Search

Acute lymphoblastic leukemia (ALL) is the most common type of cancer in children (age 0-14 years); however, the etiology remains incompletely understood. Several environmental exposures have been linked to risk of childhood ALL, including air pollution. Closely related to air pollution and human development is artificial light at night (ALAN), which is believed to disrupt circadian rhythm and impact health. We sought to evaluate outdoor ALAN and air pollution on risk of childhood ALL. The California Linkage Study of Early-Onset Cancers is a large population-based case-control in California that identifies and links cancer diagnoses from the California Cancer Registry to birth records. For each case, 50 controls with the same year of birth were obtained from birth records. A total of 2,782 ALL cases and 139,100 controls were identified during 2000-2015. ALAN was assessed with the New World Atlas of Artificial Night Sky Brightness and air pollution with an ensemble-based air pollution model of particulate matter smaller than 2.5 microns (PM 2.5 ). After adjusting for known and suspected risk factors, the highest tertile of ALAN was associated with an increased risk of ALL in Hispanic children (odds ratio [OR] = 1.15, 95% confidence interval [CI] 1.01-1.32). There also appeared to be a borderline association between PM 2.5 level and risk of ALL among non-Hispanic White children (OR per 10 µg/m 3  = 1.24, 95% CI 0.98-1.56). We observed elevated risk of ALL in Hispanic children residing in areas of greater ALAN. Further work is needed to understand the role of ALAN and air pollution in the etiology of childhood ALL in different racial/ethnic groups., (© 2023. The Author(s).)

Only two-thirds of Americans meet the recommended 7 hours of sleep nightly. Insufficient sleep and circadian disruption have been associated with adverse health outcomes, including diabetes and cardiovascular disease. Several environmental disruptors of sleep have been reported, such as artificial light at night (ALAN) and noise. These studies tended to evaluate exposures individually. We evaluated several spatially derived environmental exposures (ALAN, noise, green space, and air pollution) and self-reported sleep outcomes obtained in 2012-2015 in a large cohort of 51,562 women in the California Teachers Study. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated for sleep duration and latency. After adjusting for age, race/ethnicity, chronotype, use of sleep medication, and self-reported trouble sleeping, ALAN (per 5 millicandela (mcd)/m2 luminance, OR = 1.13, 95% CI: 1.07, 1.20) and air pollution (per 5 μg/m3 PM2.5, OR = 1.06, 95% CI: 1.04, 1.09) were associated with shorter sleep duration (<7 hours), and noise was associated with longer latency (>15 minutes) (per 10 decibels, OR = 1.05, 95% CI: 1.01, 1.10). Green space was associated with increased duration (per 0.1 units, OR = 0.41, 95% CI: 0.28, 0.60) and decreased latency (per 0.1 units, OR = 0.55, 95% CI: 0.39, 0.78). Further research is necessary to understand how these and other exposures (e.g., diet) perturb an individuals' inherited sleep patterns and contribute to downstream health outcomes., (© The Author(s) 2022. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health.)

Due to a scarcity of routine monitoring of speciated particulate matter (PM), there has been limited capability to develop exposure models that robustly estimate component-specific concentrations. This paper presents the largest such study conducted in a single urban area. Using samples that were collected at 220 locations over two seasons, quasi-ultrafine (PM 0.2 ), accumulation mode fine (PM 0.2-2.5 ), and coarse (PM 2.5-10 ) particulate matter concentrations were used to develop spatiotemporal regression, machine learning models that enabled predictions of 24 elemental components in eight Southern California communities. We used supervised variable selection of over 150 variables, largely from publicly available sources, including meteorological, roadway and traffic characteristics, land use, and dispersion model estimates of traffic emissions. PM components that have high oxidative potential (and potentially large health effects) or are otherwise important markers for major PM sources were the primary focus. We present results for copper, iron, and zinc (as non-tailpipe vehicle emissions); elemental carbon (diesel emissions); vanadium (ship emissions); calcium (soil dust); and sodium (sea salt). Spatiotemporal linear regression models with 17 to 36 predictor variables including meteorology; distance to different classifications of roads; intersections and off ramps within a given buffer distance; truck and vehicle traffic volumes; and near-roadway dispersion model estimates produced superior predictions over the machine learning approaches (cross validation R-squares ranged from 0.76 to 0.92). Our models are easily interpretable and appear to have more effectively captured spatial gradients in the metallic portion of PM than other comparably large studies, particularly near roadways for the non-tailpipe emissions. Furthermore, we demonstrated the importance of including spatiotemporally resolved meteorology in our models as it helped to provide key insights into spatial patterns and allowed us to make temporal predictions., (Copyright © 2022 The Author(s). Published by Elsevier Ltd.. All rights reserved.)

Background: Studies demonstrated associations between maternal exposure to household air pollution from cooking and increased risk of adverse birth outcomes in offspring; however, the modifying effect of socioeconomic status (SES) on this association has not been explored., Objectives: In a cohort of pregnant women with 800 single live births between 2016 and 2017 in rural and semi urban areas of Bangladesh, we tested the hypotheses that kitchen location and cooking fuel type affect birth outcomes (birth weight, low birth weight [LBW] and small for gestational age [SGA]) and these associations vary by SES., Methods: Demographic characteristics including SES, kitchen location and fuel type were assessed in prenatal visits. Neonatal anthropometric measurements were recorded within 72 h of births. We performed multivariable linear and logistic regressions adjusting for potential confounders to test the study hypotheses., Results: For newborns from households with indoor kitchens, adjusted mean birth weight was 65.13 g (95% confidence interval [CI]: -118.37, -11.90) lower and the odds of LBW and SGA were 58% (odds ratio [OR]:1.58, 95% CI: 1.12, 2.24) and 41% (OR: 1.41, 95% CI: 1.05, 1.92) higher compared to those born in households with outdoor kitchens. We found SES significantly modified the associations between kitchen location and birth outcomes in households using biomass fuels. Newborns from low SES households with indoor kitchens had 89 g lower birth weight and a higher odds of being born with LBW (OR: 2.08, 95% CI 1.23, 3.58), and SGA (OR: 1.70, 95% CI 1.06, 2.76) than those born in high SES households using outdoor kitchens., Conclusions: In areas with poor access or affordability to clean fuel such as in our study population, cooking in an outdoor kitchen can reduce the burden of LBW and SGA, particularly for low SES households. Promoting outdoor kitchens is a possible intervention strategy to mitigate adverse birth outcomes., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2022 The Authors. Published by Elsevier B.V. All rights reserved.)

