospital cardiac arrest, even at very low levels below current regulatory standards.

Analysis of severe ozone-related human health and weather influence over China in 2019 based on a high-resolution dataset

Ozone pollution in 2019 in China is particularly severe posing a tremendous threat to the health of Chinese inhabitants. In this study, we constructed a more reliable and accurate 1-km gridded dataset for 2019 with as many sites as possible using the inverse distance weight interpolation method to analyze spatiotemporal ozone pollution characteristics and health burden attributed to ozone exposure from the perspective of different diseases and weather influence. The accuracy of this new dataset is higher than other public datasets, with the coefficient of determination of 0.84 and root-mean-square error of 8.77 ppb through the validation of 300 external sites which were never used for establishing retrieval methods by the datasets mentioned-above. The averaged MDA8 (the daily maximum 8 h average) ozone concentrations over China was 43.5 ppb, and during April-July, 83.9% of total grids occurred peak-month ozone concentrations. Overall, the highest averaged exceedance days (60 days) and population-weighted ozone concentrations (55.0 ppb) both concentrated in central-eastern China including 9 provinces (only 11.4% of the national territory); meanwhile, all-cause premature deaths attributable to ozone exposure reached up to 142,000 (54.9% of national total deaths) with higher deaths for cardiovascular and respiratory, and the provincial per capita premature mortality was 0.27~0.44‰. The six most polluted weather types in the central-eastern China are in order as follows: westerly (SW and W), cyclonic, northerly, and southerly (NW, N, and S) types, which accounts for approximately 73.2% of health burden attributed to daily ozone exposure and poses the greatest public health risk with mean daily premature deaths ranging from 466 to 610. Our findings could provide an effective support for regional ozone pollution control and public health management in China.

Community factors and excess mortality in the COVID-19 pandemic in England, Italy and Sweden

BACKGROUND

Analyses of coronavirus disease 19 suggest specific risk factors make communities more or less vulnerable to pandemic-related deaths within countries. What is unclear is whether the characteristics affecting vulnerability of small communities within countries produce similar patterns of excess mortality across countries with different demographics and public health responses to the pandemic. Our aim is to quantify community-level variations in excess mortality within England, Italy and Sweden and identify how such spatial variability was driven by community-level characteristics.

METHODS

We applied a two-stage Bayesian model to quantify inequalities in excess mortality in people aged 40 years and older at the community level in England, Italy and Sweden during the first year of the pandemic (March 2020-February 2021). We used community characteristics measuring deprivation, air pollution, living conditions, population density and movement of people as covariates to quantify their associations with excess mortality.

RESULTS

We found just under half of communities in England (48.1%) and Italy (45.8%) had an excess mortality of over 300 per 100 000 males over the age of 40, while for Sweden that covered 23.1% of communities. We showed that deprivation is a strong predictor of excess mortality across the three countries, and communities with high levels of overcrowding were associated with higher excess mortality in England and Sweden.

CONCLUSION

These results highlight some international similarities in factors affecting mortality that will help policy makers target public health measures to increase resilience to the mortality impacts of this and future pandemics.

Machine Learning Explains Long-Term Trend and Health Risk of Air Pollution during 2015-2022 in a Coastal City in Eastern China

Exposure to air pollution is one of the greatest environmental risks for human health. Air pollution level is significantly driven by anthropogenic emissions and meteorological conditions. To protect people from air pollutants, China has implemented clean air actions to reduce anthropogenic emissions, which has led to rapid improvement in air quality over China. Here, we evaluated the impact of anthropogenic emissions and meteorological conditions on trends in air pollutants in a coastal city (Lianyungang) in eastern China from 2015 to 2022 based on a random forest model. The annual mean concentration of observed air pollutants, including fine particles, inhalable particles, sulfur dioxide, nitrogen dioxide, and carbon monoxide, presented significant decreasing trends during 2015-2022, with dominant contributions (55-75%) by anthropogenic emission reduction. An increasing trend in ozone was observed with an important contribution (28%) by anthropogenic emissions. The impact of meteorological conditions on air pollution showed significant seasonality. For instance, the negative impact on aerosol pollution occurred during cold months, while the positive impact was in warm months. Health-risk-based air quality decreased by approximately 40% in 8 years, for which anthropogenic emission made a major contribution (93%).

