Development of land use regression models for nitrogen dioxide, ultrafine particles, lung deposited surface area, and four other markers of particulate matter pollution in the Swiss SAPALDIA regions

Urban green spaces, self-rated air pollution and health: A sensitivity analysis of green space characteristics and proximity in four European cities

Exploring the influence of green space characteristics and proximity on health via air pollution mitigation, our study analysed data from 1,365 participants across Porto, Nantes, Sofia, and Høje-Taastrup. Utilizing OpenStreetMap and the AID-PRIGSHARE tool, we generated nine green space indicators around residential addresses at 15 distances, ranging from 100m to 1500m. We performed a mediation analysis for these 135 green space variables and revealed significant associations between self-rated air pollution and self-rated health for specific green space characteristics. In our study, indirect positive effects on health via air pollution were mainly associated with green corridors in intermediate Euclidean distances (800-1,000m) and the amount of accessible green spaces in larger network distances (1,400-1,500m). Our results suggest that the amount of connected green spaces measured in intermediate surroundings seems to be a prime green space characteristic that could drive the air pollution mitigation pathway to health.

A causal relationship between particulate matter 2.5 and obesity and its related indicators: a Mendelian randomization study of European ancestry

Background

In recent years, the prevalence of obesity has continued to increase as a global health concern. Numerous epidemiological studies have confirmed the long-term effects of exposure to ambient air pollutant particulate matter 2.5 (PM2.5) on obesity, but their relationship remains ambiguous.

Methods

Utilizing large-scale publicly available genome-wide association studies (GWAS), we conducted univariate and multivariate Mendelian randomization (MR) analyses to assess the causal effect of PM2.5 exposure on obesity and its related indicators. The primary outcome given for both univariate MR (UVMR) and multivariate MR (MVMR) is the estimation utilizing the inverse variance weighted (IVW) method. The weighted median, MR-Egger, and maximum likelihood techniques were employed for UVMR, while the MVMR-Lasso method was applied for MVMR in the supplementary analyses. In addition, we conducted a series of thorough sensitivity studies to determine the accuracy of our MR findings.

Results

The UVMR analysis demonstrated a significant association between PM2.5 exposure and an increased risk of obesity, as indicated by the IVW model (odds ratio [OR]: 6.427; 95% confidence interval [CI]: 1.881-21.968; P FDR = 0.005). Additionally, PM2.5 concentrations were positively associated with fat distribution metrics, including visceral adipose tissue (VAT) (OR: 1.861; 95% CI: 1.244-2.776; P FDR = 0.004), particularly pancreatic fat (OR: 3.499; 95% CI: 2.092-5.855; PFDR =1.28E-05), and abdominal subcutaneous adipose tissue (ASAT) volume (OR: 1.773; 95% CI: 1.106-2.841; P FDR = 0.019). Furthermore, PM2.5 exposure correlated positively with markers of glucose and lipid metabolism, specifically triglycerides (TG) (OR: 19.959; 95% CI: 1.269-3.022; P FDR = 0.004) and glycated hemoglobin (HbA1c) (OR: 2.462; 95% CI: 1.34-4.649; P FDR = 0.007). Finally, a significant negative association was observed between PM2.5 concentrations and levels of the novel obesity-related biomarker fibroblast growth factor 21 (FGF-21) (OR: 0.148; 95% CI: 0.025-0.89; P FDR = 0.037). After adjusting for confounding factors, including external smoke exposure, physical activity, educational attainment (EA), participation in sports clubs or gym leisure activities, and Townsend deprivation index at recruitment (TDI), the MVMR analysis revealed that PM2.5 levels maintained significant associations with pancreatic fat, HbA1c, and FGF-21.

Conclusion

Our MR study demonstrates conclusively that higher PM2.5 concentrations are associated with an increased risk of obesity-related indicators such as pancreatic fat content, HbA1c, and FGF-21. The potential mechanisms require additional investigation.

Harnessing AI to unmask Copenhagen's invisible air pollutants: A study on three ultrafine particle metrics

Ultrafine particles (UFPs) are airborne particles with a diameter of less than 100 nm. They are emitted from various sources, such as traffic, combustion, and industrial processes, and can have adverse effects on human health. Long-term mean ambient average particle size (APS) in the UFP range varies over space within cities, with locations near UFP sources having typically smaller APS. Spatial models for lung deposited surface area (LDSA) within urban areas are limited and currently there is no model for APS in any European city. We collected particle number concentration (PNC), LDSA, and APS data over one-year monitoring campaign from May 2021 to May 2022 across 27 locations and estimated annual mean in Copenhagen, Denmark, and obtained additionally annual mean PNC data from 6 state-owned continuous monitors. We developed 94 predictor variables, and machine learning models (random forest and bagged tree) were developed for PNC, LDSA, and APS. The annual mean PNC, LDSA, and APS were, respectively, 5523 pt/cm3, 12.0 μm2/cm3, and 46.1 nm. The final R2 values by random forest (RF) model were 0.93 for PNC, 0.88 for LDSA, and 0.85 for APS. The 10-fold, repeated 10-times cross-validation R2 values were 0.65, 0.67, and 0.60 for PNC, LDSA, and APS, respectively. The root mean square error for final RF models were 296 pt/cm3, 0.48 μm2/cm3, and 1.60 nm for PNC, LDSA, and APS, respectively. Traffic-related variables, such as length of major roads within buffers 100-150 m and distance to streets with various speed limits were amongst the highly-ranked predictors for our models. Overall, our ML models achieved high R2 values and low errors, providing insights into UFP exposure in a European city where average PNC is quite low. These hyperlocal predictions can be used to study health effects of UFPs in the Danish Capital.

A machine learning-based ensemble model for estimating diurnal variations of nitrogen oxide concentrations in Taiwan

Air pollution is inextricable from human activity patterns. This is especially true for nitrogen oxide (NOx), a pollutant that exists naturally and also as a result of anthropogenic factors. Assessing exposure by considering diurnal variation is a challenge that has not been widely studied. Incorporating 27 years of data, we attempted to estimate diurnal variations in NOx across Taiwan. We developed a machine learning-based ensemble model that integrated hybrid kriging-LUR, machine-learning, and an ensemble learning approach. Hybrid kriging-LUR was performed to select the most influential predictors, and machine-learning algorithms were applied to improve model performance. The three best machine-learning algorithms were suited and reassessed to develop ensemble learning that was designed to improve model performance. Our ensemble model resulted in estimates of daytime, nighttime, and daily NOx with high explanatory powers (Adj-R2) of 0.93, 0.98, and 0.94, respectively. These explanatory powers increased from the initial model that used only hybrid kriging-LUR. Additionally, the results depicted the temporal variation of NOx, with concentrations higher during the daytime than the nighttime. Regarding spatial variation, the highest NOx concentrations were identified in northern and western Taiwan. Model evaluations confirmed the reliability of the models. This study could serve as a reference for regional planning supporting emission control for environmental and human health.

Inequalities in urban air pollution in sub-Saharan Africa: an empirical modeling of ambient NO and NO2 concentrations in Accra, Ghana

Road traffic has become the leading source of air pollution in fast-growing sub-Saharan African cities. Yet, there is a dearth of robust city-wide data for understanding space-time variations and inequalities in combustion related emissions and exposures. We combined nitrogen dioxide (NO2) and nitric oxide (NO) measurement data from 134 locations in the Greater Accra Metropolitan Area (GAMA), with geographical, meteorological, and population factors in spatio-temporal mixed effects models to predict NO2 and NO concentrations at fine spatial (50 m) and temporal (weekly) resolution over the entire GAMA. Model performance was evaluated with 10-fold cross-validation (CV), and predictions were summarized as annual and seasonal (dusty [Harmattan] and rainy [non-Harmattan]) mean concentrations. The predictions were used to examine population distributions of, and socioeconomic inequalities in, exposure at the census enumeration area (EA) level. The models explained 88% and 79% of the spatiotemporal variability in NO2 and NO concentrations, respectively. The mean predicted annual, non-Harmattan and Harmattan NO2 levels were 37 (range: 1-189), 28 (range: 1-170) and 50 (range: 1-195) µg m-3, respectively. Unlike NO2, NO concentrations were highest in the non-Harmattan season (41 [range: 31-521] µg m-3). Road traffic was the dominant factor for both pollutants, but NO2 had higher spatial heterogeneity than NO. For both pollutants, the levels were substantially higher in the city core, where the entire population (100%) was exposed to annual NO2 levels exceeding the World Health Organization (WHO) guideline of 10 µg m-3. Significant disparities in NO2 concentrations existed across socioeconomic gradients, with residents in the poorest communities exposed to levels about 15 µg m-3 higher compared with the wealthiest (p < 0.001). The results showed the important role of road traffic emissions in air pollution concentrations in the GAMA, which has major implications for the health of the city's poorest residents. These data could support climate and health impact assessments as well as policy evaluations in the city.