Many approaches to time series classification rely on machine learning methods. However, there is growing interest in going beyond black box prediction models to understand discriminatory features of the time series and their associations with outcomes. One promising method is time-series shapelets (TSS), which identifies maximally discriminative subsequences of time series. For example, in environmental health applications TSS could be used to identify short-term patterns in exposure time series (shapelets) associated with adverse health outcomes. Identification of candidate shapelets in TSS is computationally intensive. The original TSS algorithm used exhaustive search. Subsequent algorithms introduced efficiencies by trimming/aggregating the set of candidates or training candidates from initialized values, but these approaches have limitations. In this paper, we introduce Wavelet-TSS (W-TSS) a novel intelligent method for identifying candidate shapelets in TSS using wavelet transformation discovery. We tested W-TSS on two datasets: (1) a synthetic example used in previous TSS studies and (2) a panel study relating exposures from residential air pollution sensors to symptoms in participants with asthma. Compared to previous TSS algorithms, W-TSS was more computationally efficient, more accurate, and was able to discover more discriminative shapelets. W-TSS does not require pre-specification of shapelet length.

Background: Accurately assessing individual ambient air pollution exposure is a crucial part of epidemiological studies looking at the adverse health effect of poor air quality. This is particularly challenging in developing countries with high levels of air pollution, mostly due to sparse monitoring networks with a lack of consistent data., Methods: We evaluated the performance of six different machine learning algorithms in predicting fine particulate matter (PM 2.5 ) concentrations in Ulaanbaatar, Mongolia using data between 2010 and 2018. We found that the algorithms produce robust results based on performance metrics., Results: Random forest (RF) and gradient boosting models performed the best with leave-one-location-out cross-validated R 2 of 0.82 for when using data from the entire study period. After applying tuned models on the hold-out test set, R 2 increased to 0.96 for the RF and 0.90 for the gradient boosting model. We also predicted PM 2.5 concentrations for each administrative area (khoroo) of the city using RF and maps of predictions show spatiotemporal variations that are in line with the location of the high-emission area (ger district), city center, and population density., Conclusion: Our results provide evidence of the advantage and feasibility of machine learning approaches in predicting ambient PM 2.5 levels in a setting with limited resources and extreme air pollution levels.

Introduction: Estimating PM 2.5 concentrations and their prediction uncertainties at a high spatiotemporal resolution is important for air pollution health effect studies. This is particularly challenging for California, which has high variability in natural (e.g, wildfires, dust) and anthropogenic emissions, meteorology, topography (e.g. desert surfaces, mountains, snow cover) and land use., Methods: Using ensemble-based deep learning with big data fused from multiple sources we developed a PM 2.5 prediction model with uncertainty estimates at a high spatial (1 km × 1 km) and temporal (weekly) resolution for a 10-year time span (2008-2017). We leveraged autoencoder-based full residual deep networks to model complex nonlinear interrelationships among PM 2.5 emission, transport and dispersion factors and other influential features. These included remote sensing data (MAIAC aerosol optical depth (AOD), normalized difference vegetation index, impervious surface), MERRA-2 GMI Replay Simulation (M2GMI) output, wildfire smoke plume dispersion, meteorology, land cover, traffic, elevation, and spatiotemporal trends (geo-coordinates, temporal basis functions, time index). As one of the primary predictors of interest with substantial missing data in California related to bright surfaces, cloud cover and other known interferences, missing MAIAC AOD observations were imputed and adjusted for relative humidity and vertical distribution. Wildfire smoke contribution to PM 2.5 was also calculated through HYSPLIT dispersion modeling of smoke emissions derived from MODIS fire radiative power using the Fire Energetics and Emissions Research version 1.0 model., Results: Ensemble deep learning to predict PM 2.5 achieved an overall mean training RMSE of 1.54 μg/m 3 (R 2 : 0.94) and test RMSE of 2.29 μg/m 3 (R 2 : 0.87). The top predictors included M2GMI carbon monoxide mixing ratio in the bottom layer, temporal basis functions, spatial location, air temperature, MAIAC AOD, and PM 2.5 sea salt mass concentration. In an independent test using three long-term AQS sites and one short-term non-AQS site, our model achieved a high correlation (>0.8) and a low RMSE (<3 μg/m 3 ). Statewide predictions indicated that our model can capture the spatial distribution and temporal peaks in wildfire-related PM 2.5 . The coefficient of variation indicated highest uncertainty over deciduous and mixed forests and open water land covers., Conclusion: Our method can be generalized to other regions, including those having a mix of major urban areas, deserts, intensive smoke events, snow cover and complex terrains, where PM 2.5 has previously been challenging to predict. Prediction uncertainty estimates can also inform further model development and measurement error evaluations in exposure and health studies., (Copyright © 2020 The Author(s). Published by Elsevier Ltd.. All rights reserved.)