Susceptible windows of exposure to fine particulate matter and fetal growth trajectories in the Spanish INMA (INfancia y Medio Ambiente) birth cohort

While prior studies report associations between fine particulate matter (PM2.5) exposure and fetal growth, few have explored temporally refined susceptible windows of exposure. We included 2328 women from the Spanish INMA Project from 2003 to 2008. Longitudinal growth curves were constructed for each fetus using ultrasounds from 12, 20, and 34 gestational weeks. Z-scores representing growth trajectories of biparietal diameter, femur length, abdominal circumference (AC), and estimated fetal weight (EFW) during early (0-12 weeks), mid- (12-20 weeks), and late (20-34 weeks) pregnancy were calculated. A spatio-temporal random forest model with back-extrapolation provided weekly PM2.5 exposure estimates for each woman during her pregnancy. Distributed lag non-linear models were implemented within the Bayesian hierarchical framework to identify susceptible windows of exposure for each outcome and cumulative effects [βcum, 95% credible interval (CrI)] were aggregated across adjacent weeks. For comparison, general linear models evaluated associations between PM2.5 averaged across multi-week periods (i.e., weeks 1-11, 12-19, and 20-33) and fetal growth, mutually adjusted for exposure during each period. Results are presented as %change in z-scores per 5 μg/m3 in PM2.5, adjusted for covariates. Weeks 1-6 [βcum = -0.77%, 95%CrI (-1.07%, -0.47%)] were identified as a susceptible window of exposure for reduced late pregnancy EFW while weeks 29-33 were positively associated with this outcome [βcum = 0.42%, 95%CrI (0.20%, 0.64%)]. A similar pattern was observed for AC in late pregnancy. In linear regression models, PM2.5 exposure averaged across weeks 1-11 was associated with reduced late pregnancy EFW and AC; but, positive associations between PM2.5 and EFW or AC trajectories in late pregnancy were not observed. PM2.5 exposures during specific weeks may affect fetal growth differentially across pregnancy and such associations may be missed by averaging exposure across multi-week periods, highlighting the importance of temporally refined exposure estimates when studying the associations of air pollution with fetal growth.

Ensemble averaging using remote sensing data to model spatiotemporal PM10 concentrations in sparsely monitored South Africa

There is a paucity of air quality data in sub-Saharan African countries to inform science driven air quality management and epidemiological studies. We investigated the use of available remote-sensing aerosol optical depth (AOD) data to develop spatially and temporally resolved models to predict daily particulate matter (PM₁₀) concentrations across four provinces of South Africa (Gauteng, Mpumalanga, KwaZulu-Natal and Western Cape) for the year 2016 in a two-staged approach. In stage 1, a Random Forest (RF) model was used to impute Multiangle Implementation of Atmospheric Correction AOD data for days where it was missing. In stage 2, the machine learner algorithms RF, Gradient Boosting and Support Vector Regression were used to model the relationship between ground-monitored PM₁₀ data, AOD and other spatial and temporal predictors. These were subsequently combined in an ensemble model to predict daily PM₁₀ concentrations at 1 km × 1 km spatial resolution across the four provinces. An out-of-bag R² of 0.96 was achieved for the first stage model. The stage 2 cross-validated (CV) ensemble model captured 0.84 variability in ground-monitored PM₁₀ with a spatial CV R² of 0.48 and temporal CV R² of 0.80. The stage 2 model indicated an optimal performance of the daily predictions when aggregated to monthly and annual means. Our results suggest that a combination of remote sensing data, chemical transport model estimates and other spatiotemporal predictors has the potential to improve air quality exposure data in South Africa's major industrial provinces. In particular, the use of a combined ensemble approach was found to be useful for this area with limited availability of air pollution ground monitoring data.

Optimized neural network for daily-scale ozone prediction based on transfer learning

Tropospheric ozone (O₃) pollution is worsening in China, and an accurate forecast is a prerequisite to lower the O₃ peak level. In recent years, machine learning techniques have attracted increasing attention in O₃ prediction owing to their high efficiency and simple operation. However, the accuracy of predicting the daily O₃ level is low. This study proposed a novel model by coupling long short-term memory neural network with transfer learning (TL-LSTM), with meteorology and pollutant concentration information as the model input. L2 regularization was applied to reduce the risk of overfitting and to improve the accuracy and generalization ability of the model prediction. Our results indicated that by transferring the knowledge in the model configuration from the hourly LSTM module, TL-LSTM greatly improves the predictability of the daily maximum 8 h average (MDA8) of O₃ in Hong Kong. The coefficient of determination (R²) increased from 0.684 to 0.783 and the mean square error (MSE) reduced from 1.36 × 10^-2 to 1.05 × 10^-2. Furthermore, R² and MSE were the highest in summer, indicating an under-prediction of peak O₃ levels. This was a result of the limited number of high O₃ days, which did not provide sufficient knowledge for the model to make an accurate prediction. Sobol analysis indicated that wind speed was the most sensitive factor in O₃ prediction, largely due to the development of land-sea breeze circulation which effectively traps pollutants and expedites O₃ formation. The results clearly demonstrate the effectiveness of the TL-LSTM in predicting the daily O₃ concentration in Hong Kong. Thus, TL-LSTM can be promulgated into other photochemically active regions to assist in O₃ pollution forecasting and management.