Outdoor Ultrafine Particulate Matter and Risk of Lung Cancer in Southern California

Rationale: Particulate matter ⩽2.5 μm in aerodynamic diameter (PM2.5) is an established cause of lung cancer, but the association with ultrafine particulate matter (UFP; aerodynamic diameter < 0.1 μm) is unclear. Objectives: To investigate the association between UFP and lung cancer overall and by histologic subtype. Methods: The Los Angeles Ultrafines Study includes 45,012 participants aged ⩾50 years in southern California at enrollment (1995-1996) followed through 2017 for incident lung cancer (n = 1,770). We estimated historical residential ambient UFP number concentrations via land use regression and back extrapolation using PM2.5. In Cox proportional hazards models adjusted for smoking and other confounders, we estimated associations between 10-year lagged UFP (per 10,000 particles/cm3 and quartiles) and lung cancer overall and by major histologic subtype (adenocarcinoma, squamous cell carcinoma, and small cell carcinoma). We also evaluated relationships by smoking status, birth cohort, and historical duration at the residence. Measurements and Main Results: UFP was modestly associated with lung cancer risk overall (hazard ratio [HR], 1.03 [95% confidence interval (CI), 0.99-1.08]). For adenocarcinoma, we observed a positive trend among men; risk was increased in the highest exposure quartile versus the lowest (HR, 1.39 [95% CI, 1.05-1.85]; P for trend = 0.01) and was also increased in continuous models (HR per 10,000 particles/cm3, 1.09 [95% CI, 1.00-1.18]), but no increased risk was apparent among women (P for interaction = 0.03). Adenocarcinoma risk was elevated among men born between 1925 and 1930 (HR, 1.13 [95% CI, 1.02-1.26] per 10,000) but not for other birth cohorts, and was suggestive for men with ⩾10 years of residential duration (HR, 1.11 [95% CI, 0.98-1.26]). We found no consistent associations for women or other histologic subtypes. Conclusions: UFP exposure was modestly associated with lung cancer overall, with stronger associations observed for adenocarcinoma of the lung.

Development of Spatio-Temporal Land Use Regression Models for Fine Particulate Matter and Wood-Burning Tracers in Temuco, Chile

Biomass burning is common in much of the world, and in some areas, residential wood-burning has increased. However, air pollution resulting from biomass burning is an important public health problem. A sampling campaign was carried out between May 2017 and July 2018 in over 64 sites in four sessions, to develop a spatio-temporal land use regression (LUR) model for fine particulate matter (PM) and wood-burning tracers levoglucosan and soluble potassium (Ksol) in a city heavily impacted by wood-burning. The mean (sd) was 46.5 (37.4) μg m-3 for PM2.5, 0.607 (0.538) μg m-3 for levoglucosan, and 0.635 (0.489) μg m-3 for Ksol. LUR models for PM2.5, levoglucosan, and Ksol had a satisfactory performance (LOSOCV R2), explaining 88.8%, 87.4%, and 87.3% of the total variance, respectively. All models included sociodemographic predictors consistent with the pattern of use of wood-burning in homes. The models were applied to predict concentrations surfaces and to estimate exposures for an epidemiological study.

The causal relationship between air pollution, obesity, and COVID-19 risk: a large-scale genetic correlation study

Objective

Observational evidence reported that air pollution is a significant risk element for numerous health problems, such as obesity and coronavirus disease 2019 (COVID-19), but their causal relationship is currently unknown. Our objective was to probe the causal relationship between air pollution, obesity, and COVID-19 and to explore whether obesity mediates this association.

Methods

We obtained instrumental variables strongly correlated to air pollutants [PM2.5, nitrogen dioxide (NO2) and nitrogen oxides (NOx)], 9 obesity-related traits (abdominal subcutaneous adipose tissue volume, waist-to-hip ratio, body mass index, hip circumference, waist circumference, obesity class 1-3, visceral adipose tissue volume), and COVID-19 phenotypes (susceptibility, hospitalization, severity) from public genome-wide association studies. We used clinical and genetic data from different public biological databases and performed analysis by two-sample and two-step Mendelian randomization.

Results

PM2.5 genetically correlated with 5 obesity-related traits, which obesity class 1 was most affected (beta = 0.38, 95% CI = 0.11 - 0.65, p = 6.31E-3). NO2 genetically correlated with 3 obesity-related traits, which obesity class 1 was also most affected (beta = 0.33, 95% CI = 0.055 - 0.61, p = 1.90E-2). NOx genetically correlated with 7 obesity-related traits, which obesity class 3 was most affected (beta = 1.16, 95% CI = 0.42-1.90, p = 2.10E-3). Almost all the obesity-related traits genetically increased the risks for COVID-19 phenotypes. Among them, body mass index, waist circumference, hip circumference, waist-to-hip ratio, and obesity class 1 and 2 mediated the effects of air pollutants on COVID-19 risks (p < 0.05). However, no direct causal relationship was observed between air pollution and COVID-19.

Conclusion

Our study suggested that exposure to heavy air pollutants causally increased risks for obesity. Besides, obesity causally increased the risks for COVID-19 phenotypes. Attention needs to be paid to weight status for the population who suffer from heavy air pollution, as they are more likely to be susceptible and vulnerable to COVID-19.

Causal relationship between particulate matter 2.5 and diabetes: two sample Mendelian randomization

Backgrounds

Many studies have shown particulate matter has emerged as one of the major environmental risk factors for diabetes; however, studies on the causal relationship between particulate matter 2.5 (PM2.5) and diabetes based on genetic approaches are scarce. The study estimated the causal relationship between diabetes and PM2.5 using two sample mendelian randomization (TSMR).

Methods

We collected genetic data from European ancestry publicly available genome wide association studies (GWAS) summary data through the MR-BASE repository. The IEU GWAS information output PM2.5 from the Single nucleotide polymorphisms (SNPs) GWAS pipeline using pheasant-derived variables (Consortium = MRC-IEU, sample size: 423,796). The annual relationship of PM2.5 (2010) were modeled for each address using a Land Use Regression model developed as part of the European Study of Cohorts for Air Pollution Effects. Diabetes GWAS information (Consortium = MRC-IEU, sample size: 461,578) were used, and the genetic variants were used as the instrumental variables (IVs). We performed three representative Mendelian Randomization (MR) methods: Inverse Variance Weighted regression (IVW), Egger, and weighted median for causal relationship using genetic variants. Furthermore, we used a novel method called MR Mixture to identify outlier SNPs.

Results

From the IVW method, we revealed the causal relationship between PM2.5 and diabetes (Odds ratio [OR]: 1.041, 95% CI: 1.008-1.076, P = 0.016), and the finding was substantiated by the absence of any directional horizontal pleiotropy through MR-Egger regression (β = 0.016, P = 0.687). From the IVW fixed-effect method (i.e., one of the MR machine learning mixture methods), we excluded outlier SNP (rs1537371) and showed the best predictive model (AUC = 0.72) with a causal relationship between PM2.5 and diabetes (OR: 1.028, 95% CI: 1.006-1.049, P = 0.012).

Conclusion

We identified the hypothesis that there is a causal relationship between PM2.5 and diabetes in the European population, using MR methods.

Spatiotemporal Heterogeneity of Lung-Deposited Surface Area in Zurich Switzerland: Lung-Deposited Surface Area as a New Routine Metric for Ambient Particle Monitoring

Objective: To assess the spatiotemporal heterogeneity of lung-deposited particle surface area concentration (LDSA), while testing the long-term performance of a prototype of low-cost-low-maintenance LDSA sensors. One factor hampering epidemiological studies on fine to ultrafine particles (F-to-UFP) exposure is exposure error due to their high spatiotemporal heterogeneity, not reflected in particle mass. Though LDSA shows consistent associations between F-to-UFP exposure and health effects, LDSA data are limited. Methods: We measured LDSA in a network of ten sensors, including urban, suburban, and rural environments in Zurich, Switzerland. With traffic counts, traffic co-pollutant concentrations, and meteorological parameters, we assessed the drivers of the LDSA observations. Results: LDSA reflected the high spatiotemporal heterogeneity of F-to-UFP. With micrometeorological influences, local sources like road traffic, restaurants, air traffic, and residential combustion drove LDSA. The temporal pattern of LDSA reflected that of the local sources. Conclusion: LDSA may be a viable metric for inexpensively characterizing F-to-UFP exposure. The tested devices generated sound data and may significantly contribute to filling the LDSA exposure data gap, providing grounds for more statistically significant epidemiological studies and regulation of F-to-UFP.

Assessing the Spatiotemporal Characteristics, Factor Importance, and Health Impacts of Air Pollution in Seoul by Integrating Machine Learning into Land-Use Regression Modeling at High Spatiotemporal Resolutions

Previous studies have characterized spatial patterns of air pollution with land-use regression (LUR) models. However, the spatiotemporal characteristics of air pollution, the contribution of various factors to them, and the resultant health impacts have yet to be evaluated comprehensively. This study integrates machine learning (random forest) into LUR modeling (LURF) with intensive evaluations to develop high spatiotemporal resolution prediction models to estimate daily and diurnal PM2.5 and NO2 in Seoul, South Korea, at the spatial resolution of 500 m for a year (2019) and to then evaluate the contribution of driving factors and quantify the resultant premature mortality. Our results show that incorporating the random forest algorithm into our LUR model improves the model performance. Meteorological conditions have a great influence on daily models, while land-use factors play important roles in diurnal models. Our health assessment using dynamic population data estimates that PM2.5 and NO2 pollution, when combined, causes a total of 11,183 (95% CI: 5837-16,354) premature mortalities in Seoul in 2019, of which 64.9% are due to PM2.5, while the remaining are attributable to NO2. The air pollution-attributable health impacts in Seoul are largely caused by cardiovascular diseases including stroke. This study pinpoints the significant spatiotemporal variations and health impact of PM2.5 and NO2 in Seoul, providing essential data for epidemiological research and air quality management.

Revisiting the atmospheric particles: Connecting lines and changing paradigms

Historically, the atmospheric particles constitute the most primitive and recent class of air pollutants. The science of atmospheric particles erupted more than a century ago covering more than four decades of size, with past few years experiencing major advancements on both theoretic and data-based observational grounds. More recently, the plausible recognition between particulate matter (PM) and the diffusion of the COVID-19 pandemic has led to the accretion of interest in particle science. With motivation from diverse particle research interests, this paper is an 'old engineer's survey' beginning with the evolution of atmospheric particles and identifies along the way many of the global instances signaling the 'size concept' of PM. A theme that runs through the narrative is a 'previously known' generational evolution of particle science to the 'newly procured' portfolio of knowledge, with important gains on the application of unmet concepts and future approaches to PM exposure and epidemiological research.