Unconventional extraction techniques including hydraulic fracturing or "fracking" have led to a boom in oil and gas production in the Eagle Ford shale play, Texas, one of the most productive regions in the United States. Nearly 400000 people live within 5 km of an unconventional oil or gas well in this largely rural area. Flaring is associated primarily with unconventional oil wells and is an increasingly common practice in the Eagle Ford to dispose of excess gas through combustion. Flares can operate continuously for months and release hazardous air pollutants such as particulate matter and volatile organic compounds in addition to causing light and noise pollution and noxious odors. We estimated ethnic disparities in exposure to flaring using satellite observations from the Visible Infrared Imaging Spectroradiometer between March 2012-December 2016. Census blocks with majority Hispanic (>60%) populations were exposed to twice as many nightly flare events within 5 km as those with <20% Hispanics. We found that Hispanics were exposed to more flares despite being less likely than non-Hispanic White residents to live near unconventional oil and gas wells. Our findings suggest Hispanics are disproportionately exposed to flares in the Eagle Ford shale, a pattern known as environmental injustice, which could contribute to disparities in air pollution and other nuisance exposures.

Background: Although the association between exposure to particulate matter and health is well established, there remains uncertainty as to whether certain chemical components are more harmful than others. We explored whether the association between cause-specific hospital admissions and PM(2.5) was modified by PM(2.5) chemical composition., Methods: We estimated the association between daily PM(2.5) and emergency hospital admissions for cardiac causes (CVD), myocardial infarction (MI), congestive heart failure (CHF), respiratory disease, and diabetes in 26 US communities, for the years 2000-2003. Using meta-regression, we examined how this association was modified by season- and community-specific PM(2.5) composition, controlling for seasonal temperature as a surrogate for ventilation., Results: For a 10 microg/m3 increase in 2-day averaged PM(2.5) concentration we found an increase of 1.89% (95% CI: 1.34- 2.45) in CVD, 2.25% (95% CI: 1.10- 3.42) in MI, 1.85% (95% CI: 1.19- 2.51) in CHF, 2.74% (95% CI: 1.30- 4.2) in diabetes, and 2.07% (95% CI: 1.20- 2.95) in respiratory admissions. The association between PM2.5 and CVD admissions was significantly modified when the mass was high in Br, Cr, Ni, and Na(+), while mass high in As, Cr, Mn, OC, Ni, and Na(+) modified MI, and mass high in As, OC, and SO(4)(2-) modified diabetes admissions. For these species, an interquartile range increase in their relative proportion was associated with a 1-2% additional increase in daily admissions per 10 microg/m(3) increase in mass., Conclusions: We found that PM(2.5) mass higher in Ni, As, and Cr, as well as Br and OC significantly increased its effect on hospital admissions. This result suggests that particles from industrial combustion sources and traffic may, on average, have greater toxicity.

Background: Although several previous studies have found a positive association between ambient ozone and mortality, the observed effect may be confounded by other secondary pollutants that are produced concurrently with ozone., Objectives: We addressed the question of whether the ozone-mortality relationship is entirely a reflection of the adverse effect of ozone, or whether it is, at least in part, an effect of other secondary pollutants., Methods: Separate time-series models were fit to 3-6 years of data between 2000 and 2005 from 18 U.S. communities. The association between nonaccidental mortality was examined with ozone alone and with ozone after adjustment for fine particle mass, sulfate, organic carbon, or nitrate concentrations. The effect estimates from each of these models were pooled using a random-effects meta-analysis to obtain an across-community average., Results: We found a 0.89% [95% confidence interval (CI), 0.45-1.33%] increase in nonaccidental mortality with a 10-ppb increase in same-day 24-hr summertime ozone across the 18 communities. After adjustment for PM(2.5) (particulate matter with aerodynamic diameter

IntroductionEstimating PM2.5 concentrations and their prediction uncertainties at a high spatiotemporal resolution is important for air pollution health effect studies. This is particularly challenging for California, which has high variability in natural (e.g, wildfires, dust) and anthropogenic emissions, meteorology, topography (e.g. desert surfaces, mountains, snow cover) and land use.MethodsUsing ensemble-based deep learning with big data fused from multiple sources we developed a PM2.5 prediction model with uncertainty estimates at a high spatial (1 km × 1 km) and temporal (weekly) resolution for a 10-year time span (2008-2017). We leveraged autoencoder-based full residual deep networks to model complex nonlinear interrelationships among PM2.5 emission, transport and dispersion factors and other influential features. These included remote sensing data (MAIAC aerosol optical depth (AOD), normalized difference vegetation index, impervious surface), MERRA-2 GMI Replay Simulation (M2GMI) output, wildfire smoke plume dispersion, meteorology, land cover, traffic, elevation, and spatiotemporal trends (geo-coordinates, temporal basis functions, time index). As one of the primary predictors of interest with substantial missing data in California related to bright surfaces, cloud cover and other known interferences, missing MAIAC AOD observations were imputed and adjusted for relative humidity and vertical distribution. Wildfire smoke contribution to PM2.5 was also calculated through HYSPLIT dispersion modeling of smoke emissions derived from MODIS fire radiative power using the Fire Energetics and Emissions Research version 1.0 model.ResultsEnsemble deep learning to predict PM2.5 achieved an overall mean training RMSE of 1.54 μg/m3 (R2: 0.94) and test RMSE of 2.29 μg/m3 (R2: 0.87). The top predictors included M2GMI carbon monoxide mixing ratio in the bottom layer, temporal basis functions, spatial location, air temperature, MAIAC AOD, and PM2.5 sea salt mass concentration. In an independent test using three long-term AQS sites and one short-term non-AQS site, our model achieved a high correlation (>0.8) and a low RMSE (

BackgroundAccurate 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.ObjectivesIn 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.MethodsWe 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.ResultsWe 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.ConclusionsOur 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.