Development and Evaluation of Statistical Models Based on Machine Learning Techniques for Estimating Particulate Matter (PM2.5 and PM10) Concentrations

Despite extensive research on air pollution estimation/prediction, inter-country models for estimating air pollutant concentrations in Southeast Asia have not yet been fully developed and validated owing to the lack of air quality (AQ), emission inventory and meteorological data from different countries in the region. The purpose of this study is to develop and evaluate two machine learning (ML)-based models (i.e., analysis of covariance (ANCOVA) and random forest regression (RFR)) for estimating daily PM_2.5 and PM₁₀ concentrations in Brunei Darussalam. These models were first derived from past AQ and meteorological measurements in Singapore and then tested with AQ and meteorological data from Brunei Darussalam. The results show that the ANCOVA model (R2 = 0.94 and RMSE = 0.05 µg/m³ for PM_2.5, and R2 = 0.72 and RMSE = 0.09 µg/m³ for PM₁₀) could describe daily PM concentrations over 18 µg/m³ in Brunei Darussalam much better than the RFR model (R2 = 0.92 and RMSE = 0.04 µg/m³ for PM_2.5, and R2 = 0.86 and RMSE = 0.08 µg/m³ for PM₁₀). In conclusion, the derived models provide a satisfactory estimation of PM concentrations for both countries despite some limitations. This study shows the potential of the models for inter-country PM estimations in Southeast Asia.

Random Forests Assessment of the Role of Atmospheric Circulation in PM10 in an Urban Area with Complex Topography

This study presents the assessment of the quantitative influence of atmospheric circulation on the pollutant concentration in the area of Kraków, Southern Poland, for the period 2000–2020. The research has been realized with the application of different statistical parameters, synoptic meteorology tools, the Random Forests machine learning method, and multilinear regression analyses. Another aim of the research was to evaluate the types of atmospheric circulation classification methods used in studies on air pollution dispersion and to assess the possibility of their application in air quality management, including short-term PM10 daily forecasts. During the period analyzed, a significant decreasing trend of pollutants’ concentrations and varying atmospheric circulation conditions was observed. To understand the relation between PM10 concentration and meteorological conditions and their significance, the Random Forests algorithm was applied. Observations from meteorological stations, air quality measurements and ERA-5 reanalysis were used. The meteorological database was used as an input to models that were trained to predict daily PM10 concentration and its day-to-day changes. This study made it possible to distinguish the dominant circulation types with the highest probability of occurrence of poor air quality or a significant improvement in air quality conditions. Apart from the parameters whose significant influence on air quality is well established (air temperature and wind speed at the ground and air temperature gradient), the key factor was also the gradient of relative air humidity and wind shear in the lowest troposphere. Partial dependence calculated with the use of the Random Forests model made it possible to better analyze the impact of individual meteorological parameters on the PM10 daily concentration. The analysis has shown that, for areas with a diversified topography, it is crucial to use the variability of the atmospheric circulation during the day to better forecast air quality.

Data-Driven Machine Learning in Environmental Pollution: Gains and Problems

The complexity and dynamics of the environment make it extremely difficult to directly predict and trace the temporal and spatial changes in pollution. In the past decade, the unprecedented accumulation of data, the development of high-performance computing power, and the rise of diverse machine learning (ML) methods provide new opportunities for environmental pollution research. The ML methodology has been used in satellite data processing to obtain ground-level concentrations of atmospheric pollutants, pollution source apportionment, and spatial distribution modeling of water pollutants. However, unlike the active practices of ML in chemical toxicity prediction, advanced algorithms such as deep neural networks in environmental process studies of pollutants are still deficient. In addition, over 40% of the environmental applications of ML go to air pollution, and its application range and acceptance in other aspects of environmental science remain to be increased. The use of ML methods to revolutionize environmental science and its problem-solving scenarios has its own challenges. Several issues should be taken into consideration, such as the tradeoff between model performance and interpretability, prerequisites of the machine learning model, model selection, and data sharing.

