Air Pollution Exposure Estimates Research Articles

Predicting real-time within-vehicle air pollution exposure with mass-balance and machine learning approaches using on-road and air quality data

Modelling the air pollutant concentrations within-vehicles is an essential step to estimate our daily exposure to air pollution. This is a challenging issue however, since the processes that affect the exposures within-vehicles change with different driving patterns and ventilation settings. This study introduces an innovative approach that combines mass-balance principles and machine learning techniques, leveraging ambient air quality, on-road and within-vehicle measurements of particulate matter (PM10, PM2.5, PM1), nitrogen dioxide (NO2), nitrogen oxides (NOx), aerosol lung surface deposited area (LSDA) and ultrafine particles (UFP) under different ventilation settings to estimate air pollution exposure levels within vehicles. The first model (MB) includes basic physical and chemical processes and follows a mass-balance approach to estimate the within-vehicle concentrations. The second model (ML) applies data driven machine learning algorithms to a training set of observations to predict unseen within-vehicle concentrations. By using a number generator, the whole observational dataset was divided to 80:20 and 80% was used to build and train the ML model, while 20% was used for validation. Both models demonstrated good predictions of observations apart from an underestimation in UFP and LSDA. The ML model showed better predictive power than the MB model and had skill in predicting the unseen within-vehicle exposures. The ML model predictions were as good as the MB model for most of the species and improved for NO2. The ML model demonstrated good index of agreement (IOA >0.69) and Pearson correlation coefficient (r > 0.80) for all the species. The inclusion of air quality data from nearby monitoring stations instead of on-road (sampled while driving), in the ML model showed promising and new capabilities to within-vehicle exposure predictions. In an era where air pollution is a growing concern, understanding and predicting within-vehicle air pollution exposure is of great importance for public health and environmental research. This research not only advances the field of exposure assessment but (at no extra cost) also demonstrates practical implications for real-time exposure mapping and health impact assessment of vehicle occupants with existing infrastructure.

Geophysical Uncertainties in Air Pollution Exposure and Benefits of Emissions Reductions for Global Health

AbstractExposure to fine particulate matter (PM2.5) air pollution is associated with large‐scale health consequences, but the ranges in estimates of global air pollution exposure and PM2.5‐related global premature mortality remain understudied. Using four model/observation‐based PM2.5 data sets and six Coupled Model Intercomparison Project Phase 6 (CMIP6) climate models, we compare ranges in current PM2.5‐related mortality estimates based on model/observation combinations to the range of projected emissions reductions and the resulting impacts on global health from CMIP6 models. Although estimates of current mortality are sensitive to the PM2.5 data set (6.54–9.23 million/year using the central estimate from the Global Exposure Mortality Model), the projected near‐term benefits of emissions reductions for reduced mortality are much more certain. Specifically, until the middle of the century, uncertainty ranges in projected avoided deaths are less than the magnitude of ranges in present‐day mortality estimates, even when projected future changing socio‐demographic factors are considered. Under a low‐emissions scenario, avoided cumulative early deaths could reach 275 million (range: 243–351 million) by 2100, assuming no future socio‐demographic changes. Avoided cumulative deaths could reach 124 million (range: 74–246 million) if changing future socio‐demographic factors are considered. As the global population ages and becomes more sensitive to breathing polluted air, there will be an increasing need for policies that both reduce air pollution and improve public health.

Air pollution exposure is associated with rhinitis in older US adults via specific immune mechanisms.