Existing methods for fine-scale air quality assessment have significant gaps in their reliability. Purely data-driven methods lack any physically-based mechanisms to simulate the interactive process of air pollution, potentially leading to physically inconsistent or implausible results. Here, we report a hybrid multilevel graph neural network that encodes fluid physics to capture spatial and temporal dynamic characteristics of air pollutants. On a multi-air pollutant test in China, our method consistently improved extrapolation accuracy by an average of 11–22% compared to several baseline machine learning methods, and generated physically consistent spatiotemporal trends of air pollutants at fine spatial and temporal scales. [ABSTRACT FROM AUTHOR]

Exposure to traffic-related air pollution (TRAP) in the near-roadway environment is associated with multiple adverse health effects. To characterize the relative contribution of tailpipe and non-tailpipe TRAP sources to particulate matter (PM) in the quasi-ultrafine (PM0.2), fine (PM2.5) and coarse (PM2.5–10) size fractions and identify their spatial determinants in southern California (CA). Month-long integrated PM0.2, PM2.5 and PM2.5–10 samples (n = 461, 265 and 298, respectively) were collected across cool and warm seasons in 8 southern CA communities (2008–9). Concentrations of PM mass, elements, carbons and major ions were obtained. Enrichment ratios (ER) in PM0.2 and PM10 relative to PM2.5 were calculated for each element. The Positive Matrix Factorization model was used to resolve and estimate the relative contribution of TRAP sources to PM in three size fractions. Generalized additive models (GAMs) with bivariate loess smooths were used to understand the geographic variation of TRAP sources and identify their spatial determinants. EC, OC, and B had the highest median ER in PM0.2 relative to PM2.5. Six, seven and five sources (with characteristic species) were resolved in PM0.2, PM2.5 and PM2.5–10, respectively. Combined tailpipe and non-tailpipe traffic sources contributed 66%, 32% and 18% of PM0.2, PM2.5 and PM2.5–10 mass, respectively. Tailpipe traffic emissions (EC, OC, B) were the largest contributor to PM0.2 mass (58%). Distinct gasoline and diesel tailpipe traffic sources were resolved in PM2.5. Others included fuel oil, biomass burning, secondary inorganic aerosol, sea salt, and crustal/soil. CALINE4 dispersion model nitrogen oxides, trucks and intersections were most correlated with TRAP sources. The influence of smaller roadways and intersections became more apparent once Long Beach was excluded. Non-tailpipe emissions constituted ~8%, 11% and 18% of PM0.2, PM2.5 and PM2.5–10, respectively, with important exposure and health implications. Future efforts should consider non-linear relationships amongst predictors when modeling exposures. Implications: Vehicle emissions result in a complex mix of air pollutants with both tailpipe and non-tailpipe components. As mobile source regulations lead to decreased tailpipe emissions, the relative contribution of non-tailpipe traffic emissions to near-roadway exposures is increasing. This study documents the presence of non-tailpipe abrasive vehicular emissions (AVE) from brake and tire wear, catalyst degradation and resuspended road dust in the quasi-ultrafine (PM0.2), fine and coarse particulate matter size fractions, with contributions reaching up to 30% in PM0.2 in some southern California communities. These findings have important exposure and policy implications given the high metal content of AVE and the efficiency of PM0.2 at reaching the alveolar region of the lungs and other organ systems once inhaled. This work also highlights important considerations for building models that can accurately predict tailpipe and non-tailpipe exposures for population health studies. [ABSTRACT FROM AUTHOR]

Biomass fuel burning is a significant contributor of household fine particulate matter (PM 2.5) in the low to middle income countries (LMIC) and assessing PM 2.5 levels is essential to investigate exposure-related health effects such as pregnancy outcomes and acute lower respiratory infection in infants. However, measuring household PM 2.5 requires significant investments of labor, resources, and time, which limits the ability to conduct health effects studies. It is therefore imperative to leverage lower-cost measurement techniques to develop exposure models coupled with survey information about housing characteristics. Between April 2017 and March 2018, we continuously sampled PM 2.5 in three seasonal waves for approximately 48-h (range 46 to 52-h) in 74 rural and semi-urban households among the participants of the Bangladesh Cook Stove Pregnancy Cohort Study (CSPCS). Measurements were taken simultaneously in the kitchen, bedroom, and open space within the household. Structured questionnaires captured household-level information related to the sources of air pollution. With data from two waves, we fit multivariate mixed effect models to estimate 24-h average, cooking time average, daytime and nighttime average PM 2.5 in each of the household locations. Households using biomass cookstoves had significantly higher PM 2.5 concentrations than those using electricity/liquefied petroleum gas (626 μg/m3 vs. 213 μg/m3). Exposure model performances showed 10-fold cross validated R2 ranging from 0.52 to 0.76 with excellent agreement in independent tests against measured PM 2.5 from the third wave of monitoring and ambient PM 2.5 from a separate satellite-based model (correlation coefficient, r = 0.82). Significant predictors of household PM 2.5 included ambient PM 2.5 , season, and types of fuel used for cooking. This study demonstrates that we can predict household PM 2.5 with moderate to high confidence using ambient PM 2.5 and household characteristics. Our results present a framework for estimating household PM 2.5 exposures in LMICs, which are often understudied and underrepresented due to resource limitations. [Display omitted] • Characterization of household air pollution is lacking, particularly in LMICs. • Simultaneous PM 2.5 monitoring was done in multiple locations within each household. • Monitored data coupled with household characteristics enabled successful modelling. • Ambient PM 2.5 plays an important role in predicting household PM 2.5. [ABSTRACT FROM AUTHOR]