Tracing out the effect of transportation infrastructure on NO2 concentration levels with Kernel Density Estimation by investigating successive COVID-19-induced lockdowns

This study aims to investigate the effect of transportation infrastructure on the decrease of NO₂ air pollution during three COVID-19-induced lockdowns in a vast region of France. For this purpose, using Sentinel-5P satellite data, the relative change in tropospheric NO₂ air pollution during the three lockdowns was calculated. The estimation of regional infrastructure intensity was performed using Kernel Density Estimation, being the predictor variable. By performing hotspot-coldspot analysis on the relative change in NO₂ air pollution, significant spatial clusters of decreased air pollution during the three lockdowns were identified. Based on the clusters, a novel spatial index, the Clustering Index (CI) was developed using its Coldspot Clustering Index (CCI) variant as a predicted variable in the regression model between infrastructure intensity and NO₂ air pollution decline. The analysis revealed that during the three lockdowns there was a strong and statistically significant relationship between the transportation infrastructure and the decline index, CCI (r = 0.899, R² = 0.808). The results showed that the largest decrease in NO₂ air pollution was recorded during the first lockdown, and in this case, there was the strongest inverse correlation with transportation infrastructure (r = -0.904, R² = 0.818). Economic and population predictors also explained with good fit the decrease in NO₂ air pollution during the first lockdown: GDP (R² = 0.511), employees (R² = 0.513), population density (R² = 0.837). It is concluded that not only economic-population variables determined the reduction of near-surface air pollution but also the transportation infrastructure. Further studies are recommended to investigate other pollutant gases as predicted variables.

Characterization of Annual Average Traffic-Related Air Pollution Concentrations in the Greater Seattle Area from a Year-Long Mobile Monitoring Campaign

Growing evidence links traffic-related air pollution (TRAP) to adverse health effects. We designed an innovative and extensive mobile monitoring campaign to characterize TRAP exposure levels for the Adult Changes in Thought (ACT) study, a Seattle-based cohort. The campaign measured particle number concentration (PNC) to capture ultrafine particles (UFP), black carbon (BC), nitrogen dioxide (NO2), fine particulate matter (PM2.5), and carbon dioxide (CO2) at 309 roadside sites within a large, 1200 land km2 (463 mi2) area representative of the cohort. We collected about 29 two-minute measurements at each site during all seasons, days of the week, and most times of the day over a 1-year period. Validation showed good agreement between our BC, NO2, and PM2.5 measurements and monitoring agency sites (R2 = 0.68-0.73). Universal kriging-partial least squares models of annual average pollutant concentrations had cross-validated mean square error-based R2 (and root mean square error) values of 0.77 (1177 pt/cm3) for PNC, 0.60 (102 ng/m3) for BC, 0.77 (1.3 ppb) for NO2, 0.70 (0.3 μg/m3) for PM2.5, and 0.51 (4.2 ppm) for CO2. Overall, we found that the design of this extensive campaign captured the spatial pollutant variations well and these were explained by sensible land use features, including those related to traffic.

Developing air pollution concentration fields for health studies using multiple methods: Cross-comparison and evaluation

Past air pollution epidemiological studies have used a wide range of methods to develop concentration fields for health analyses. The fields developed differ considerably among these methods. The reasons for these differences and comparisons of their strengths, as well as the limitations for estimating exposures, remains under-investigated. Here, we applied nine methods to develop fields of eight pollutants (carbon monoxide (CO), nitrogen dioxide (NO₂), sulfur dioxide (SO₂), ozone (O₃), fine particulate matter (PM_2.5), and three speciated PM_2.5 constituents including elemental carbon (EC), organic carbon (OC), and sulfate (SO₄)) for the metropolitan Atlanta region for five years. The nine methods are Central Monitor (CM), Site Average (SA), Inverse Distance Weighting (IDW), Kriging (KRIG), Land Use Regression (LUR), satellite Aerosol Optical Depth (AOD), CMAQ model, CMAQ with kriging adjustment (CMAQ-KRIG), and CMAQ based data fusion (CMAQ-DF). Additionally, we applied an increasingly popular method, Random Forest (RF), and compared its results for NO₂ and PM_2.5 with other methods. For statistical evaluation, we focused our discussion on the temporal coefficient of determination, although other metrics are also calculated. Raw output from the CMAQ model contains modeling biases and errors, which are partially mitigated by fusing observational data in the CMAQ-KRIG and CMAQ-DF methods. Based on analyses that simulated model responses to more limited input data, the RF model is more robust and outperforms LUR for PM_2.5. These results suggest RF may have potential in air pollution health studies, especially when limited measurement data are available. The RF method has several important weaknesses, including a relatively poor performance for NO₂, diagnostic challenges, and computational intensiveness. The results of this study will help to improve our understanding of the strengths and weaknesses of different methods for estimating air pollutant exposures in epidemiological studies.

Spatialization and Prediction of Seasonal NO2 Pollution Due to Climate Change in the Korean Capital Area through Land Use Regression Modeling

Urbanization is causing an increase in air pollution leading to serious health issues. However, even though the necessity of its regulation is acknowledged, there are relatively few monitoring sites in the capital metropolitan city of the Republic of Korea. Furthermore, a significant relationship between air pollution and climate variables is expected, thus the prediction of air pollution under climate change should be carefully attended. This study aims to predict and spatialize present and future NO₂ distribution by using existing monitoring sites to overcome deficiency in monitoring. Prediction was conducted through seasonal Land use regression modeling using variables correlated with NO₂ concentration. Variables were selected through two correlation analyses and future pollution was predicted under HadGEM-AO RCP scenarios 4.5 and 8.5. Our results showed a relatively high NO₂ concentration in winter in both present and future predictions, resulting from elevated use of fossil fuels in boilers, and also showed increments of NO₂ pollution due to climate change. The results of this study could strengthen existing air pollution management strategies and mitigation measures for planning concerning future climate change, supporting proper management and control of air pollution.

High spatial resolution land-use regression model for urban ultrafine particle exposure assessment in Shanghai, China

BACKGROUND

Little is currently known about long-term health effects of ambient ultrafine particles (UFPs) due to the lack of exposure assessment metrics suitable for use in large population-based studies. Land use regression (LUR) models have been used increasingly for modeling small-scale spatial variation in UFPs concentrations in European and American, but have never been applied in developing countries with heavy air pollution.

OBJECTIVE

This study developed a land-use regression (LUR) model for UFP exposure assessment in Shanghai, a typic mega city of China, where dense population resides.

METHOD

A 30-minute measurement of particle number concentrations of UFPs was collected at each visit at 144 fixed sites, and each was visited three times in each season of winter, spring, and summer. The annual adjusted average was calculated and regressed against pre-selected geographic information system-derived predictor variables using a stepwise variable selection method.

RESULT

The final LUR model explained 69% of the spatial variability in UFP with a root mean square error of 6008 particles cm^-3. The 10-fold cross validation R² reached 0.68, revealing the robustness of the model. The final predictors included traffic-related NO_x emissions, number of restaurants, building footprint area, and distance to the nearest national road. These predictors were within a relatively small buffer size, ranging from 50 m to 100 m, indicating great spatial variations of UFP particle number concentration and the need of high-resolution models for UFP exposure assessment in urban areas.

CONCLUSION

We concluded that based on a purpose-designed short-term monitoring network, LUR model can be applied to predict UFPs spatial surface in a mega city of China. Majority of the spatial variability in the annual mean of ambient UFP was explained in the model comprised primarily of traffic-, building-, and restaurant-related predictors.

Residential greenness-related DNA methylation changes

BACKGROUND

Residential greenness has been associated with health benefits, but its biological mechanism is largely unknown. Investigation of greenness-related DNA methylation profiles can contribute to mechanistic understanding of the health benefits of residential greenness.

OBJECTIVE

To identify DNA methylation profiles associated with greenness in the immediate surroundings of the residence.

METHODS

We analyzed genome-wide DNA methylation in 1938 blood samples (982 participants) from the Swiss Cohort Study on Air Pollution and Lung and Heart Diseases in Adults (SAPALDIA). We estimated residential greenness based on normalized difference vegetation index at 30 × 30 m cell (green30) and 500 m buffer (green500) around the residential address. We conducted epigenome-wide association study (EWAS) to identify differentially methylated CpGs and regions, and enrichment tests by comparing to the CpGs that previous EWAS identified as associated with allergy, physical activity, and allostatic load-relevant biomarkers.

RESULTS

We identified no genome-wide significant CpGs, but 163 and 56 differentially methylated regions for green30 and green500, respectively. Green30-related DNA methylation profiles showed enrichments in allergy, physical activity, and allostatic load, while green500-related methylation was enriched in allergy and allostatic load.

CONCLUSIONS

Residential greenness may have health impacts through allergic sensitization, stress coping, or behavioral changes. Exposure to more proximal greenness may be more health-relevant.

Spatial and temporal variability of airborne ultrafine particles in the Greater Montreal area: Results of monitoring campaigns in two seasons

It has been hypothesized that ultrafine particles (UFP) in air pollution may cause lung cancer. In preparation for an epidemiologic case-control study to assess this hypothesis in Montreal, Canada, we conducted a UFP measurement campaign in order to create an exposure surface with which we could assign UFP exposure to subjects corresponding to their residential addresses. The purpose of this paper is to describe the temporal and spatial variability that underlies the creation of an exposure surface in the Montreal area, and to consider the implications for epidemiological exposure assessment. We identified 249 fixed sampling sites, selected to provide a dense spatial representation of the areas of residence of Montreal residents. We conducted a winter campaign and a summer campaign, and each of the sites was visited three times during each seasonal campaign. Each visit entailed a 20-minute measurement period for UFPs with a separate measurement each second. This provided data for temporal comparisons at each site between seasons, between visits and between seconds. The median of UFP measurements was 16,593 particles/cm³ in winter and 8919 particles/cm³ in summer. Across the 249 sampling sites the Spearman correlation coefficient between the UFP measurements of winter and summer was 0.35. Within each visit, correlation was below 0.50 between pairs of UFP measurements taken more than 60 s apart, and there was hardly any correlation among measurements taken more than 300 s apart. When sites were grouped by proximity to certain types of pollution sources, and the seven resulting groups compared, there were modest, albeit statistically significant, differences in UFP levels. There was moderate positive spatial autocorrelation in UFPs over the study area. High temporal variability of UFPs from short-term measurements campaigns will likely compromise the predictive validity of the exposure surface, and will eventually attenuate the epidemiologic risk estimates.

