Ozone trends and their relationship to characteristic weather patterns

Warm season ambient ozone and children's health in the USA

BACKGROUND

Over 120 million people in the USA live in areas with unsafe ozone (O3) levels. Studies among adults have linked exposure to worse lung function and higher risk of asthma and chronic obstructive pulmonary disease (COPD). However, few studies have examined the effects of O3 in children, and existing studies are limited in terms of their geographic scope or outcomes considered.

METHODS

We leveraged a dataset of encounters at 42 US children's hospitals from 2004-2015. We used a one-stage case-crossover design to quantify the association between daily maximum 8-hour O3 in the county in which the hospital is located and risk of emergency department (ED) visits for any cause and for respiratory disorders, asthma, respiratory infections, allergies and ear disorders.

RESULTS

Approximately 28 million visits were available during this period. Per 10 ppb increase, warm-season (May through September) O3 levels over the past three days were associated with higher risk of ED visits for all causes (risk ratio [RR]: 0.3% [95% confidence interval (CI): 0.2%, 0.4%]), allergies (4.1% [2.5%, 5.7%]), ear disorders (0.8% [0.3%, 1.3%]) and asthma (1.3% [0.8%, 1.9%]). When restricting to levels below the current regulatory standard (70 ppb), O3 was still associated with risk of ED visits for all-cause, allergies, ear disorders and asthma. Stratified analyses suggest that the risk of O3-related all-cause ED visits may be higher in older children.

CONCLUSIONS

Results from this national study extend prior research on the impacts of daily O3 on children's health and reinforce the presence of important adverse health impacts even at levels below the current regulatory standard in the USA.

Observational evidence for detrimental impact of inhaled ozone on human respiratory system

The detrimental influence of inhaled ozone on human respiratory system is ambiguous due to the complexity of dose response relationship between ozone and human respiratory system. This study collects inhaled ozone concentration and respiratory disease data from Shenzhen City to reveal the impact of ozone on respiratory diseases using the Generalized Additive Models (GAM) and Convergent Cross Mapping (CCM) method at the 95% confidence level. The result of GAM exhibits a partially significant lag effect on acute respiratory diseases in cumulative mode. Since the traditional correlation analysis is incapable of capturing causality, the CCM method is applied to examine whether the inhaled ozone affects human respiratory system. The results demonstrate that the inhaled ozone has a significant causative impact on hospitalization rates of both upper and lower respiratory diseases. Furthermore, the harmful causative effects of ozone to the human health are varied with gender and age. Females are more susceptible to inhaled ozone than males, probably because of the estrogen levels and the differential regulation of lung immune response. Adults are more sensitive to ozone exposure than children, potentially due to the fact that children need longer time to react to ozone stress than adults, and the elderly are more tolerant than adults and children, which may be related to pulmonary hypofunction of the elderly while has little correlation with ozone exposure.

Air stagnation in China: Spatiotemporal variability and differing impact on PM2.5 and O3 during 2013-2018

In recent years, winter PM_2.5 and summer O₃ pollution which often occurred with air stagnation condition has become a major concern in China. Thus, it is imperative to understand the air stagnation distribution in China and elucidate its impact on air pollution. In this study, three air stagnation indices were calculated according to atmospheric thermal and dynamics parameters using ERA5 data. Two improved indices were more suitable in China, and they displayed similar characteristics: most of the air stagnant days were found in winter, and seasonal distributions showed substantial regional heterogeneity. During stagnation events, flat west or northwest winds at 500 hPa and high pressure at surface dominated, with high relative humidity (RH) and temperature (T), weak winds in most regions. The pollutants concentrations on stagnant days were higher than those on non-stagnant days in most studied areas, with the largest difference of the 90th percentiles of maximum daily 8-h average (MDA8) O₃ up to 62.2 μg m^-3 in Pearl River Delta (PRD) and PM_2.5 up to 95.8 μg m^-3 in North China Plain (NCP). During the evolution of stagnation events, the MDA8 O₃ concentrations showed a significant increase (6.0 μg m^-3 day^-1) in PRD and a slight rise in other regions; the PM_2.5 concentrations and the frequency of extreme PM_2.5 days increased, especially in NCP. Furthermore, O₃ was simultaneously controlled by temperature and stagnation except for Xinjiang (XJ), with the average growth rate of 19.5 μg m^-3 every 3 °C at 19 °C-31 °C. PM_2.5 was dominated by RH and stagnation in northern China while mainly controlled by stagnation in southern China. Notably, the extremes of summer O₃ (winter PM_2.5) pollution was most associated with air stagnation and T at 25 °C-31 °C (air stagnation and RH >50%). The results are expected to provide important reference information for air pollution control in China.