Pathophysiology of rhinitis in older adults is largely unknown. We tested whether air pollution is associated with this condition and how immune mechanisms may play a role in this relationship. We analyzed cross-sectional data from the National Social Life, Health, and Aging Project, a nationally representative study of older adults born between 1920 and 1947. Particulate matter ≤2.5μm (PM2.5 ) air pollution exposure estimates were generated using validated spatiotemporal models. Presence of rhinitis was defined based on medication use (≥1: intranasal medications: steroids, antihistamines, lubricants, and/or decongestants, and/or oral medications: antihistamines and/or decongestants). K-means cluster analysis (Jaccard method) was used to group 13 peripheral blood cytokines into 3 clusters to facilitate functional determination. We fitted multivariate logistic regressions to correlate PM2.5 exposure with presence of rhinitis, controlling for confounders, and then determined the role of cytokines in this relationship. Long- (but not short-) term exposure to PM2.5 was associated with presence of rhinitis: 3-year exposure window, odds ratio (OR)=1.32, 95% confidence interval (CI): 0.98, 1.80, per 1 standard deviation (SD) PM2.5 increase. Inclusion of cytokine cluster in the model led to a modestly stronger effect of PM2.5 exposure on rhinitis (OR=1.37; 95% CI: 1.00, 1.87; 3-year exposure window). The particular immune profile responsible for this result was composed of elevated IL-3, IL-12, and IFN-γ (OR=4.86, 95% CI: 1.10, 21.58, immune profile-PM2.5 exposure interaction term). We show for the first time that IL-3, IL-12, and IFN-γ explain in part the relationship between PM2.5 exposure and rhinitis in older US adults. If confirmed, these immune pathways may be used as therapeutic targets.

How Differential Privacy Will Affect Estimates of Air Pollution Exposure and Disparities in the United States

Census data is crucial to understand energy and environmental justice outcomes such as poor air quality which disproportionately impact people of color in the U.S. Wwith the advent of sophisticated personal datasets and analysis, Census Bureau is considering adding top-down noise (differential privacy) and post-processing 2020 census data to reduce the risk of identification of individual respondents. Using 2010 demonstration census and pollution data, I find that compared to the original census, differentially private (DP) census significantly changes ambient pollution exposure in areas with sparse populations. White Americans have lowest variability, followed by Latinos, Asian, and Black Americans. DP underestimates pollution disparities for SO2 and PM2.5 while overestimates the pollution disparities for PM10.

Impacts of distinct travel behaviors on potential air pollution exposure measurement error

Prior studies typically estimate air pollution exposure at places of residence only; however, people are likely to conduct daily activities at out-of-home locations. Overlooking human mobility in exposure estimation can introduce measurement error in health analyses and may lead to ineffective public policies. We simulated travel mobility patterns for 100,784 residents of Los Angeles County and their exposure to outdoor PM2.5 on a typical weekday. We used K-means clustering via principal components analysis (PCA) to subdivide subjects into seven travel behavior clusters and compared their mobility-based dynamic PM2.5 exposures to their residence-based static exposure using our previously developed spatiotemporal model. Travel behavior correlated with exposure to outdoor PM2.5 and with the degree of PM2.5 exposure measurement error. Workers with long commuting distance (and highest income) had the lowest dynamic PM2.5 exposure (8.1 μg/m3), while workers who relied on public transit (and were generally lower income) had the highest exposure (8.8 μg/m3). In general, lower income individuals lived in more polluted neighborhoods and took longer duration (but shorter distance) commutes to work locations with similarly elevated PM2.5 concentrations, while higher income individuals took longer distance commutes (similar duration) from less polluted residential to more polluted occupational neighborhoods. “Worker” and frequent traveler clusters with highest mobility had the highest exposure measurement error, and the non-worker “stay-at-home” cluster and student clusters had the lowest. The extent of exposure measurement error generally correlated with pollution levels at the residential location and distance (less so duration) of travel away from the residential neighborhood. Our results contribute to the literature on effects of travel behavior on exposure measurement error. The generated activity clusters combined with sociodemographic information provide additional information for researchers to understand which groups of people may be more challenging to model or predict their exposures accurately.