Potential for developing independent daytime/nighttime LUR models based on short-term mobile monitoring to improve model performance

Land use regression model (LUR) is a widespread method for predicting air pollution exposure. Few studies have explored the performance of independently developed daytime/nighttime LUR models. In this study, fine particulate matter (PM_2.5), inhalable particulate matter (PM₁₀), and nitrogen dioxide (NO₂) concentrations were measured by mobile monitoring during non-heating and heating seasons in Taiyuan. Pollutant concentrations were higher in the nighttime than the daytime, and higher in the heating season than the non-heating season. Daytime/nighttime and full-day LUR models were developed and validated for each pollutant to examine variations in model performance. Adjusted coefficients of determination (adjusted R²) for the LUR models ranged from 0.53-0.87 (PM_2.5), 0.53-0.85 (PM₁₀), and 0.33-0.67 (NO₂). The performance of the daytime/nighttime LUR models for PM_2.5 and PM₁₀ was better than that of the full-day models according to the results of model adjusted R² and validation R². Consistent results were confirmed in the non-heating and heating seasons. Effectiveness of developing independent daytime/nighttime models for NO₂ to improve performance was limited. Surfaces based on the daytime/nighttime models revealed variations in concentrations and spatial distribution. In conclusion, the independent development of daytime/nighttime LUR models for PM_2.5/PM₁₀ has the potential to replace full-day models for better model performance. The modeling strategy is consistent with the residential activity patterns and contributes to achieving reliable exposure predictions for PM_2.5 and PM₁₀. Nighttime could be a critical exposure period, due to high pollutant concentrations.

Comparison of Spatial Modelling Approaches on PM10 and NO2 Concentration Variations: A Case Study in Surabaya City, Indonesia

Because of fast-paced industrialization, urbanization, and population growth in Indonesia, there are serious health issues in the country resulting from air pollution. This study uses geospatial modelling technologies, namely land-use regression (LUR), geographically weighted regression (GWR), and geographic and temporal weighted regression (GTWR) models, to assess variations in particulate matter (PM₁₀) and nitrogen dioxide (NO₂) concentrations in Surabaya City, Indonesia. This is the first study to implement spatiotemporal variability of air pollution concentrations in Surabaya City, Indonesia. To develop the prediction models, air pollution data collected from seven monitoring stations from 2010 to 2018 were used as dependent variables, while land-use/land cover allocations within a 250 m to 5000 m circular buffer range surrounding the monitoring stations were collected as independent variables. A supervised stepwise variable selection procedure was applied to identify the important predictor variables for developing the LUR, GWR, and GTWR models. The developed models of LUR, GWR, and GTWR accounted for 49%, 50%, and 51% of PM₁₀ variations and 46%, 47%, and 48% of NO₂ variations, respectively. The GTWR model performed better (R² = 0.51 for PM₁₀ and 0.48 for NO₂) than the other two models (R² = 0.49–0.50 for PM₁₀ and 0.46–0.47 for NO₂), LUR and GWR. In the PM₁₀ model four predictor variables, public facility, industry and warehousing, paddy field, and normalized difference vegetation index (NDVI), were selected during the variable selection procedure. Meanwhile, paddy field, residential area, rainfall, and temperature played important roles in explaining NO₂ variations. Because of biomass burning issues in South Asia, the paddy field, which has a positive correlation with PM₁₀ and NO₂, was selected as a predictor. By using long-term monitoring data to establish prediction models, this model may better depict PM₁₀ and NO₂ concentration variations within areas across Asia.

Incidence of depression in relation to transportation noise exposure and noise annoyance in the SAPALDIA study

Prospective evidence on the risk of depression in relation to transportation noise exposure and noise annoyance is limited and mixed. We aimed to investigate the associations of long-term exposure to source-specific transportation noise and noise annoyance with incidence of depression in the SAPALDIA (Swiss cohort study on air pollution and lung and heart diseases in adults) cohort. We investigated 4,581 SAPALDIA participants without depression in the year 2001/2002. Corresponding one-year mean road, railway and aircraft day-evening-night noise (Lden) was calculated at the most exposed façade of the participants' residential floors, and transportation noise annoyance was assessed on an 11-point scale. Incident cases of depression were identified in 2010/2011, and comprised participants reporting physician diagnosis, intake of antidepressant medication or having a short form-36 mental health score < 50. We used robust Poisson regressions to estimate the mutually adjusted relative risks (RR) and 95% confidence intervals (CI) of depression, independent of traffic-related air pollution and other potential confounders. Incidence of depression was 11 cases per 1,000 person-years. In single exposure models, we observed positive but in part, statistically non-significant associations (per 10 dB) of road traffic Lden [RR: 1.06 (0.93, 1.22)] and aircraft Lden [RR: 1.19 (0.93, 1.53)], and (per 1-point difference) of noise annoyance [RR: 1.05 (1.02, 1.08)] with depression risk. In multi-exposure model, noise annoyance effect remained unchanged, with weaker effects of road traffic Lden [(RR: 1.02 (0.89, 1.17)] and aircraft Lden [(RR: 1.17 (0.90, 1.50)]. However, there were statistically significant indirect effects of road traffic Lden [(β: 0.02 (0.01, 0.03)] and aircraft Lden [β: 0.01 (0.002, 0.02)] via noise annoyance. There were no associations with railway Lden in the single and multi-exposure models [(RR_{both models}: 0.88 (0.75, 1.03)]. We made similar findings among 2,885 non-movers, where the effect modification and cumulative risks were more distinct. Noise annoyance effect in non-movers was stronger among the insufficiently active (RR: 1.09; 95%CI: 1.02, 1.17; p_interaction = 0.07) and those with daytime sleepiness [RR: 1.07 (1.02, 1.12); p_interaction = 0.008]. Cumulative risks of Lden in non-movers showed additive tendencies for the linear cumulative risk [(RR_{per 10dB of combined sources}: 1.31 (0.90, 1.91)] and the categorical cumulative risk [(RR_{triple- vs. zero-source ≥45 dB}: 2.29 (1.02, 5.14)], and remained stable to noise annoyance. Transportation noise level and noise annoyance may jointly and independently influence the risk of depression. Combined long-term exposures to noise level seems to be most detrimental, largely acting via annoyance. The moderation of noise annoyance effect by daytime sleepiness and physical activity further contribute to clarifying the involved mechanisms. More evidence is needed to confirm these findings for effective public health control of depression and noise exposure burden.

Development and transferability of ultrafine particle land use regression models in London

Due to a lack of routine monitoring, bespoke measurements are required to develop ultrafine particle (UFP) land use regression (LUR) models, which is especially challenging in megacities due to their large area. As an alternative, for London, we developed separate models for three urban residential areas, models combining two areas, and models using all three areas. Models were developed against annual mean ultrafine particle count cm^-3 estimated from repeated 30-min fixed-site measurements, in different seasons (2016-2018), at forty sites per area, that were subsequently temporally adjusted using continuous measurements from a single reference site within or close to each area. A single model and 10 models were developed for each individual area and combination of areas. Within each area, sites were split into 10 groups using stratified random sampling. Each of the 10 models were developed using 90% of sites. Hold-out validation was performed by pooling the 10% of sites held-out each time. The transferability of models was tested by applying individual and two-area models to external area(s). In model evaluation, within-area mean squared error (MSE) R² ranged from 14% to 48%. Transferring individual- and combined-area models to external areas without calibration yielded MSE-R² ranging from -18 to 0. MSE-R² was in the range 21% to 41% when using particle number count (PNC) measurements in external areas to calibrate models. Our results suggest that the UFP models could be transferred to other areas without calibration in London to assess relative ranking in exposures but not for estimating absolute values of PNC.

Genome-Wide DNA Methylation in Peripheral Blood and Long-Term Exposure to Source-Specific Transportation Noise and Air Pollution: The SAPALDIA Study

BACKGROUND

Few epigenome-wide association studies (EWAS) on air pollutants exist, and none have been done on transportation noise exposures, which also contribute to environmental burden of disease.

OBJECTIVE

We performed mutually independent EWAS on transportation noise and air pollution exposures.

METHODS

We used data from two time points of the Swiss Cohort Study on Air Pollution and Lung and Heart Diseases in Adults (SAPALDIA) from 1,389 participants contributing 2,542 observations. We applied multiexposure linear mixed-effects regressions with participant-level random intercept to identify significant Cytosine-phosphate-Guanine (CpG) sites and differentially methylated regions (DMRs) in relation to 1-y average aircraft, railway, and road traffic day-evening-night noise (Lden); nitrogen dioxide (NO 2 ); and particulate matter (PM) with aerodynamic diameter < 2.5 μ m (PM 2.5 ). We performed candidate (CpG-based; cross-systemic phenotypes, combined into "allostatic load") and agnostic (DMR-based) pathway enrichment tests, and replicated previously reported air pollution EWAS signals.