Urban Air Chemistry in Changing Times

Urban air chemistry is characterized by measurements of gas and aerosol composition. These measurements are interpreted from a long history for laboratory and theoretical studies integrating chemical processes with reactant (or emissions) sources, meteorology and air surface interaction. The knowledge of these latter elements and their changes have enabled chemists to quantitatively account for the averages and variability of chemical indicators. To date, the changes are consistent with dominating energy-related emissions for more than 50 years of gas phase photochemistry and associated reactions forming and evolving aerosols. Future changes are expected to continue focusing on energy resources and transportation in most cities. Extreme meteorological conditions combined with urban surface exchange are also likely to become increasingly important factors affecting atmospheric composition, accounting for the past leads to projecting future conditions. The potential evolution of urban air chemistry can be followed with three approaches using observations and chemical transport modeling. The first approach projects future changes using long term indicator data compared with the emission estimates. The second approach applies advanced measurement analysis of the ambient data. Examples include statistical modeling or evaluation derived from chemical mechanisms. The third method, verified with observations, employs a comparison of the deterministic models of chemistry, emission futures, urban meteorology and urban infrastructure changes for future insight.

Real-time indoor PM2.5 monitoring in an urban cohort: Implications for exposure disparities and source control

Fine particulate matter (PM2.5) concentrations are highly variable indoors, with evidence for exposure disparities. Real-time monitoring coupled with novel statistical approaches can better characterize drivers of elevated PM2.5 indoors. We collected real-time PM2.5 data in 71 homes in an urban community of Greater Boston, Massachusetts using Alphasense OPC-N2 monitors. We estimated indoor PM2.5 concentrations of non-ambient origin using mass balance principles, and investigated their associations with indoor source activities at the 0.50 to 0.95 exposure quantiles using mixed effects quantile regressions, overall and by homeownership. On average, the majority of indoor PM2.5 concentrations were of non-ambient origin (≥77%), with a higher proportion at increasing quantiles of the exposure distribution. Major source predictors of non-ambient PM2.5 concentrations at the upper quantile (0.95) were cooking (1.4-23 μg/m3) and smoking (15 μg/m3, only among renters), with concentrations also increasing with range hood use (3.6 μg/m3) and during the heating season (5.6 μg/m3). Across quantiles, renters in multifamily housing experienced a higher proportion of PM2.5 concentrations from non-ambient sources than homeowners in single- and multifamily housing. Renters also more frequently reported cooking, smoking, spray air freshener use, and second-hand smoke exposure, and lived in units with higher air exchange rate and building density. Accounting for these factors explained observed PM2.5 exposure disparities by homeownership, particularly in the upper exposure quantiles. Our results suggest that renters in multifamily housing may experience higher PM2.5 exposures due to a combination of behavioral and building factors that are amenable to intervention.

Site-scale modeling of surface ozone in Northern Bavaria using machine learning algorithms, regional dynamic models, and a hybrid model

Ozone (O₃) is a harmful pollutant when present in the lowermost layer of the atmosphere. Therefore, the European Commission formulated directives to regulate O₃ concentrations in near-surface air. However, almost 50% of the 5068 air quality stations in Europe do not monitor O₃ concentrations. This study aims to provide a hybrid modeling system that fills these gaps in the hourly surface O₃ observations on a site scale with much higher accuracy than existing O₃ models. This hybrid model was developed using estimations from multiple linear regression-based eXtreme Gradient Boosting Machines (MLR-XGBM) and O₃ reanalysis from European regional air quality models (CAMS-EU). The binary classification of extremely high O₃ events and the 1- and 24-h forecasts of hourly O₃ were investigated as secondary aims. In this study thirteen stations in Northern Bavaria, out of which six do not monitor O₃, were chosen as test sites. Considering the computational complexity of machine learning algorithms (MLAs), we also applied two recent MLA interpretation methods, namely SHapley Additive exPlanations (SHAP) and Local interpretable model-agnostic explanations (LIME). With SHAP, we showed an increasing effect of temperature on O₃ concentrations which intensifies for temperatures exceeding 17 °C. According to LIME, O₃ concentration peaks are mainly governed by meteorological factors under dry and warm conditions on a regional scale, whereas local nitrogen oxide concentrations control base O₃ concentrations during cold and wet periods. While recently developed MLAs for the spatial estimation of hourly O₃ concentrations had a station-based root-mean-square error (RMSE) above 27 μg/m³, our proposed model significantly reduced the estimation errors by about 66% with an RMSE of 9.49 μg/m³. We also found that logistic regression (LR) and MLR-XGBM performed best in the site-scale classification and 24-h forecast of O₃ concentrations (with a station-averaged accuracy and RMSE of 0.95 and 19.34 μg/m³, respectively).