Ethnic disparities in ambient air and traffic-related pollution exposure and ethnic-specific impacts on clinical biomarker levels

BackgroundAlthough characterizing the inequality in pollution exposure burden across ethnic groups and the ethnic-specific exposure associations is of great social and public health importance, it has not been systematically investigated in large population studies. MethodsThe UK Biobank data (N = 485, 806) of individual-level air ambient and traffic-related pollution exposure, biomarkers routinely used in clinical practice, genotype, life-style factors, and socioeconomic status were analyzed. Air pollution exposure estimates were compared among six genetically inferred ethnic groups. We also quantified the association between exposure and biomarkers within and across ethnicities. ResultsNon-European participants (defined by genetics) disproportionately bear a higher burden of exposure than their European counterparts even after adjusting for covariables including socioeconomic status. For example, exposure to NO2 in people with African ancestry was 30.7 % higher (p = 1.5E-786) than European subjects. Within the genetically defined African group, larger African genetic ancestry proportion (AGAP) was linked to higher ambient air pollutant exposure. Trans-Ethnic analysis identified 32 clinical biomarkers associated with environmental exposure. For 13 biomarkers, the association with exposure was significantly different or even in opposing directions across ethnic groups. ConclusionsSubstantial disparities in air pollution exposure was observed among genetically-defined ethnic groups. Most importantly, we show that the impact of exposure on biomarkers varies by ethnicity. Reducing the disproportionally high exposure burden on non-European populations and alleviating the adverse consequences in an ethnic-specific manner are of great urgency and significance.

Modeling county-level benzene emissions using transportation analysis zones in the Denver metro area

The link between exposure to benzene (an important toxic air pollutant) and physical health is well-documented as benzene is a known carcinogen. Understanding benzene concentration distributions in urban areas is very important and can be done using air dispersion models. An essential input requirement for dispersion models is the spatial and temporal allocation of emissions. Since most county-level benzene emissions come from gasoline powered motor vehicles, the goal of this work was to develop a new method of county-level benzene emission allocation using transportation analysis zones (TAZs) commonly used in travel demand models. The high spatial resolution provided by the TAZ allocation method helped visualize different flux patterns generated from five emission sectors and identified where more effort is needed to reduce benzene concentrations. The new allocation method was also tested using the AERMOD dispersion model to estimate benzene concentrations in the Denver metropolitan area. The model performance was assessed using 2014 morning benzene measurements from a monitoring station located in downtown Denver; results showed an acceptable model-to-monitor ratio. The model was also used to investigate the effect of temperature inversions on morning benzene concentrations. As expected, predicted benzene concentrations were higher during temperature inversions and model-to-monitor ratios during investigated days were within a factor of two. The TAZ allocation modeling methodology demonstrated in this work can improve the estimation of air pollution exposure in future health-related studies. • County-level benzene emissions were allocated based on transportation analysis zones. • This new allocation method identified on-road, non-road, and oil and gas as major emission sources. • Urban benzene concentrations were reasonably predicted using the updated emission allocations and a dispersion model. • Model-predicted and measured benzene concentrations were elevated during morning temperature inversions.

Long-term Air Pollution Exposure on COVID-19 Risk and Severity

Since the emergence of the COVID-19 pandemic, long-term exposure to air pollution has emerged as a possible risk factor for increased COVID-19 disease risk, severity, and mortality. The objective of this study was to examine the association between COVID-19 and long-term exposure to air pollution. This retrospective cohort study included patients who had a COVID-19 test at the University of Chicago Medicine Center from March 18, 2020, to March 30, 2021. Air pollution exposure was assessed using models developed for Chicago by the MESA Air Study and data from the Environmental Protection Agency’s EJSCREEN. Air pollution exposure estimates were linked to the University of Chicago Medicine patients using residential addresses. Air pollution measurements included PM2.5 (MESA Air and EPA), NO2, diesel PM, the respiratory hazard index, traffic proximity, and ozone. The outcome was COVID-19 risk and severity (ICU admission). Mixed effect analyses were performed to test whether poor air quality status was associated with COVID-19 risk and severity, controlling for demographic, smoking, and comorbidity indicators collected from the medical records. A total of 74,004 patients (mean [SD] age 44.0 [22.51] years; 42,979 [58.09%] women; and 32,870 [49.09%] African Americans) were linked to air quality exposure based on residence and other covariates using electronic medical record data in addition to COVID-19 test results. Overall, 8,463 participants (11.4%) tested positive for COVID-19. In the mixed effect analysis, long-term PM2.5 exposure was associated with increased risk of testing positive for COVID-19 (odds ratio [OR], 1.30; 95% CI, 1.23-1.38; P <0.01) and increased severity from COVID-19 (OR= 2.28; 95% CI, 2.00-2.60; P = <0.01). The results of this retrospective cohort study provide some evidence of the effect of long-term air pollution exposure on COVID-19 risk and severity.