RESULTS

We found no statistically significant CpGs at false discovery rate < 0.05 . However, 14, 48, 183, 8, and 71 DMRs independently associated with aircraft, railway, and road traffic Lden; NO 2 ; and PM 2.5 , respectively, with minimally overlapping signals. Transportation Lden and air pollutants tendentially associated with decreased and increased methylation, respectively. We observed significant enrichment of candidate DNA methylation related to C-reactive protein and body mass index (aircraft, road traffic Lden, and PM 2.5 ), renal function and "allostatic load" (all exposures). Agnostic functional networks related to cellular immunity, gene expression, cell growth/proliferation, cardiovascular, auditory, embryonic, and neurological systems development were enriched. We replicated increased methylation in cg08500171 (NO 2 ) and decreased methylation in cg17629796 (PM 2.5 ).

CONCLUSIONS

Mutually independent DNA methylation was associated with source-specific transportation noise and air pollution exposures, with distinct and shared enrichments for pathways related to inflammation, cellular development, and immune responses. These findings contribute in clarifying the pathways linking these exposures and age-related diseases but need further confirmation in the context of mediation analyses. https://doi.org/10.1289/EHP6174.

A hybrid kriging/land-use regression model with Asian culture-specific sources to assess NO2 spatial-temporal variations

Kriging interpolation and land use regression (LUR) have characterized the spatial variability of long-term nitrogen dioxide (NO₂), but there has been little research on combining these two methods to capture small-scale spatial variation. Furthermore, studies predicting NO₂ exposure are almost exclusively based on traffic-related variables, which may not be transferable to Taiwan, a typical Asian country with diverse local emission sources, where densely distributed temples and restaurants may be important for NO₂ levels. To advance the exposure estimates in Taiwan, a hybrid kriging/LUR model incorporates culture-specific sources as potential predictors. Based on 14-year NO₂ observations from 73 monitoring stations across Taiwan, a set of interpolated NO₂ values were generated through a leave-one-out ordinary kriging algorithm, and this was included as an explanatory variable in the stepwise LUR procedures. Kriging interpolated NO₂ and culture-specific predictors were entered in the final models, which captured 90% and 87% of NO₂ variation in annual and monthly resolution, respectively. Results from 10-fold cross-validation and external data verification demonstrate robust performance of the developed models. This study demonstrates the value of incorporating the kriging-interpolated estimates and culture-specific emission sources into the traditional LUR model structure for predicting NO₂, which can be particularly useful for Asian countries.

Assessing NO2 Concentration and Model Uncertainty with High Spatiotemporal Resolution across the Contiguous United States Using Ensemble Model Averaging

NO₂ is a combustion byproduct that has been associated with multiple adverse health outcomes. To assess NO₂ levels with high accuracy, we propose the use of an ensemble model to integrate multiple machine learning algorithms, including neural network, random forest, and gradient boosting, with a variety of predictor variables, including chemical transport models. This NO₂ model covers the entire contiguous U.S. with daily predictions on 1-km-level grid cells from 2000 to 2016. The ensemble produced a cross-validated R² of 0.788 overall, a spatial R² of 0.844, and a temporal R² of 0.729. The relationship between daily monitored and predicted NO₂ is almost linear. We also estimated the associated monthly uncertainty level for the predictions and address-specific NO₂ levels. This NO₂ estimation has a very high spatiotemporal resolution and allows the examination of the health effects of NO₂ in unmonitored areas. We found the highest NO₂ levels along highways and in cities. We also observed that nationwide NO₂ levels declined in early years and stagnated after 2007, in contrast to the trend at monitoring sites in urban areas, where the decline continued. Our research indicates that the integration of different predictor variables and fitting algorithms can achieve an improved air pollution modeling framework.

Assessing NO2 Concentration and Model Uncertainty with High Spatiotemporal Resolution across the Contiguous United States Using Ensemble Model Averaging

NO₂ is a combustion byproduct that has been associated with multiple adverse health outcomes. To assess NO₂ levels with high accuracy, we propose the use of an ensemble model to integrate multiple machine learning algorithms, including neural network, random forest, and gradient boosting, with a variety of predictor variables, including chemical transport models. This NO₂ model covers the entire contiguous U.S. with daily predictions on 1-km-level grid cells from 2000 to 2016. The ensemble produced a cross-validated R² of 0.788 overall, a spatial R² of 0.844, and a temporal R² of 0.729. The relationship between daily monitored and predicted NO₂ is almost linear. We also estimated the associated monthly uncertainty level for the predictions and address-specific NO₂ levels. This NO₂ estimation has a very high spatiotemporal resolution and allows the examination of the health effects of NO₂ in unmonitored areas. We found the highest NO₂ levels along highways and in cities. We also observed that nationwide NO₂ levels declined in early years and stagnated after 2007, in contrast to the trend at monitoring sites in urban areas, where the decline continued. Our research indicates that the integration of different predictor variables and fitting algorithms can achieve an improved air pollution modeling framework.

Environmental data treatment to support exposure studies: The statistical behavior for NO2, O3, PM10 and PM2.5 air concentrations in Europe

In determining and assessing external exposure, there is a need for extensive environmental data sets of sufficient time and space resolution. It is unlikely that a complete set of those data exist for a specific study. Therefore, there will be a need to fill the necessary data gaps. As a first step towards this direction, the statistical behavior of the parameters involved can be estimated so that such parameters can be statistically reconstructed in finer scales. In this study the methodology has been applied to the air concentrations of the priority pollutants NO₂, O₃, PM10 (particulate matter with an aerodynamic diameter of<10 μm) and PM2.5 (particulate matter with an aerodynamic diameter of<2.5 μm). More specifically, the hourly and the daily concentrations at a given site of those pollutants can be statistically reconstructed assuming known (a) the concentration annual average (m), (b) the pdf of the ratio of the standard deviation over the annual average (σ/m) for the hourly/daily concentrations and (c) the pdf for hourly/daily concentrations themselves. In the case that PM2.5 annual value is missing, it is estimated statistically from the PM10 annual value and the PM2.5/PM10 ratio statistics. As a first test, the proposed methodology is applied for the year 2012 arriving to concrete proposals concerning the statistical behavior of the above-mentioned parameters.

A Robust Deep Learning Approach for Spatiotemporal Estimation of Satellite AOD and PM2.5

Accurate estimation of fine particulate matter with diameter ≤2.5 μm (PM2.5) at a high spatiotemporal resolution is crucial for the evaluation of its health effects. Previous studies face multiple challenges including limited ground measurements and availability of spatiotemporal covariates. Although the multiangle implementation of atmospheric correction (MAIAC) retrieves satellite aerosol optical depth (AOD) at a high spatiotemporal resolution, massive non-random missingness considerably limits its application in PM2.5 estimation. Here, a deep learning approach, i.e., bootstrap aggregating (bagging) of autoencoder-based residual deep networks, was developed to make robust imputation of MAIAC AOD and further estimate PM2.5 at a high spatial (1 km) and temporal (daily) resolution. The base model consisted of autoencoder-based residual networks where residual connections were introduced to improve learning performance. Bagging of residual networks was used to generate ensemble predictions for better accuracy and uncertainty estimates. As a case study, the proposed approach was applied to impute daily satellite AOD and subsequently estimate daily PM2.5 in the Jing-Jin-Ji metropolitan region of China in 2015. The presented approach achieved competitive performance in AOD imputation (mean test R2: 0.96; mean test RMSE: 0.06) and PM2.5 estimation (test R2: 0.90; test RMSE: 22.3 μg/m3). In the additional independent tests using ground AERONET AOD and PM2.5 measurements at the monitoring station of the U.S. Embassy in Beijing, this approach achieved high R2 (0.82–0.97). Compared with the state-of-the-art machine learning method, XGBoost, the proposed approach generated more reasonable spatial variation for predicted PM2.5 surfaces. Publically available covariates used included meteorology, MERRA2 PBLH and AOD, coordinates, and elevation. Other covariates such as cloud fractions or land-use were not used due to unavailability. The results of validation and independent testing demonstrate the usefulness of the proposed approach in exposure assessment of PM2.5 using satellite AOD having massive missing values.

Land use regression models for ultrafine particles, fine particles, and black carbon in Southern California

Exposure models are needed to evaluate health effects of long-term exposure to ambient ultrafine particles (UFP; <0.1 μm) and to disentangle their association from other pollutants, particularly PM_2.5 (<2.5 μm). We developed land use regression (LUR) models to support UFP exposure assessment in the Los Angeles Ultrafines Study, a cohort in Southern California. We conducted a short-term measurement campaign in Los Angeles and parts of Riverside and Orange counties to measure UFP, PM_2.5, and black carbon (BC), collecting three 30-minute average measurements at 215 sites across three seasons. We averaged concentrations for each site and evaluated geographic predictors including traffic intensity, distance to airports, land use, and population and building density by supervised stepwise selection to develop models. UFP and PM_2.5 measurements (r = 0.001) and predictions (r = 0.05) were uncorrelated at the sites. UFP model explained variance was robust (R² = 0.66) and 10-fold cross-validation indicated good performance (R² = 0.59). Explained variation was moderate for PM_2.5 (R² = 0.47) and BC (R² = 0.38). In the cohort, we predicted a 2.3-fold exposure contrast from the 5^th to 95^th percentiles for all three pollutants. The correlation between modeled UFP and PM_2.5 at cohort residences was weak (r = 0.28), although higher than between measured levels. LUR models, particularly for UFP, were successfully developed and predicted reasonable exposure contrasts.