Mounting evidence indicates that early-life exposure to particulate air pollutants pose threats to children’s cognitive development, but studies about the neurotoxic effects associated with exposures during adolescence remain unclear. We examined whether exposure to ambient fine particles (PM2.5) at residential locations affects intelligence quotient (IQ) during pre-/early- adolescence (ages 9–11) and emerging adulthood (ages 18–20) in a demographically-diverse population (N = 1,360) residing in Southern California. Increased ambient PM2.5 levels were associated with decreased IQ scores. This association was more evident for Performance IQ (PIQ), but less for Verbal IQ, assessed by the Wechsler Abbreviated Scale of Intelligence. For each inter-quartile (7.73 μg/m3) increase in one-year PM2.5 preceding each assessment, the average PIQ score decreased by 3.08 points (95% confidence interval = [-6.04, -0.12]) accounting for within-family/within-individual correlations, demographic characteristics, family socioeconomic status (SES), parents’ cognitive abilities, neighborhood characteristics, and other spatial confounders. The adverse effect was 150% greater in low SES families and 89% stronger in males, compared to their counterparts. Better understanding of the social disparities and sexual dimorphism in the adverse PM2.5–IQ effects may help elucidate the underlying mechanisms and shed light on prevention strategies. [ABSTRACT FROM AUTHOR]

The author convey their response to a feedback about their report on how environmental exposures and their inequities contribute to sleep disruptions. Topics mentioned include the limitations of satellite imagery in assessing artificial light at night, the ability to examine short-wavelength blue light, and measurement of the correlation between outdoor light and indoor exposure.

Although it has been shown that traffic-related air pollution adversely affects children's lung function, few studies have examined the influence of traffic noise on this association, despite both sharing a common source. Estimates of noise exposure (L dn, dB), and freeway and non-freeway emission concentrations of oxides of nitrogen (NO x , ppb) were spatially assigned to children in Southern California who were tested for forced vital capacity (FVC, n=1345), forced expiratory volume in 1 s, (FEV 1, n=1332), and asthma. The associations between traffic-related NO x and these outcomes, with and without adjustment for noise, were examined using mixed effects models. Adjustment for noise strengthened the association between NO x and reduced lung function. A 14.5 mL (95% CI −40.0, 11.0 mL) decrease in FVC per interquartile range (13.6 ppb) in freeway NO x was strengthened to a 34.6 mL decrease after including a non-linear function of noise (95% CI −66.3, −2.78 mL). Similarly, a 6.54 mL decrease in FEV 1 (95% CI −28.3, 15.3 mL) was strengthened to a 21.1 mL decrease (95% CI −47.6, 5.51) per interquartile range in freeway NO x . Our results indicate that where possible, noise should be included in epidemiological studies of the association between traffic-related air pollution on lung function. Without taking noise into account, the detrimental effects of traffic-related pollution may be underestimated. [ABSTRACT FROM AUTHOR]

There is significant public health interest in gaining a better understanding of the health effects associated with particulate matter (PM) of differing sizes and compositions. Unfortunately the PM monitoring data needed for such studies are limited. Satellite observations of aerosol optical depth (AOD) have been used to overcome this limitation, but existing techniques using AOD to estimate surface concentrations of PM have limited ability to differentiate particle size and type, and the spatial resolution is often too coarse to evaluate PM variability on local scales. Fortunately there has been a significant breakthrough in aerosol product from the Multi-angle Imaging SpectroRadiometer (MISR) instrument onboard the NASA Terra satellite. Recent computational advances have made it possible to derive a “local mode” MISR aerosol product at a spatial resolution of 4.4 km that maintains the quality of the 17.6 km resolution Version 22 operational aerosol product. MISR is uniquely capable of providing total AOD as well as AOD fractionated by size (small, medium, and large), enabling estimation of size-resolved surface PM concentrations. Total and size-fractionated AOD from the MISR 4.4 km aerosol product were used to develop spatio-temporal models to estimate surface concentrations of PM 2.5 , PM 10 , and PM 2.5 chemical species over Southern California. We found, and confirmed by leave-one-site-out cross validation, that PM 2.5 was best estimated with a spatio-temporal model of AOD small + medium that also included adjustment for relative humidity and wind speed (R 2 = 0.67, CV R 2 = 0.51), while PM 10 was best estimated from AOD large with adjustment for dew point and wind speed (R 2 = 0.76, CV R 2 = 0.44). Total AOD was most strongly associated with PM 2.5 components SO 4 2 − (R 2 = 0.74, CV R 2 = 0.69) and NO 3 − (R 2 = 0.72, CV R 2 = 0.39). The best fitting models were applied to all available MISR aerosol retrievals over the region, generating surfaces of estimated size-resolved PM concentrations that will be a great asset to the environmental science and public health communities. [ABSTRACT FROM AUTHOR]