An Ensemble Learning Approach for Estimating High Spatiotemporal Resolution of Ground-Level Ozone in the Contiguous United States

In this paper, we integrated multiple types of predictor variables and three types of machine learners (neural network, random forest, and gradient boosting) into a geographically weighted ensemble model to estimate the daily maximum 8 h O₃ with high resolution over both space (at 1 km × 1 km grid cells covering the contiguous United States) and time (daily estimates between 2000 and 2016). We further quantify monthly model uncertainty for our 1 km × 1 km gridded domain. The results demonstrate high overall model performance with an average cross-validated R² (coefficient of determination) against observations of 0.90 and 0.86 for annual averages. Overall, the model performance of the three machine learning algorithms was quite similar. The overall model performance from the ensemble model outperformed those from any single algorithm. The East North Central region of the United States had the highest R², 0.93, and performance was weakest for the western mountainous regions (R² of 0.86) and New England (R² of 0.87). For the cross validation by season, our model had the best performance during summer with an R² of 0.88. This study can be useful for the environmental health community to more accurately estimate the health impacts of O₃ over space and time, especially in health studies at an intra-urban scale.

An Ensemble Learning Approach for Estimating High Spatiotemporal Resolution of Ground-Level Ozone in the Contiguous United States

In this paper, we integrated multiple types of predictor variables and three types of machine learners (neural network, random forest, and gradient boosting) into a geographically weighted ensemble model to estimate the daily maximum 8 h O₃ with high resolution over both space (at 1 km × 1 km grid cells covering the contiguous United States) and time (daily estimates between 2000 and 2016). We further quantify monthly model uncertainty for our 1 km × 1 km gridded domain. The results demonstrate high overall model performance with an average cross-validated R² (coefficient of determination) against observations of 0.90 and 0.86 for annual averages. Overall, the model performance of the three machine learning algorithms was quite similar. The overall model performance from the ensemble model outperformed those from any single algorithm. The East North Central region of the United States had the highest R², 0.93, and performance was weakest for the western mountainous regions (R² of 0.86) and New England (R² of 0.87). For the cross validation by season, our model had the best performance during summer with an R² of 0.88. This study can be useful for the environmental health community to more accurately estimate the health impacts of O₃ over space and time, especially in health studies at an intra-urban scale.

A hybrid model for spatially and temporally resolved ozone exposures in the continental United States

Ground-level ozone is an important atmospheric oxidant, which exhibits considerable spatial and temporal variability in its concentration level. Existing modeling approaches for ground-level ozone include chemical transport models, land-use regression, Kriging, and data fusion of chemical transport models with monitoring data. Each of these methods has both strengths and weaknesses. Combining those complementary approaches could improve model performance. Meanwhile, satellite-based total column ozone, combined with ozone vertical profile, is another potential input. The authors propose a hybrid model that integrates the above variables to achieve spatially and temporally resolved exposure assessments for ground-level ozone. The authors used a neural network for its capacity to model interactions and nonlinearity. Convolutional layers, which use convolution kernels to aggregate nearby information, were added to the neural network to account for spatial and temporal autocorrelation. The authors trained the model with the Air Quality System (AQS) 8-hr daily maximum ozone in the continental United States from 2000 to 2012 and tested it with left out monitoring sites. Cross-validated R² on the left out monitoring sites ranged from 0.74 to 0.80 (mean 0.76) for predictions on 1 km × 1 km grid cells, which indicates good model performance. Model performance remains good even at low ozone concentrations. The prediction results facilitate epidemiological studies to assess the health effect of ozone in the long term and the short term.

IMPLICATIONS

Ozone monitors do not provide full data coverage over the United States, which is an obstacle to assess the health effect of ozone when monitoring data are not available. This paper used a hybrid approach to combine satellite-based ozone measurements, chemical transport model simulations, land-use terms, and other auxiliary variables to obtain spatially and temporally resolved ground-level ozone estimation.