Source profiling of air pollution and its association with acute respiratory infections in the Himalayan-bound region of India

The studies related to air pollutants and their association with human health over the mountainous region are of utmost importance and are sparse especially over the Himalayan region of India. The linkages between various atmospheric variables and clinically validated data have been done using various datasets procured from satellite, model reanalysis, and surface observations during 2013-2017. Aerosol optical depth, air temperature, and wind speed are significantly related (p < 0.001) to the incidence of acute respiratory infections with its peak during winter. Model-derived particulate matter (PM_2.5) shows high contributions of black carbon, organic carbon, and sulfate during winter. The wind roses show the passage of winds from the south-west and southern side of the region. Back trajectory density plot along with bivariate polar plot analyses have shown that most of the winds coming from the western side are taking a southward direction before reaching the study area and may be bringing pollutants from the Indo-Gangetic Plain and other surrounding regions. Our study shows that the accumulation of pollutants in the Himalayan valley is owing to the meteorological stability with significant local emissions from burning of biomass and biofuels along with long-range and mid-range transport during the winter season that significantly correlated with the incidence of acute respiratory infections in the region.

A robust approach to deriving long-term daily surface NO2 levels across China: Correction to substantial estimation bias in back-extrapolation

BACKGROUND

Long-term surface NO₂ data are essential for retrospective policy evaluation and chronic human exposure assessment. In the absence of NO₂ observations for Mainland China before 2013, training a model with 2013-2018 data to make predictions for 2005-2012 (back-extrapolation) could cause substantial estimation bias due to concept drift.

OBJECTIVE

This study aims to correct the estimation bias in order to reconstruct the spatiotemporal distribution of daily surface NO₂ levels across China during 2005-2018.

METHODS

On the basis of ground- and satellite-based data, we proposed the robust back-extrapolation with a random forest (RBE-RF) to simulate the surface NO₂ through intermediate modeling of the scaling factors. For comparison purposes, we also employed a random forest (Base-RF), as a representative of the commonly used approach, to directly model the surface NO₂ levels.

RESULTS

The validation against Taiwan's NO₂ observations during 2005-2012 showed that RBE-RF adequately corrected the substantial underestimation by Base-RF. The RMSE decreased from 10.1 to 8.2 µg/m³, 7.1 to 4.3 µg/m³, and 6.1 to 2.9 µg/m³ in predicting daily, monthly, and annual levels, respectively. For North China with the most severe pollution, the population-weighted NO₂ ([NO₂]_pw) during 2005-2012 was estimated as 40.2 and 50.9 µg/m³ by Base-RF and RBE-RF, respectively, i.e., 21.0% difference. While both models predicted that the national annual [NO₂]_pw increased during 2005-2011 and then decreased, the interannual trends were underestimated by >50.2% by Base-RF relative to RBE-RF. During 2005-2018, the nationwide population that lived in the areas with NO₂ > 40 µg/m³ were estimated as 259 and 460 million by Base-RF and RBE-RF, respectively.

CONCLUSION

With RBE-RF, we corrected the estimation bias in back-extrapolation and obtained a full-coverage dataset of daily surface NO₂ across China during 2005-2018, which is valuable for environmental management and epidemiological research.

A New Random Forest Algorithm Based on Learning Automata

The goal of aggregating the base classifiers is to achieve an aggregated classifier that has a higher resolution than individual classifiers. Random forest is one of the types of ensemble learning methods that have been considered more than other ensemble learning methods due to its simple structure, ease of understanding, as well as higher efficiency than similar methods. The ability and efficiency of classical methods are always influenced by the data. The capabilities of independence from the data domain, and the ability to adapt to problem space conditions, are the most challenging issues about the different types of classifiers. In this paper, a method based on learning automata is presented, through which the adaptive capabilities of the problem space, as well as the independence of the data domain, are added to the random forest to increase its efficiency. Using the idea of reinforcement learning in the random forest has made it possible to address issues with data that have a dynamic behaviour. Dynamic behaviour refers to the variability in the behaviour of a data sample in different domains. Therefore, to evaluate the proposed method, and to create an environment with dynamic behaviour, different domains of data have been considered. In the proposed method, the idea is added to the random forest using learning automata. The reason for this choice is the simple structure of the learning automata and the compatibility of the learning automata with the problem space. The evaluation results confirm the improvement of random forest efficiency.