Comparison of PM2.5 air pollution exposure estimates at WHIMS participant locations based on different modelling approaches and impact on health effects estimation

Background and Aim: Many approaches to quantifying air pollution exposures for use in epidemiologic analyses have been developed. However, whether exposure estimates from different approaches are comparable and whether choice of approach affects reported associations with health-related outcomes remains unclear. We compare PM2.5 concentrations at addresses from participants of the Women’s Health Initiative Memory Study (WHIMS) derived using multiple estimation approaches and assess differences in associated health effects estimates. Methods: We linked annual PM2.5 exposure estimates from 1999 to 2004 from 12 different approaches to WHIMS participant addresses from across the continental United States and restricted analyses to data from participants with stable addresses within each calendar year. Approaches included geo-statistical interpolation approaches, geographic information system (GIS)-based statistical models, air dispersion and chemical transport models, and hybrid models. We used descriptive statistics and visualizations to assess relative and absolute agreement of exposure estimates from different approaches across space and time and examined the impact of estimation approach on associations between PM2.5 and non-accidental mortality, CVD-related mortality, and incident CVD events. Results: Annual PM2.5 exposure estimates were reasonably similar across approaches, although the specific shapes and ranges of the distributions varied. Interquartile ranges tended to be similar, though ranges were more variable. Relative agreement was consistently higher than absolute agreement. After controlling for putative confounders, associations of PM2.5 with mortality and CVD were similar regardless of estimation approach. Conclusions: Agreement across different air pollution estimation approaches for PM2.5 was generally high when considering participants drawn from a broad geographic area, suggesting that all models predicted similar between-area differences. The implications of fine-scale variability between exposure estimation approaches warrants further investigation. Keywords: Air pollution, exposure assessment

Europe-wide air pollution modelling from 2000 to 2019 using geographically weighted regression

BACKGROUND AND AIM Land-use regression (LUR) empirically relates air pollution concentrations to predictors such as traffic data, population, and land use data. Previously used regression methods in Europe wide models assumed spatially-fixed linear relationships between predictors and air pollution concentrations. Thus, to include spatial heterogeneity in relationships, we used geographically weighted regression (GWR) and the machine learning method Random Forest (RF), with the aim to further improve the accuracy of air pollution exposure estimates. METHODS We built separate LUR models for each year from 2000 to 2019 for NO2, O3, PM2.5 and PM10 using annual average concentrations from routine monitoring stations across Europe. The potential predictors included satellite retrievals, chemical transport model estimates and land-use variables. Supervised linear regression (SLR) was first used to select predictors, and then GWR estimated the potentially spatially-varying regression coefficients. We evaluated the model performance using five-fold cross-validation (CV) for each year and compared GWR with SLR and RF. RESULTS GWR explained measured concentrations generally well, with annual CV-R2 values of 0.62-0.67, 0.43-0.66, 0.48-0.71, 0.69-0.82, for NO2, O3, PM10, and PM2.5 respectively. GWR improved the R2 values compared to SLR by 4-7%, 7-15%, 7-14%, 0-6% and compared to RF by 1-4%, 0-10%, 0-11%, 7-19%, for NO2, O3, PM10, and PM2.5 respectively. Predictors selected and the spatially-varying coefficient values varied between years. CONCLUSIONS Including spatial heterogeneity using geographically weighted regression improved European air pollution LUR models. These models allow time-varying exposure-risk models. KEYWORDS Geographically weighted regression, spatial heterogeneity, land-use regression, air pollution, supervised linear regression, random forests

Functional Kriging for Spatiotemporal Modeling of Nitrogen Dioxide in a Middle Eastern Megacity