Spatially and temporally resolved aerosol data are essential for conducting air quality studies and assessing the health effects associated with exposure to air pollution. As these data are often expensive to acquire and time consuming to estimate, computationally efficient methods are desirable. When coarse-scale data or imagery are available, fine-scale data can be generated through downscaling methods. We developed an Artificial Neural Network Sequential Downscaling Method (ASDM) with Transfer Learning Enhancement (ASDMTE) to translate time-series data from coarse- to fine-scale while maintaining between-scale empirical associations as well as inherent within-scale correlations. Using assimilated aerosol optical depth (AOD) from the GEOS-5 Nature Run (G5NR) (2 years, daily, 7 km resolution) and Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2) (20 years, daily, 50 km resolution), coupled with elevation (1 km resolution), we demonstrate the downscaling capability of ASDM and ASDMTE and compare their performances against a deep learning downscaling method, Super Resolution Deep Residual Network (SRDRN), and a traditional statistical downscaling framework called dissever ASDM/ASDMTE utilizes empirical between-scale associations, and accounts for within-scale temporal associations in the fine-scale data. In addition, within-scale temporal associations in the coarse-scale data are integrated into the ASDMTE model through the use of transfer learning to enhance downscaling performance. These features enable ASDM/ASDMTE to be trained on short periods of data yet achieve a good downscaling performance on a longer time-series. Among all the test sets, ASDM and ASDMTE had mean maximum image-wise R 2 of 0.735 and 0.758, respectively, while SRDRN, dissever GAM and dissever LM had mean maximum image-wise R 2 of 0.313, 0.106 and 0.095, respectively. [ABSTRACT FROM AUTHOR]

Air quality has emerged as a key determinant of important health outcomes in children and adults. This study aims to identify factors that influence local, within-community air quality, and to build a model for traffic-related air pollution (TRP).We utilized concentrations of NO2, NO, and total oxides of nitrogen (NOx), which were measured at 942 locations in 12 southern California communities. For each location, population density, elevation, land-use, and several indicators of traffic were calculated. A spatial random effects model was used to study the relationship of these predictors to each TRP.Variation in TRP was strongly correlated with traffic on nearby freeways and other major roads, and also with population density and elevation. After accounting for traffic, categories of land-use were not associated with the pollutants. Traffic had a larger relative impact in small urban (low regional pollution) communities than in large urban (high regional pollution) communities. For example, our best fitting model explained 70% of the variation in NOx in large urban areas and 76% in small urban areas. Compared with living at least 1,500 m from a freeway, living within 250 m of a freeway was associated with up to a 41% increase in TRP in a large urban area, and up to a 75% increase in small urban areas.Thus, traffic strongly affects local air quality in large and small urban areas, which has implications for exposure assessment and estimation of health risks. [ABSTRACT FROM AUTHOR]

While fine mode particulate matter (PM2.5) forms the basis for regulating particles in the US and other countries, there is a serious paucity of large population-based studies of its acute effect on mortality. To address this issue, we examined the association between PM2.5 and both all-cause and specific-cause mortality using over 1.3 million deaths in 27 US communities between 1997 and 2002. A two-stage approach was used. First, the association between PM2.5 and mortality in each community was quantified using a case-crossover design. Second, meta-analysis was used to estimate a summary effect over all 27 communities. Effect modification of age and gender was examined using interaction terms in the case-crossover model, while effect modification of community-specific characteristics including geographic location, annual PM2.5 concentration above 15 μg/m3 and central air conditioning prevalence was examined using meta-regression. We observed a 1.21% (95% CI 0.29, 2.14%) increase in all-cause mortality, a 1.78% (95% CI 0.20, 3.36%) increase in respiratory related mortality and a 1.03% (95% CI 0.02, 2.04%) increase in stroke related mortality with a 10 μg/m3 increase in previous day's PM2.5. The magnitude of these associations is more than triple that recently reported for PM10, suggesting that combustion and traffic related particles are more toxic than larger sized particles. Effect modification occurred in all-cause and specific-cause deaths with greater effects in subjects ≥75 years of age. There was suggestive evidence that women may be more susceptible to PM2.5 effects than men, and that effects were larger in the East than in the West. Increased prevalence of central air conditioning was associated with a decreased effect of PM2.5. Our findings describe the magnitude of the effect on all-cause and specific-cause mortality, the modifiers of this association, and suggest that PM2.5 may pose a public health risk even at or below current ambient levels.Journal of Exposure Science and Environmental Epidemiology (2007) 17, 279–287. doi:10.1038/sj.jes.7500530; published online 27 September 2006 [ABSTRACT FROM AUTHOR]

Using two general linear regression models, we compared the ability of the aerosol optical thickness (AOT) retrieved by the Multiangle Imaging SpectroRadiometer (MISR) and the Moderate Resolution Imaging Spectroradiometer (MODIS) to predict ground-level PM2. 5 concentrations in St. Louis, MO and its surrounding areas . The models included meteorological parameters obtained from the National Oceanic and Atmospheric Administration (NOAA)''s Rapid Update Cycle (RUC20) model as covariates. Both MISR and MODIS AOT values were highly significant predictors of PM2. 5 concentrations. The MISR and MODIS models have overall comparable predictability of ground-level PM2. 5 concentrations. The MISR model explained a slightly greater percentage (62%) of the variability in PM2. 5 concentrations than the MODIS model (51%), and thus was a better fit. Over the entire data range, the MISR model underpredicts PM2. 5 concentrations by approximately 12%, whereas the MODIS model underpredicts PM2. 5 concentrations by approximately 18%. This underestimation occurred primarily at higher PM2. 5 concentrations in both models. The regression coefficients from two models were highly comparable, suggesting that combining MISR and MODIS AOT data might benefit from the higher predicting accuracy of MISR and the better spatial coverage of MODIS. The newly developed particle size/shape indicators in MISR and MODIS aerosol product did not significantly improve our ability to predict PM2. 5 concentrations using AOT measurements. Finally, using hourly PM2. 5 concentrations did not seem to improve its association with AOT for the current study region. [Copyright &y& Elsevier]