Comparing Methods to Impute Missing Daily Ground-Level PM10 Concentrations between 2010-2017 in South Africa

Good quality and completeness of ambient air quality monitoring data is central in supporting actions towards mitigating the impact of ambient air pollution. In South Africa, however, availability of continuous ground-level air pollution monitoring data is scarce and incomplete. To address this issue, we developed and compared different modeling approaches to impute missing daily average particulate matter (PM₁₀) data between 2010 and 2017 using spatiotemporal predictor variables. The random forest (RF) machine learning method was used to explore the relationship between average daily PM₁₀ concentrations and spatiotemporal predictors like meteorological, land use and source-related variables. National (8 models), provincial (32) and site-specific (44) RF models were developed to impute missing daily PM₁₀ data. The annual national, provincial and site-specific RF cross-validation (CV) models explained on average 78%, 70% and 55% of ground-level PM₁₀ concentrations, respectively. The spatial components of the national and provincial CV RF models explained on average 22% and 48%, while the temporal components of the national, provincial and site-specific CV RF models explained on average 78%, 68% and 57% of ground-level PM₁₀ concentrations, respectively. This study demonstrates a feasible approach based on RF to impute missing measurement data in areas where data collection is sparse and incomplete.

Application of Bayesian Additive Regression Trees for Estimating Daily Concentrations of PM2.5 Components

Bayesian additive regression tree (BART) is a recent statistical method that combines ensemble learning and nonparametric regression. BART is constructed under a probabilistic framework that also allows for model-based prediction uncertainty quantification. We evaluated the application of BART in predicting daily concentrations of four fine particulate matter (PM_2.5) components (elemental carbon, organic carbon, nitrate, and sulfate) in California during the period 2005 to 2014. We demonstrate in this paper how BART can be tuned to optimize prediction performance and how to evaluate variable importance. Our BART models included, as predictors, a large suite of land-use variables, meteorological conditions, satellite-derived aerosol optical depth parameters, and simulations from a chemical transport model. In cross-validation experiments, BART demonstrated good out-of-sample prediction performance at monitoring locations (R² from 0.62 to 0.73). More importantly, prediction intervals associated with concentration estimates from BART showed good coverage probability at locations with and without monitoring data. In our case study, major PM_2.5 components could be estimated with good accuracy, especially when collocated PM_2.5 total mass observations were available. In conclusion, BART is an attractive approach for modeling ambient air pollution levels, especially for its ability to provide uncertainty in estimates that may be useful for subsequent health impact and health effect analyses.

Short-term effects of air pollutants on daily mortality in the Stockholm county - A spatiotemporal analysis

Short-term exposure to air pollutants has been extensively related to daily mortality, however most of the evidence comes from studies conducted in major cities, and little is known on the extent of the spatial heterogeneity in the effects within areas including both urban and non-urban settings. We aimed to investigate the short-term association of air pollutants with daily cause-specific mortality in the Stockholm county, and to test whether an association exists also outside the metropolitan area. We used a spatiotemporal random forest model to predict daily concentrations of fine and inhalable particulate matter (PM_2.5 and PM₁₀), nitrogen dioxide (NO₂) and ozone (O₃) at 1-km spatial resolution over Sweden for 2005-2016. We collected data on daily mortality for each small area for market statistics (SAMS) of the Stockholm county, to which we matched daily exposures to air pollutants and air temperature. We applied a case-crossover design to investigate the short-term association between the four pollutants and mortality from non-accidental, cardiovascular and respiratory causes. We compared the associations in and out the Stockholm urban area, by SAMS population density and across the 26 municipalities of the county. We found weak effects of most air pollutants on cause-specific mortality in the full year analysis, with estimates much larger and significant only during the warmer months (April to September): non-accidental mortality increased by 4.58% (95% confidence interval - 95% CI: 0.89%, 8.41%) and by 2.21% (95% CI: 0.71%, 3.73%) per 10 μg/m³ increase in lag 0-1 PM_2.5 and O₃, respectively. Associations were in general higher in the Stockholm city and in SAMS with high population density. When comparing the 26 municipalities, we didn't detect a significant heterogeneity in the short-term associations with air pollutants. In conclusion, we found a suggestion of a harmful role of air pollution also in non-urban areas, but the study was underpowered to draw firm conclusions. We consider this study as a pilot to investigate the spatial heterogeneity of the association between daily air pollution and mortality at the national level in Sweden.