Long-term hour-specific air pollution exposure estimates have rarely been of interest in epidemiological research. However, this can be relevant for studies that aim to estimate the residential exposure for the hours that subjects mostly spend time there, or for those hours that they may work in another location. Here, we developed a model by spatially predicting the long-term diurnal curves of nitrogen dioxide (NO2) in Tehran, Iran, one of the most polluted and populated megacities in the Middle East. We used the statistical framework of functional data analysis (FDA) including ordinary kriging for functional data (OKFD) and functional analysis of variance (fANOVA) for modeling. The long-term NO2 diurnal curves had two distinct maxima and minima. The absolute minimum value of the city average was 40.6 ppb (around 4:00 p.m.) and the absolute maximum value was 52.0 ppb (around 10:00 p.m.). The OKFD showed the concentrations, the diurnal maximum/minimum values, and their corresponding occurring times varied across the city. The fANOVA highlighted that the effect of population density on the NO2 concentrations is not constant and depends on time within the diurnal period. The provided estimation of long-term hour-specific maps can inform future epidemiological studies to use the long-term mean for specific hour(s) of the day. Moreover, the demonstrated FDA framework can be used as a set of flexible statistical methods.

Mapping Mobility: Utilizing Local-Knowledge-Derived Activity Space to Estimate Exposure to Ambient Air Pollution among Individuals Experiencing Unsheltered Homelessness.

Individuals experiencing homelessness represent a growing population in the United States. Air pollution exposure among individuals experiencing homelessness has not been quantified. Utilizing local knowledge mapping, we generated activity spaces for 62 individuals experiencing homelessness residing in a semi-rural county within the United States. Satellite derived measurements of fine particulate matter (PM2.5) were utilized to estimate annual exposure to air pollution experienced by our participants, as well as differences in the variation in estimated PM2.5 at the local scale compared with stationary monitor data and point location estimates for the same period. Spatial variation in exposure to PM2.5 was detected between participants at both the point and activity space level. Among all participants, annual median PM2.5 exposure was 16.22 μg/m3, exceeding the National Air Quality Standard. Local knowledge mapping represents a novel mechanism to capture mobility patterns and investigate exposure to air pollution within vulnerable populations. Reliance on stationary monitor data to estimate air pollution exposure may lead to exposure misclassification, particularly in rural and semirural regions where monitoring is limited.

Exposure to Elevated Nitrogen Dioxide Concentrations and Cardiac Remodeling in Patients With Dilated Cardiomyopathy

BackgroundEmpirical evidence suggests a strong link between exposure to air pollution and heart failure incidence, hospitalizations, and mortality, but the biological basis of this remains unclear. We sought to determine the relationship between differential air pollution levels and changes in cardiac structure and function in patients with dilated cardiomyopathy. Methods and ResultsWe undertook a prospective longitudinal observational cohort study of patients in England with dilated cardiomyopathy (enrollment 2009–2015, n = 716, 66% male, 85% Caucasian) and conducted cross sectional analysis at the time of study enrollment. Annual average air pollution exposure estimates for nitrogen dioxide (NO2) and particulate matter with diameter of 2.5 µm or less (PM2.5) at enrolment were assigned to each residential postcode (on average 12 households). The relationship between air pollution and cardiac morphology was assessed using linear regression modelling. Greater ambient exposure to NO2 was associated with higher indexed left ventricular (LV) mass (4.3 g/m2 increase per interquartile range increase in NO2, 95% confidence interval 1.9–7.0 g/m2) and lower LV ejection fraction (–1.5% decrease per interquartile range increase in NO2, 95% confidence interval –2.7% to –0.2%), independent of age, sex, socioeconomic status, and clinical covariates. The associations were robust to adjustment for smoking status and geographical clustering by postcode area. The effect of air pollution on LV mass was greatest in women. These effects were specific to NO2 exposure. ConclusionsExposure to air pollution is associated with raised LV mass and lower LV ejection fraction, with the strongest effect in women. Although epidemiological associations between air pollution and heart failure have been established and supported by preclinical studies, our findings provide novel empirical evidence of cardiac remodeling and exposure to air pollution with important clinical and public health implications.