Exposure to fine particulate matter (PM 2.5 ) air pollution has been shown in numerous studies to be associated with detrimental health effects. However, the ability to conduct epidemiological assessments can be limited due to challenges in generating reliable PM 2.5 estimates, particularly in parts of the world such as the Middle East where measurements are scarce and extreme meteorological events such as sandstorms are frequent. In order to supplement exposure modeling efforts under such conditions, satellite-retrieved aerosol optical depth (AOD) has proven to be useful due to its global coverage. By using AODs from the Multiangle Implementation of Atmospheric Correction (MAIAC) of the MODerate Resolution Imaging Spectroradiometer (MODIS) and the Multiangle Imaging Spectroradiometer (MISR) combined with meteorological and assimilated aerosol information from the Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2), we constructed machine learning models to predict PM 2.5 in the area surrounding the Persian Gulf, including Kuwait, Bahrain, and the United Arab Emirates (U.A.E). Our models showed regional differences in predictive performance, with better results in the U.A.E. (median test R 2 = 0.66) than Kuwait (median test R 2 = 0.51). Variable importance also differed by region, where satellite-retrieved AOD variables were more important for predicting PM 2.5 in Kuwait than in the U.A.E. Divergent trends in the temporal and spatial autocorrelations of PM 2.5 and AOD in the two regions offered possible explanations for differences in predictive performance and variable importance. In a test of model transferability, we found that models trained in one region and applied to another did not predict PM 2.5 well, even if the transferred model had better performance. Overall the results of our study suggest that models developed over large geographic areas could generate PM 2.5 estimates with greater uncertainty than could be obtained by taking a regional modeling approach. Furthermore, development of methods to better incorporate spatial and temporal autocorrelations in machine learning models warrants further examination. [ABSTRACT FROM AUTHOR]

Studies of the association between air pollution and children's health typically rely on fixed-site monitors to determine exposures, which have spatial and temporal limitations. Satellite observations of aerosols provide the coverage that fixed-site monitors lack, enabling more refined exposure assessments. Using aerosol optical depth (AOD) data from the Multiangle Imaging SpectroRadiometer (MISR) instrument, we predicted fine particulate matter, PM 2.5 , and PM 2.5 speciation concentrations and linked them to the residential locations of 1206 children enrolled in the Southern California Children's Health Study. We fitted mixed-effects models to examine the relationship between the MISR-derived exposure estimates and lung function, measured as forced expiratory volume in 1 second (FEV 1 ) and forced vital capacity (FVC), adjusting for study community and biological factors. Gradient Boosting and Support Vector Machines showed excellent predictive performance for PM 2.5 (test R 2 = 0.68 ) and its chemical components (test R 2 = –0.71). In single-pollutant models, FEV 1 decreased by 131 mL (95% CI: − 232 , − 35 ) per 10.7-µg/m 3 increase in PM 2.5 , by 158 mL (95% CI: − 273 , − 43 ) per 1.2-µg/m 3 in sulfates (SO 4 2 − ), and by 177 mL (95% CI: − 306 , − 56 ) per 1.6-µg/m 3 increase in dust; FVC decreased by 175 mL (95% CI: − 310 , − 29 ) per 1.2-µg/m 3 increase in SO 4 2 − and by 212 mL (95% CI: − 391 , − 28 ) per 2.5-µg/m 3 increase in nitrates (NO 3 − ). These results demonstrate that satellite observations can strengthen epidemiological studies investigating air pollution health effects by providing spatially and temporally resolved exposure estimates. [ABSTRACT FROM AUTHOR]

Many air pollution health effects studies rely on exposure estimates of particulate matter (PM) concentrations derived from remote sensing observations of aerosol optical depth (AOD). Simple but robust calibration models between AOD and PM are therefore important for generating reliable PM exposures. We conduct an in-depth examination of the spatial and temporal characteristics of the AOD-PM2.5 relationship by leveraging data from the Distributed Regional Aerosol Gridded Observation Networks (DRAGON) field campaign where eight NASA Aerosol Robotic Network (AERONET) sites were co-located with EPA Air Quality System (AQS) monitoring sites in California's Central Valley from November 2012 to April 2013. With this spatiotemporally rich data we found that linear calibration models (R2 = 0.35, RMSE = 10.38 μg/m3) were significantly improved when spatial (R2 = 0.45, RMSE = 9.54 μg/m3), temporal (R2 = 0.62, RMSE = 8.30 μg/m3), and spatiotemporal (R2 = 0.65, RMSE = 7.58 μg/m3) functions were included. As a use-case we applied the best spatiotemporal model to convert space-borne MultiAngle Imaging Spectroradiometer (MISR) AOD observations to predict PM2.5 over the region (R2 = 0.60, RMSE = 8.42 μg/m3). Our results imply that simple AERONET AOD-PM2.5 calibrations are robust and can be reliably applied to space-borne AOD observations, resulting in PM2.5 prediction surfaces for use in downstream applications. [ABSTRACT FROM AUTHOR]