Use of Citizen Science-Derived Data for Spatial and Temporal Modeling of Particulate Matter near the US/Mexico Border

This paper describes the use of citizen science-derived data for the creation of a land-use regression (LUR) model for particulate matter (PM2.5 and PMcoarse) for a vulnerable community in Imperial County, California (CA), near the United States (US)/Mexico border. Data from the Imperial County Community Air Monitoring Network community monitors were calibrated and added to a LUR, along with meteorology and land use. PM2.5 and PMcoarse were predicted across the county at the monthly timescale. Model types were compared by cross-validated (CV) R2 and root-mean-square error (RMSE). The Bayesian additive regression trees model (BART) performed the best for both PM2.5 (CV R2 = 0.47, RMSE = 1.5 µg/m3) and PMcoarse (CV R2 = 0.65, RMSE = 8.07 µg/m3). Model predictions were also compared to measurements from the regulatory monitors. RMSE for the monthly models was 3.6 µg/m3 for PM2.5 and 17.7 µg/m3 for PMcoarse. Variable importance measures pointed to seasonality and length of roads as drivers of PM2.5, and seasonality, type of farmland, and length of roads as drivers of PMcoarse. Predicted PM2.5 was elevated near the US/Mexico border and predicted PMcoarse was elevated in the center of Imperial Valley. Both sizes of PM were high near the western edge of the Salton Sea. This analysis provides some of the initial evidence for the utility of citizen science-derived pollution measurements to develop spatial and temporal models which can make estimates of pollution levels throughout vulnerable communities.

Urban Ultrafine Particle Exposure Assessment with Land-Use Regression: Influence of Sampling Strategy

Sampling strategies in the collection of ultrafine particle (UFP) data to develop land-use regression (LUR) models can strongly influence the resulting exposure estimates. Here, we systematically examine how much sampling is needed to develop robust and stable UFP LUR models. To address this question, we collected 3-6 weeks of continuous measurements of UFP concentrations at 32 sites in Pittsburgh, Pennsylvania covering a wide range of urban land-use attributes. Through systematic subsampling of this data set, we evaluate the performance of hundreds of LUR models with varying numbers of sampling days and daily sampling durations. Our base LUR model derived from wintertime average concentrations explained about 80% of the spatial variability in the data (adjusted R² ∼ 0.8). The performance of the LUR models degrades with decreasing number of sampling days and sampling duration per day. For our data set, 1-3 h of sampling per day for 10-15 days provided UFP concentration estimates comparable to models derived from the entire data set. Small numbers of repeated sampling per site (1-3 days) at short duration (∼15-60 min per day) result in poor performance ( R² < 0.5), similar to previous UFP LUR models. This study provides guidelines for the design of future measurement campaigns and monitoring networks to generate robust UFP LUR models for exposure assessments. Further study in other locations with more sites is needed to evaluate these guidelines over a broader range of conditions.

Cluster-based bagging of constrained mixed-effects models for high spatiotemporal resolution nitrogen oxides prediction over large regions

BACKGROUND

Accurate estimation of nitrogen dioxide (NO2) and nitrogen oxide (NOx) concentrations at high spatiotemporal resolutions is crucial for improving evaluation of their health effects, particularly with respect to short-term exposures and acute health outcomes. For estimation over large regions like California, high spatial density field campaign measurements can be combined with more sparse routine monitoring network measurements to capture spatiotemporal variability of NO2 and NOx concentrations. However, monitors in spatially dense field sampling are often highly clustered and their uneven distribution creates a challenge for such combined use. Furthermore, heterogeneities due to seasonal patterns of meteorology and source mixtures between sub-regions (e.g. southern vs. northern California) need to be addressed.

OBJECTIVES

In this study, we aim to develop highly accurate and adaptive machine learning models to predict high-resolution NO2 and NOx concentrations over large geographic regions using measurements from different sources that contain samples with heterogeneous spatiotemporal distributions and clustering patterns.

METHODS

We used a comprehensive Kruskal-K-means method to cluster the measurement samples from multiple heterogeneous sources. Spatiotemporal cluster-based bootstrap aggregating (bagging) of the base mixed-effects models was then applied, leveraging the clusters to obtain balanced and less correlated training samples for less bias and improvement in generalization. Further, we used the machine learning technique of grid search to find the optimal interaction of temporal basis functions and the scale of spatial effects, which, together with spatiotemporal covariates, adequately captured spatiotemporal variability in NO2 and NOx at the state and local levels.

RESULTS

We found an optimal combination of four temporal basis functions and 200 m scale spatial effects for the base mixed-effects models. With the cluster-based bagging of the base models, we obtained robust predictions with an ensemble cross validation R2 of 0.88 for both NO2 and NOx [RMSE (RMSEIQR): 3.62 ppb (0.28) and 9.63 ppb (0.37) respectively]. In independent tests of random sampling, our models achieved similarly strong performance (R2 of 0.87-0.90; RMSE of 3.97-9.69 ppb; RMSEIQR of 0.21-0.27), illustrating minimal over-fitting.

CONCLUSIONS

Our approach has important implications for fusing data from highly clustered and heterogeneous measurement samples from multiple data sources to produce highly accurate concentration estimates of air pollutants such as NO2 and NOx at high resolution over a large region.

The mediating effect of immune markers on the association between ambient air pollution and adult-onset asthma

We aim to investigate to what extent a set of immune markers mediate the association between air pollution and adult-onset asthma. We considered long-term exposure to multiple air pollution markers and a panel of 13 immune markers in peripheral blood samples collected from 140 adult cases and 199 controls using a nested-case control design. We tested associations between air pollutants and immune markers and adult-onset asthma using mixed-effects (logistic) regression models, adjusted for confounding variables. In order to evaluate a possible mediating effect of the full set of immune markers, we modelled the relationship between asthma and air pollution with a partial least square path model. We observed a strong positive association of IL-1RA [OR 1.37; 95% CI (1.09, 1.73)] with adult-onset asthma. Univariate models did not yield any association between air pollution and immune markers. However, mediation analyses indicated that 15% of the effect of air pollution on risk of adult-onset asthma was mediated through the immune system when considering all immune markers as a latent variable (path coefficient (β) = 0.09; 95% CI: (-0.02, 0.20)). This effect appeared to be stronger for allergic asthma (22%; β = 0.12; 95% CI: (-0.03, 0.27)) and overweight subjects (27%; β = 0.19; 95% CI: (-0.004, 0.38)). Our results provides supportive evidence for a mediating effect of the immune system in the association between air pollution and adult-onset asthma.

Modelling the vertical gradient of nitrogen dioxide in an urban area

INTRODUCTION

Land use regression models environmental predictors to estimate ground-floor air pollution concentration surfaces of a study area. While many cities are expanding vertically, such models typically ignore the vertical dimension.

METHODS

We took integrated measurements of NO₂ at up to three different floors on the facades of 25 buildings in the mid-sized European city of Basel, Switzerland. We quantified the decrease in NO₂ concentration with increasing height at each facade over two 14-day periods in different seasons. Using predictors of traffic load, population density and street configuration, we built conventional land use regression (LUR) models which predicted ground floor concentrations. We further evaluated which predictors best explained the vertical decay rate. Ultimately, we combined ground floor and decay models to explain the measured concentrations at all heights.

RESULTS

We found a clear decrease in mean nitrogen dioxide concentrations between measurements at ground level and those at higher floors for both seasons. The median concentration decrease was 8.1% at 10 m above street level in winter and 10.4% in summer. The decrease with height was sharper at buildings where high concentrations were measured on the ground and in canyon-like street configurations. While the conventional ground floor model was able to explain ground floor concentrations with a model R² of 0.84 (RMSE 4.1 μg/m³), it predicted measured concentrations at all heights with an R² of 0.79 (RMSE 4.5 μg/m³), systematically overpredicting concentrations at higher floors. The LUR model considering vertical decay was able to predict ground floor and higher floor concentrations with a model R² of 0.84 (RMSE 3.8 μg/m³) and without systematic bias.

DISCUSSION

Height above the ground is a relevant determinant of outdoor residential exposure, even in medium-sized European cities without much high-rise. It is likely that conventional LUR models overestimate exposure for residences at higher floors near major roads. This overestimation can be minimized by considering decay with height.

Long-term exposure to transportation noise and its association with adiposity markers and development of obesity

The contribution of different transportation noise sources to metabolic disorders such as obesity remains understudied. We evaluated the associations of long-term exposure to road, railway and aircraft noise with measures of obesity and its subphenotypes using cross-sectional and longitudinal designs. We assessed 3796 participants from the population-based Swiss Cohort Study on Air Pollution and Lung and Heart Diseases (SAPALDIA), who attended the visits in 2001 (SAP2) and 2010/2011 (SAP3) and who were aged 29-72 at SAP2. At SAP2 we measured body mass index (BMI, kg/m²). At SAP3 we measured BMI, waist circumference (centimetres) and Kyle body Fat Index (%) and derived overweight, central and general obesity. Longitudinally for BMI, we derived change in BMI, incidence of overweight and obesity and a 3-category outcome combining the latter two. We assigned source-specific 5-year mean noise levels before visits and during follow-up at the most exposed dwelling façade (Lden, dB), using Swiss noise models for 2001 and 2011 and participants' residential history. Models were adjusted for relevant confounders, including traffic-related air pollution. Exposure to road traffic noise was significantly associated with all adiposity subphenotypes, cross-sectionally (at SAP3) [e.g. beta (95% CI) per 10 dB, BMI: 0.39 (0.18; 0.59); waist circumference: 0.93 (0.37; 1.50)], and with increased risk of obesity, longitudinally (e.g. RR = 1.25, 95% CI: 1.04; 1.51, per 10 dB in 5-year mean). Railway noise was significantly related to increased risk of overweight. In cross-sectional analyses, we further identified a stronger association between road traffic noise and BMI among participants with cardiovascular disease and an association between railway noise and BMI among participants reporting bad sleep. Associations were independent of the other noise sources, air pollution and robust to all adjustment sets. No associations were observed for aircraft noise. Long-term exposure to transportation noise, particularly road traffic noise, may increase the risk of obesity and could constitute a pathway towards cardiometabolic and other diseases.

Perturbation of metabolic pathways mediates the association of air pollutants with asthma and cardiovascular diseases

BACKGROUND

Epidemiologic evidence indicates common risk factors, including air pollution exposure, for respiratory and cardiovascular diseases, suggesting the involvement of common altered molecular pathways.