Evaluating the Utility of High-Resolution Spatiotemporal Air Pollution Data in Estimating Local PM2.5 Exposures in California from 2015–2018

Air quality management is increasingly focused not only on across-the-board reductions in ambient pollution concentrations but also on identifying and remediating elevated exposures that often occur in traditionally disadvantaged communities. Remote sensing of ambient air pollution using data derived from satellites has the potential to better inform management decisions that address environmental disparities by providing increased spatial coverage, at high-spatial resolutions, compared to air pollution exposure estimates based on ground-based monitors alone. Daily PM2.5 estimates for 2015–2018 were estimated at a 1 km2 resolution, derived from NASA’s Moderate Resolution Imaging Spectroradiometer (MODIS) satellite instrument and the Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm in order to assess the utility of highly refined spatiotemporal air pollution data in 92 California cities and in the 13 communities included in the California Community Air Protection Program. The identification of pollution hot-spots within a city is typically not possible relying solely on the regulatory monitoring networks; however, day-to-day temporal variability was shown to be generally well represented by nearby ground-based monitoring data even in communities with strong spatial gradients in pollutant concentrations. An assessment of within-ZIP Code variability in pollution estimates indicates that high-resolution pollution estimates (i.e., 1 km2) are not always needed to identify spatial differences in exposure but become increasingly important for larger geographic areas (approximately 50 km2). Taken together, these findings can help inform strategies for use of remote sensing data for air quality management including the screening of locations with air pollution exposures that are not well represented by existing ground-based air pollution monitors.

‘Real-time’ air quality channels: A technology review of emerging environmental alert systems

Poor air quality is a pressing global challenge contributing to adverse health impacts around the world. In the past decade, there has been a rapid proliferation of air quality information delivered via sensors, apps, websites or other media channels in near real-time and at increasingly localized geographic scales. This paper explores the growing emphasis on self-monitoring and digital platforms to supply informational interventions for reducing pollution exposures and improving health outcomes at the individual level. It presents a technological case study that characterizes emerging air quality information communication mechanisms, or ‘AQ channels’, while drawing upon examples throughout the literature. The questions are posed: which air quality channels are ‘freely’ available to individuals in London, UK, and when and where are they accessed? Digital trace data and metadata associated with 54 air quality channels are synthesized narratively and graphically. Results reveal air quality channels derive air pollution estimates using common data sources, display disparate messaging, adopt variable geographic scales for reporting ‘readings’ and maintain psychosocial barriers to access and adoption of exposure-reducing behaviours. The results also point to a clear association between the publication of a high-profile news article about air pollution and increased air quality channel access. These findings illuminate a need for greater transparency around how air quality channels generate personalized air pollution exposure estimates and tailor messaging. The paper concludes by calling for air quality channel developers to exercise co-creative methods that can support sustainable, democratic data and knowledge production around air quality, while critically approaching disproportionate patterns of both pollution and information exposure.

A novel multi-pollutant space-time learning network for air pollution inference

Detailed information about air pollution in space and time is essential to manage risks to public health. In this paper we propose a multi-pollutant space-time learning network (Multi-AP learning network), which estimates pixel-wise (grid-level) concentrations of multiple air pollutant species based on fixed-station measurements and multi-source urban features, including land use information, traffic data, and meteorological conditions. We infer concentrations of multiple pollutants within one integrated learning network, which is applied to and evaluated on a case study in Chengdu (4900 km2, 26 April - 12 June 2019), where air pollutant (PM2.5, PM10 and O3) measurements from 40 monitoring sites are used to train the network to estimate pollutant concentrations in 4900 grid-cells (1 km2). The Multi-AP learning network allows us to estimate highly-resolved (1 km × 1 km, hourly) air pollution maps based on pollutant measurements which cover less than 1% of the grid-cells with better accuracy compared to other approaches, and with significant computational efficiency improvements. The time-cost is 1/3 of the time-cost of modelling each pollutant individually. Furthermore, we evaluate the relative importance of features and find that the meteorological feature set is the most important followed the land use features. The proposed Multi-AP method could be used to estimate air pollution exposure across a city using a limited set of air pollution monitoring sites.