Ulaanbaatar (UB), the capital city of Mongolia, has extremely poor wintertime air quality with fine particulate matter concentrations frequently exceeding 500 μg/m3, over 20 times the daily maximum guideline set by the World Health Organization. Intensive use of sulfur-rich coal for heating and cooking coupled with an atmospheric inversion amplified by the mid-continental Siberian anticyclone drive these high levels of air pollution. Ground-based air quality monitoring in Mongolia is sparse, making use of satellite observations of aerosol optical depth (AOD) instrumental for characterizing air pollution in the region. We harnessed data from the Multi-angle Imaging SpectroRadiometer (MISR) Version 23 (V23) aerosol product, which provides total column AOD and component-particle optical properties for 74 different aerosol mixtures at 4.4 km spatial resolution globally. To test the performance of the V23 product over Mongolia, we compared values of MISR AOD with spatially and temporally matched AOD from the Dalanzadgad AERONET site and find good agreement (correlation r = 0.845, and root-mean-square deviation RMSD = 0.071). Over UB, exploratory principal component analysis indicates that the 74 MISR AOD mixture profiles consisted primarily of small, spherical, non-absorbing aerosols in the wintertime, and contributions from medium and large dust particles in the summertime. Comparing several machine learning methods for relating the 74 MISR mixtures to ground-level pollutants, including particulate matter with aerodynamic diameters smaller than 2.5 μm ( PM 2.5 ) and 10 μm ( PM 10 ), as well as sulfur dioxide ( SO 2 ), a proxy for sulfate particles, we find that Support Vector Machine regression consistently has the highest predictive performance with median test R 2 for PM 2.5 , PM 10 , and SO 2 equal to 0.461, 0.063, and 0.508, respectively. These results indicate that the high-dimensional MISR AOD mixture set can provide reliable predictions of air pollution and can distinguish dominant particle types in the UB region. [ABSTRACT FROM AUTHOR]

Recent use of satellite observations of aerosol optical depth (AOD) to characterize surface concentrations of particulate matter (PM) air pollution has proven extremely valuable in estimating exposures for health effects studies. While the spatial resolutions of satellite data provide far better coverage than existing fixed site surface monitoring stations, they are not able to capture atmospheric processes such as dilution of primary pollutants that vary at small spatial scales. As a result, small-scale variability due to highly localized sources such as traffic may be poorly represented, which in turn may lead to exposure measurement error in epidemiological analyses. Using a fixed spatial grid representing 4.4 km Multiangle Imaging SpectroRadiometer (MISR) aerosol observations, we examined the spatial variability in fine and coarse mode PM (PM2.5 and PM2.5–10 respectively) measured at ground monitors from a unique spatially-dense sampling campaign in Southern California. We found that while the variance in measured PM2.5 differed seasonally (warm 6.82 μg2/m6 and cool 24.5 μg2/m6) across the study region, the average subgrid (<4.4 km) variance did not (warm 2.03 μg2/m6 and cool 2.43 μg2/m6) and was significantly smaller. On the other hand, ground monitor PM2.5–10 concentrations showed large variance in warm (18.6 μg2/m6) and cool (20.6 μg2/m6) seasons, as well as seasonal differences in subgrid variance (warm 8.90 μg2/m6 and cool 3.28 μg2/m6). Geostatistical analysis of the semivariance as a function of distance indicated that variability in measured PM2.5 and PM2.5–10 concentrations was relatively constant for spatial scales of one to five kilometers, but there was evidence of small-scale (~500 m) variability in PM2.5–10 concentrations in the cool season. The lack of small-scale spatial variability in the warm season was likely due to large photochemical contributions to regional PM2.5, and greater regional contributions to PM2.5–10 from windblown dust. In contrast, in the cool season there tends to be greater localized concentrations from primary traffic sources due to stronger nocturnal inversions and delayed morning winds reducing dilution that contribute to greater spatial heterogeneity. Overall, these results suggest that regional contributions tend to dominate PM2.5, and spatial resolutions of satellite observations including the 4.4 km MISR and 3 km MODIS aerosol products aptly capture relevant spatial variability. Coarse PM2.5–10 can have seasonally dependent localized contributions, leading to small-scale variability below current satellite aerosol product resolutions. [ABSTRACT FROM AUTHOR]

Unlike air pollution, traffic-related noise remains unregulated and has been under-studied despite evidence of its deleterious health impacts. To characterize population exposure to traffic noise, both acoustic-based numerical models and data-driven statistical approaches can generate estimates over large urban areas. The aim of this work is to formally compare the performances of the most common traffic noise models by evaluating their estimates for different categories of roads and validating them against a unique dataset of measured noise in Long Beach, California. Specifically, a statistical land use regression model, an extreme gradient boosting machine learning model (XGB), and three numerical/acoustic traffic noise models: the US Noise Model (FHWA-TNM2.5), a commercial noise model (CadnaA), and an open-source European model (Harmonoise) were optimized and compared. The results demonstrate that XGB and CadnaA were the most effective models for estimating traffic noise, and they are particularly adept at differentiating noise levels on different categories of road. [ABSTRACT FROM AUTHOR]

Abstract: To characterize exposures to particulate matter (PM) and its components, we performed a large sampling study of small-scale spatial variation in size-resolved particle mass and composition. PM was collected in size ranges of <0.2, 0.2-to-2.5, and 2.5-to-10 μm on a scale of 100s to 1000s of meters to capture local sources. Within each of eight Southern California communities, up to 29 locations were sampled for rotating, month-long integrated periods at two different times of the year, six months apart, from Nov 2008 through Dec 2009. Additional sampling was conducted at each community's regional monitoring station to provide temporal coverage over the sampling campaign duration. Residential sampling locations were selected based on a novel design stratified by high- and low-predicted traffic emissions and locations over- and under-predicted from previous dispersion model and sampling comparisons. Primary vehicle emissions constituents, such as elemental carbon (EC), showed much stronger patterns of association with traffic than pollutants with significant secondary formation, such as PM2.5 or water soluble organic carbon. Associations were also stronger during cooler times of the year (Oct through Mar). Primary pollutants also showed greater within-community spatial variation compared to pollutants with secondary formation contributions. For example, the average cool-season community mean and standard deviation (SD) for EC were 1.1 and 0.17 μg m−3, respectively, giving a coefficient of variation (CV) of 18%. For PM2.5, average mean and SD were 14 and 1.3 μg m−3, respectively, with a CV of 9%. We conclude that within-community spatial differences are important for accurate exposure assessment of traffic-related pollutants. [Copyright &y& Elsevier]