OBJECTIVES

The goal was to find intermediate metabolites or metabolic pathways that could be associated with both air pollutants and health outcomes ("meeting-in-the-middle"), thus shedding light on mechanisms and reinforcing causality.

METHODS

We applied a statistical approach named 'meet-in-the-middle' to untargeted metabolomics in two independent case-control studies nested in cohorts on adult-onset asthma (AOA) and cardio-cerebrovascular diseases (CCVD). We compared the results to identify both common and disease-specific altered metabolic pathways.

RESULTS

A novel finding was a strong association of AOA with ultrafine particles (UFP; odds ratio 1.80 [1.26, 2.55] per increase by 5000 particles/cm3). Further, we have identified several metabolic pathways that potentially mediate the effect of air pollution on health outcomes. Among those, perturbation of Linoleate metabolism pathway was associated with air pollution exposure, AOA and CCVD.

CONCLUSIONS

Our results suggest common pathway perturbations may occur as a consequence of chronic exposure to air pollution leading to increased risk for both AOA and CCVD.

Development of land-use regression models for fine particles and black carbon in peri-urban South India

Land-use regression (LUR) has been used to model local spatial variability of particulate matter in cities of high-income countries. Performance of LUR models is unknown in less urbanized areas of low-/middle-income countries (LMICs) experiencing complex sources of ambient air pollution and which typically have limited land use data. To address these concerns, we developed LUR models using satellite imagery (e.g., vegetation, urbanicity) and manually-collected data from a comprehensive built-environment survey (e.g., roads, industries, non-residential places) for a peri-urban area outside Hyderabad, India. As part of the CHAI (Cardiovascular Health effects of Air pollution in Telangana, India) project, concentrations of fine particulate matter (PM_2.5) and black carbon were measured over two seasons at 23 sites. Annual mean (sd) was 34.1 (3.2) μg/m³ for PM_2.5 and 2.7 (0.5) μg/m³ for black carbon. The LUR model for annual black carbon explained 78% of total variance and included both local-scale (energy supply places) and regional-scale (roads) predictors. Explained variance was 58% for annual PM_2.5 and the included predictors were only regional (urbanicity, vegetation). During leave-one-out cross-validation and cross-holdout validation, only the black carbon model showed consistent performance. The LUR model for black carbon explained a substantial proportion of the spatial variability that could not be captured by simpler interpolation technique (ordinary kriging). This is the first study to develop a LUR model for ambient concentrations of PM_2.5 and black carbon in a non-urban area of LMICs, supporting the applicability of the LUR approach in such settings. Our results provide insights on the added value of manually-collected built-environment data to improve the performance of LUR models in settings with limited data availability. For both pollutants, LUR models predicted substantial within-village variability, an important feature for future epidemiological studies.

Spatiotemporal land use random forest model for estimating metropolitan NO2 exposure in Japan

Adequate spatial and temporal estimates of NO₂ concentrations are essential for proper prenatal exposure assessment. Here, we develop a spatiotemporal land use random forest (LURF) model of the monthly mean NO₂ over four years in a metropolitan area of Japan. The overall objective is to obtain accurate NO₂ estimates for use in prenatal exposure assessments. We use random forests to convey the non-linear relationship between NO₂ concentrations and predictor variables, and compare the prediction accuracy with that of a linear regression. In addition, we include the distance decay effect of emission sources on NO₂ concentrations for more efficient model construction. The prediction accuracy of the LURF model is evaluated through a leave-one-monitor-out cross validation. We obtain a high R² value of 0.79, which is better than that of the conventional land use regression model using linear regression (R² of 0.73). We also evaluate the LURF model via a temporal and overall cross validation and obtain R² values of 0.84 and 0.92, respectively. We successfully integrate temporal and spatial components into our model, which exhibits higher accuracy than spatial models constructed individually for each month. Our findings illustrate the advantage of using a LURF to model the spatiotemporal variability of NO₂ concentrations.

Combining Measurements from Mobile Monitoring and a Reference Site To Develop Models of Ambient Ultrafine Particle Number Concentration at Residences

Significant spatial and temporal variation in ultrafine particle (UFP; <100 nm in diameter) concentrations creates challenges in developing predictive models for epidemiological investigations. We compared the performance of land-use regression models built by combining mobile and stationary measurements (hybrid model) with a regression model built using mobile measurements only (mobile model) in Chelsea and Boston, MA (USA). In each study area, particle number concentration (PNC; a proxy for UFP) was measured at a stationary reference site and with a mobile laboratory driven along a fixed route during an ∼1-year monitoring period. In comparing PNC measured at 20 residences and PNC estimates from hybrid and mobile models, the hybrid model showed higher Pearson correlations of natural log-transformed PNC ( r = 0.73 vs 0.51 in Chelsea; r = 0.74 vs 0.47 in Boston) and lower root-mean-square error in Chelsea (0.61 vs 0.72) but no benefit in Boston (0.72 vs 0.71). All models overpredicted log-transformed PNC by 3-6% at residences, yet the hybrid model reduced the standard deviation of the residuals by 15% in Chelsea and 31% in Boston with better tracking of overnight decreases in PNC. Overall, the hybrid model considerably outperformed the mobile model and could offer reduced exposure error for UFP epidemiology.

Methods for Assessing Long-Term Exposures to Outdoor Air Pollutants

Purpose of Review

Epidemiological studies of health effects of long-term exposure to outdoor air pollution rely on different exposure assessment methods. This review discusses widely used methods with a special focus on new developments.

Recent Findings

New data and study designs have been applied, including satellite measurements of fine particles and nitrogen dioxide (NO₂). The methods to apply satellite data for epidemiological studies are improving rapidly and have already contributed significantly to national-, continental- and global-scale models. Spatiotemporal models have been developed allowing more detailed temporal resolution compared to spatial models. The development of hybrid models combining dispersion models, satellite observations, land use and surface monitoring has improved models substantially. Mobile monitoring designs to develop models for long-term UFP exposure have been conducted.

Summary

Methods to assess long-term exposure to outdoor air pollution have improved significantly over the past decade. Application of satellite data and mobile monitoring designs is promising new methods.

Robustness of intra urban land-use regression models for ultrafine particles and black carbon based on mobile monitoring

Land-use regression (LUR) models for ultrafine particles (UFP) and Black Carbon (BC) in urban areas have been developed using short-term stationary monitoring or mobile platforms in order to capture the high variability of these pollutants. However, little is known about the comparability of predictions of mobile and short-term stationary models and especially the validity of these models for assessing residential exposures and the robustness of model predictions developed in different campaigns. We used an electric car to collect mobile measurements (n = 5236 unique road segments) and short-term stationary measurements (3 × 30min, n = 240) of UFP and BC in three Dutch cities (Amsterdam, Utrecht, Maastricht) in 2014-2015. Predictions of LUR models based on mobile measurements were compared to (i) measured concentrations at the short-term stationary sites, (ii) LUR model predictions based on short-term stationary measurements at 1500 random addresses in the three cities, (iii) externally obtained home outdoor measurements (3 × 24h samples; n = 42) and (iv) predictions of a LUR model developed based upon a 2013 mobile campaign in two cities (Amsterdam, Rotterdam). Despite the poor model R² of 15%, the ability of mobile UFP models to predict measurements with longer averaging time increased substantially from 36% for short-term stationary measurements to 57% for home outdoor measurements. In contrast, the mobile BC model only predicted 14% of the variation in the short-term stationary sites and also 14% of the home outdoor sites. Models based upon mobile and short-term stationary monitoring provided fairly high correlated predictions of UFP concentrations at 1500 randomly selected addresses in the three Dutch cities (R² = 0.64). We found higher UFP predictions (of about 30%) based on mobile models opposed to short-term model predictions and home outdoor measurements with no clear geospatial patterns. The mobile model for UFP was stable over different settings as the model predicted concentration levels highly correlated to predictions made by a previously developed LUR model with another spatial extent and in a different year at the 1500 random addresses (R² = 0.80). In conclusion, mobile monitoring provided robust LUR models for UFP, valid to use in epidemiological studies.

Characterizing prolonged heat effects on mortality in a sub-tropical high-density city, Hong Kong

Extreme hot weather events are likely to increase under future climate change, and it is exacerbated in urban areas due to the complex urban settings. It causes excess mortality due to prolonged exposure to such extreme heat. However, there is lack of universal definition of prolonged heat or heat wave, which leads to inadequacies of associated risk preparedness. Previous studies focused on estimating temperature-mortality relationship based on temperature thresholds for assessing heat-related health risks but only several studies investigated the association between types of prolonged heat and excess mortality. However, most studies focused on one or a few isolated heat waves, which cannot demonstrate typical scenarios that population has experienced. In addition, there are limited studies on the difference between daytime and nighttime temperature, resulting in insufficiency to conclude the effect of prolonged heat. In sub-tropical high-density cities where prolonged heat is common in summer, it is important to obtain a comprehensive understanding of prolonged heat for a complete assessment of heat-related health risks. In this study, six types of prolonged heat were examined by using a time-stratified analysis. We found that more consecutive hot nights contribute to higher mortality risk while the number of consecutive hot days does not have significant association with excess mortality. For a day after five consecutive hot nights, there were 7.99% [7.64%, 8.35%], 7.74% [6.93%, 8.55%], and 8.14% [7.38%, 8.88%] increases in all-cause, cardiovascular, and respiratory mortality, respectively. Non-consecutive hot days or nights are also found to contribute to short-term mortality risk. For a 7-day-period with at least five non-consecutive hot days and nights, there was 15.61% [14.52%, 16.70%] increase in all-cause mortality at lag 0-1, but only -2.00% [-2.83%, -1.17%] at lag 2-3. Differences in the temperature-mortality relationship caused by hot days and hot nights imply the need to categorize prolonged heat for public health surveillance. Findings also contribute to potential improvement to existing heat-health warning system.