Development of high-resolution spatio-temporal models for ambient air pollution in a metropolitan area of China from 2013 to 2019

Modeling high-resolution air pollution concentrations is essential to accurately assess exposure for population studies. The aim of this study is to establish an advanced exposure model to predict spatiotemporal changes in fine particulate matter (PM2.5), nitrogen dioxides (NO2), and ozone (O3) concentrations in Shanghai, China. The model is constructed on a geo-statistical modeling framework that incorporates a dimension reduction regression approach and a spatial smoothing function to deal with fine-scale exposure variations. We used a dataset with comprehensive observational and predictor variables that included monitoring data from both national and local agencies from 2013 to 2019, a high-resolution geographical dataset of predictor variables, and a full-coverage weekly satellite data of the aerosol optical depth at a 1 × 1 km2 resolution. Our model performed well in terms of the spatial and temporal prediction ability assessed by cross-validation (CV) for PM2.5 (spatial R2 = 0.89, temporal R2 = 0.91), NO2 (R2 = 0.49, 0.78), and O3 (R2 = 0.67, 0.81) at the national monitors over seven years according to the leave-one-out CV. For the predictions at the local agency monitoring stations, the overall CV R2 was between 0.77 and 0.89 across the air pollutants. We visualized the long-term and seasonal averaged predictions of the PM2.5, NO2, and O3 exposure on maps with a spatial resolution of 100 × 100 m2. Our study provides a useful tool to accurately estimate air pollution exposure with high spatial and temporal resolution at the urban scale. These model predictions will be useful to assess both short-term and long-term air pollution exposure for health studies.

The UrbEm Hybrid Method to Derive High-Resolution Emissions for City-Scale Air Quality Modeling

As cities are growing in size and complexity, the estimation of air pollution exposure requires a detailed spatial representation of air pollution levels, rather than homogenous fields, provided by global- or regional-scale models. A critical input for city-scale modeling is a timely and spatially resolved emission inventory. Bottom–up approaches to create urban-scale emission inventories can be a demanding and time-consuming task, whereas local emission rates derived from a top–down approach may lack accuracy. In the frame of this study, the UrbEm approach of downscaling gridded emission inventories is developed, investing upon existing, open access, and credible emission data sources. As a proof-of-concept, the regional anthropogenic emissions by Copernicus Atmospheric Monitoring Service (CAMS) are handled with a top–down approach, creating an added-value product of anthropogenic emissions of trace gases and particulate matter for any city (or area) of Europe, at the desired spatial resolution down to 1 km. The disaggregation is based on contemporary proxies for the European area (e.g., Global Human Settlement population data, Urban Atlas 2012, Corine, OpenStreetMap data). The UrbEm approach is realized as a fully automated software tool to produce a detailed mapping of industrial (point), (road-) transport (line), and residential/agricultural/other (area) emission sources. Line sources are of particular value for air quality studies at the urban scale, as they enable explicit treatment of line sources by models capturing among others the street canyon effect and offer an overall better representation of the critical road transport sector. The UrbEm approach is an efficient solution for such studies and constitutes a fully credible option in case high-resolution emission inventories do not exist for a city (or area) of interest. The validity of UrbEm is examined through the evaluation of high-resolution air pollution predictions over Athens and Hamburg against in situ measurements. In addition to a better spatial representation of emission sources and especially hotspots, the air quality modeling results show that UrbEm outputs, when compared to a uniform spatial disaggregation, have an impact on NO2 predictions up to 70% for urban regions with complex topographies, which corresponds to a big improvement of model accuracy (FAC2 > 0.5), especially at the source-impacted sites.