Exposure to Road, Railway, and Aircraft Noise and Arterial Stiffness in the SAPALDIA Study: Annual Average Noise Levels and Temporal Noise Characteristics

BACKGROUND

The impact of different transportation noise sources and noise environments on arterial stiffness remains unknown.

OBJECTIVES

We evaluated the association between residential outdoor exposure to annual average road, railway, and aircraft noise levels, total noise intermittency (IR), and total number of noise events (NE) and brachial-ankle pulse wave velocity (baPWV) following a cross-sectional design.

METHODS

We measured baPWV (meters/second) in 2,775 participants (49-81 y old) at the second follow-up (2010-2011) of the Swiss Cohort Study on Air Pollution and Lung and Heart Diseases in Adults (SAPALDIA). We assigned annual average road, railway, and aircraft noise levels (Ldensource), total day- and nighttime NEtime and IRtime (percent fluctuation=0%, none or constant noise; percent fluctuation=100%, high fluctuation) at the most exposed façade using 2011 Swiss noise models. We applied multivariable linear mixed regression models to analyze associations.

RESULTS

Medians [interquartile ranges (IQRs)] were baPWV=13.4 (3.1) m/s; Ldenair (57.6% exposed)=32.8 (8.0) dB; Ldenrail (44.6% exposed)=30.0 (8.1) dB; Ldenroad (99.7% exposed): 54.2 (10.6) dB; NEnight=123 (179); NEday=433 (870); IRnight=73% (27); and IRday=63.8% (40.3). We observed a 0.87% (95% CI: 0.31, 1.43%) increase in baPWV per IQR of Ldenrail, which was greater with IRnight>80% or with daytime sleepiness. We observed a nonsignificant positive association between Ldenroad and baPWV in urban areas and a negative tendency in rural areas. NEnight, but not NEday, was associated with baPWV. Associations were independent of the other noise sources and air pollution.

CONCLUSIONS

Long-term exposure to railway noise, particularly in an intermittent nighttime noise environment, and to nighttime noise events, mainly related to road noise, may affect arterial stiffness, a major determinant of cardiovascular disease. Ascertaining noise exposure characteristics beyond average noise levels may be relevant to better understand noise-related health effects. https://doi.org/10.1289/EHP1136.

Is physical activity a modifier of the association between air pollution and arterial stiffness in older adults: The SAPALDIA cohort study

INTRODUCTION AND OBJECTIVES

Air pollution and insufficient physical activity have been associated with inflammation and oxidative stress, molecular mechanisms linked to arterial stiffness and cardiovascular disease. There are no studies on how physical activity modifies the association between air pollution and arterial stiffness. We examined whether the adverse cardiovascular effects of air pollution were modified by individual physical activity levels in 2823 adults aged 50-81 years from the well-characterized Swiss Cohort Study on Air Pollution and Lung and Heart Diseases (SAPALDIA).

METHODS

We assessed arterial stiffness as the brachial-ankle pulse wave velocity (baPWV [m/s]) with an oscillometric device. We administered a self-reported physical activity questionnaire to classify each subject's physical activity level. Air pollution exposure was estimated by the annual average individual home outdoor PM₁₀ and PM_2.5 (particulate matter <10μm and <2.5μm in diameter, respectively) and NO₂ (nitrogen dioxide) exposure estimated for the year preceding the survey. Exposure estimates for ultrafine particles calculated as particle number concentration (PNC) and lung deposited surface area (LDSA) were available for a subsample (N=1353). We used mixed effects logistic regression models to regress increased arterial stiffness (baPWV≥14.4m/s) on air pollution exposure and physical activity while adjusting for relevant confounders.

RESULTS

We found evidence that the association of air pollution exposure with baPWV was different between inactive and active participants. The probability of having increased baPWV was significantly higher with higher PM₁₀, PM_2.5, NO₂, PNC and LDSA exposure in inactive, but not in physically active participants. We found some evidence of an interaction between physical activity and ambient air pollution exposure for PM₁₀, PM_2.5 and NO₂ (p_interaction=0.06, 0.09, and 0.04, respectively), but not PNC and LDSA (p_interaction=0.32 and 0.35).

CONCLUSIONS

Our study provides some indication that physical activity may protect against the adverse vascular effects of air pollution in low pollution settings. Additional research in large prospective cohorts is needed to assess whether the observed effect modification translates to high pollution settings in mega-cities of middle and low-income countries.

Land Use Regression Models for Ultrafine Particles in Six European Areas

Long-term ultrafine particle (UFP) exposure estimates at a fine spatial scale are needed for epidemiological studies. Land use regression (LUR) models were developed and evaluated for six European areas based on repeated 30 min monitoring following standardized protocols. In each area; Basel (Switzerland), Heraklion (Greece), Amsterdam, Maastricht, and Utrecht ("The Netherlands"), Norwich (United Kingdom), Sabadell (Spain), and Turin (Italy), 160-240 sites were monitored to develop LUR models by supervised stepwise selection of GIS predictors. For each area and all areas combined, 10 models were developed in stratified random selections of 90% of sites. UFP prediction robustness was evaluated with the intraclass correlation coefficient (ICC) at 31-50 external sites per area. Models from Basel and The Netherlands were validated against repeated 24 h outdoor measurements. Structure and model R² of local models were similar within, but varied between areas (e.g., 38-43% Turin; 25-31% Sabadell). Robustness of predictions within areas was high (ICC 0.73-0.98). External validation R² was 53% in Basel and 50% in The Netherlands. Combined area models were robust (ICC 0.93-1.00) and explained UFP variation almost equally well as local models. In conclusion, robust UFP LUR models could be developed on short-term monitoring, explaining around 50% of spatial variance in longer-term measurements.

Land Use Regression Models for Ultrafine Particles in Six European Areas

Long-term ultrafine particle (UFP) exposure estimates at a fine spatial scale are needed for epidemiological studies. Land use regression (LUR) models were developed and evaluated for six European areas based on repeated 30 min monitoring following standardized protocols. In each area; Basel (Switzerland), Heraklion (Greece), Amsterdam, Maastricht, and Utrecht ("The Netherlands"), Norwich (United Kingdom), Sabadell (Spain), and Turin (Italy), 160-240 sites were monitored to develop LUR models by supervised stepwise selection of GIS predictors. For each area and all areas combined, 10 models were developed in stratified random selections of 90% of sites. UFP prediction robustness was evaluated with the intraclass correlation coefficient (ICC) at 31-50 external sites per area. Models from Basel and The Netherlands were validated against repeated 24 h outdoor measurements. Structure and model R² of local models were similar within, but varied between areas (e.g., 38-43% Turin; 25-31% Sabadell). Robustness of predictions within areas was high (ICC 0.73-0.98). External validation R² was 53% in Basel and 50% in The Netherlands. Combined area models were robust (ICC 0.93-1.00) and explained UFP variation almost equally well as local models. In conclusion, robust UFP LUR models could be developed on short-term monitoring, explaining around 50% of spatial variance in longer-term measurements.

Land use regression modeling of ultrafine particles, ozone, nitrogen oxides and markers of particulate matter pollution in Augsburg, Germany

Important health relevance has been suggested for ultrafine particles (UFP) and ozone, but studies on long-term effects are scarce, mainly due to the lack of appropriate spatial exposure models. We designed a measurement campaign to develop land use regression (LUR) models to predict the spatial variability focusing on particle number concentration (PNC) as indicator for UFP, ozone and several other air pollutants in the Augsburg region, Southern Germany. Three bi-weekly measurements of PNC, ozone, particulate matter (PM₁₀, PM_2.5), soot (PM_2.5abs) and nitrogen oxides (NO_x, NO₂) were performed at 20 sites in 2014/15. Annual average concentration were calculated and temporally adjusted by measurements from a continuous background station. As geographic predictors we offered several traffic and land use variables, altitude, population and building density. Models were validated using leave-one-out cross-validation. Adjusted model explained variance (R²) was high for PNC and ozone (0.89 and 0.88). Cross-validation adjusted R² was slightly lower (0.82 and 0.81) but still indicated a very good fit. LUR models for other pollutants performed well with adjusted R² between 0.68 (PM_coarse) and 0.94 (NO₂). Contrary to previous studies, ozone showed a moderate correlation with NO₂ (Pearson's r=-0.26). PNC was moderately correlated with ozone and PM_2.5, but highly correlated with NO_x (r=0.91). For PNC and NO_x, LUR models comprised similar predictors and future epidemiological analyses evaluating health effects need to consider these similarities.

Combining land use regression models and fixed site monitoring to reconstruct spatiotemporal variability of NO2 concentrations over a wide geographical area

The epidemiological research benefits from an accurate characterization of both spatial and temporal variability of exposure to air pollution. This work aims at proposing a method to combine the high spatial resolution of Land Use Regression (LUR) models with the high temporal resolution of fixed site monitoring data, to model spatiotemporal variability of NO₂ over a wide geographical area in Northern Italy. We developed seasonal LUR models to reconstruct the spatial distribution of a scaling factor that relates local concentrations to those measured at two reference central sites, one for the northern flat area and one for the southern mountain area. We calculated the daily average concentrations at 19 locations spread over the study areas as the product of the local scaling factor and the reference central site concentrations. We evaluated model performance comparing modeled and measured NO₂ data. LUR model's R² ranges from 0.76 to 0.92. The main predictors refers substantially to traffic, industrial land use, buildings volume and altitude a.s.l. The model's performance in reproducing measured concentrations was satisfactory. The temporal variability of concentrations was well captured: Spearman correlation between model and measures was >0.7 for almost all sites. Model's average absolute errors were in the order of 10μgm^-3. The model for the southern area tends to overestimate measured concentrations. Our modeling framework was able to reproduce spatiotemporal differences in NO₂ concentrations. This kind of model is less data-intensive than usual regional atmospheric models and it may be very helpful to assess population exposure within studies in which individual relevant exposure occurs along periods of days or months.