Using Satellite Data for Environmental Justice Efforts in California

BACKGROUND AND AIM: Air quality management is increasingly focused not only on across-the-board reductions in ambient pollution concentrations, but also on identifying and remediating elevated exposures that often occur in traditionally disadvantaged communities. Remote sensing of ambient air pollution using data derived from satellites has the potential to better inform management decisions that address environmental disparities by providing increased spatial coverage, at high spatial resolutions, compared to air pollution exposure estimates based on ground monitors alone. METHODS: Daily PM₂.₅ estimates for 2015-2018 were estimated at a 1 km² resolution, derived from NASA’s Moderate Resolution Imaging Spectroradiometer (MODIS) satellite instrument and the Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm, and compared to federal ground monitors in order to assess the utility of highly-refined spatiotemporal air pollution data in the 13 communities included in the California Community Air Protection Program. RESULTS:Results suggest that regulatory monitoring networks are unable to identify pollution hotspots within these communities without the addition of high-resolution satellite data. However, day-to-day temporal variability is generally well represented by nearby ground-based monitoring data even in communities with strong spatial gradients in pollutant concentrations. CONCLUSIONS:These findings can help inform strategies for use of remote sensing data in environmental justice efforts including the screening of locations with air pollution exposures that are not well-represented by existing ground-based air pollution monitors. KEYWORDS: Air Pollution, Environmental Disparities, Environmental Justice, Exposure Assessment, Particulate Matter

A land use regression model using machine learning and locally developed low cost particulate matter sensors in Uganda

The application of land use regression (LUR) modeling for estimating air pollution exposure has been used only rarely in sub-Saharan Africa (SSA). This is generally due to a lack of air quality monitoring networks in the region. Low cost air quality sensors developed locally in sub-Saharan Africa presents a sustainable operating mechanism that may help generate the air monitoring data needed for exposure estimation of air pollution with LUR models. The primary objective of our study is to investigate whether a network of locally developed low-cost air quality sensors can be used in LUR modeling for accurately predicting monthly ambient fine particulate matter (PM2.5) air pollution in urban areas of central and eastern Uganda. Secondarily, we aimed to explore whether the application of machine learning (ML) can improve LUR predictions compared to ordinary least squares (OLS) regression. We used data for the entire year of 2020 from a network of 23 PM2.5 low-cost sensors located in urban municipalities of eastern and central Uganda. Between January 1, 2020 and December 31, 2020, these sensors collected highly time-resolved measurement data of PM2.5 air concentrations. We used monthly-averaged PM2.5 concentration data for LUR prediction modeling of monthly PM2.5 concentrations. We used eight different ML base-learner algorithms as well as ensemble modeling. We applied 5-fold cross validation (80% training/20% test random splits) to evaluate the models with resampling and Root mean squared error (RMSE). The relative explanatory power and accuracy of the ML algorithms were evaluated by comparing coefficient of determination (R2) and RMSE, using OLS as the reference approach. The overall average PM2.5 concentration during the study period was 52.22 μg/m3 (IQR: 38.11, 62.84 μg/m3)—well above World Health Organization PM2.5 ambient air guidelines. From the base-learner and ensemble models, RMSE and R2 values ranged between 7.65 μg/m3 – 16.85 μg/m3 and 0.24–0.84, respectively. Extreme gradient boosting (xgbTree) performed best out of the base learner algorithms (R2 = 0.84; RMSE = 7.65 μg/m3). Model performance from ensemble modeling with Lasso and Elastic-Net Regularized Generalized Linear Models (glmnet) did not outperform xgbTree, but prediction performance was comparable to that of xgbTree. The most important temporal and spatial predictors of monthly PM2.5 levels were monthly precipitation, percent of the population using solid fuels for cooking, distance to Lake Victoria, and greenspace (NDVI) within a 500-m buffer of air monitors. In conclusion, data from locally developed low-cost PM sensors provide evidence that they can be used for spatio-temporal prediction modeling of air pollution exposures in Uganda. Moreover, the non-parametric ML and ensemble approaches to LUR modeling clearly outperformed OLS regression algorithm for the prediction of monthly PM2.5 concentrations. Deploying low-cost air quality sensors in concert with implementation of data quality control measures, can help address the critical need for expanding and improving air quality monitoring in resource-constrained settings of sub-Saharan Africa. These low-cost sensors, in conjunction with non-parametric ML algorithms, may provide a rapid path forward for PM2.5 exposure assessment and to spur air pollution epidemiology research in the region.