Land Use Regression Models Research Articles

Air pollution, surrounding green, road proximity and Parkinson's disease: A prospective cohort study

BackgroundThough growing evidence has linked air pollution to Parkinson's disease (PD), the results remain inconsistent. Less is known about the relevance of road proximity and surrounding green. We aimed to investigate the individual and joint associations of air pollution, road proximity and surrounding green with the incidence of PD in a prospective cohort study. MethodsWe used data from a prospective cohort of 47,516 participants recruited from July 2015 to January 2018 in Ningbo, China. Long-term exposure to particulate matter with aerodynamic diameter ≤2.5 μm (PM2.5) and ≤10 μm (PM10) and nitrogen dioxide (NO2) estimated by land-use regression models, road proximity and surrounding green assessed by Normalized Difference Vegetation Index (NDVI) were calculated based on the residential address for each participant. Cox proportional hazard models were used to analyze the individual and joint effects of air pollution, road proximity, and surrounding green on PD. ResultsIn single-exposure models, PM2.5, PM10, NO2 and road proximity was associated with increased risk of PD (e.g. Hazard Ratio (HR) = 1.51, 95%CI:1.02, 2.24 per interquartile range (IQR) increase for PM2.5) while surrounding green was associated with decreased risk of PD (e.g. HR = 0.80, 95%CI:0.65, 0.98 per IQR increase for NDVI in 300 m buffer). In two-exposure models, the associations of PM2.5 and surrounding green persisted while the associations of NO2 and road proximity attenuated towards unity. ConclusionsWe found that PM2.5 were associated with increased risk of incident PD while surrounding green was associated with decreased risk of PD. Future studies about PD etiology may benefit from including multiple environmental exposures to address potential joint associations.

Preconceptional and the first trimester exposure to PM2.5 and offspring neurodevelopment at 24 months of age: Examining mediation by maternal thyroid hormones in a birth cohort study

Prenatal fine particulate matter (PM2.5) exposure has been associated with impaired offspring neurodevelopment; however, the association of PM2.5 exposure during preconception with offspring’s neurodevelopment and factors responsible for this association are still unclear. This study estimated the associations of PM2.5 exposure during preconception and the first trimester with offspring neurodevelopment and evaluated whether maternal thyroid hormones mediate these associations. We recruited 1329 mother-child pairs between 2013 and 2015 in Wuhan, China. PM2.5 exposure levels of each woman during the 3 months preconception and the first trimester were estimated using land-use regression models. Offspring neurodevelopment characterized by mental developmental index (MDI) and psychomotor developmental index (PDI) were measured at 24 months of age. Maternal serum levels of free thyroxine (FT3), free triiodothyronine (FT4), and thyroid-stimulating hormone (TSH) during early pregnancy were measured of a subset of the 1329 women (551 women). Generalized estimation equation and general linear regression models were used to estimate the associations between maternal PM2.5 exposure, thyroid hormones, and offspring neurodevelopment. After adjusting for potential confounders, we found that either among all participants or the subset, PM2.5 exposure during preconception and the first trimester was negatively associated with offspring PDI. Double increment in the first trimester PM2.5 exposure was significantly associated with 3.43 and 6.48 points decrease in offspring MDI. In the subset, each doubling of PM2.5 exposure during preconception and the first trimester was significantly associated with 7.93 and 8.02 points decrease in maternal FT4 level, respectively. Increased maternal FT4, in turn, was associated with increased PDI (β = 16.69, 95% CI: 5.39, 27.99). About 7.7% (95% CI: 2.0%–19.4%) and 8.6% (95% CI: 3.0%, 22.1%) of the effect of PM2.5 exposure during preconception on offspring PDI was mediated through maternal FT4 and the FT4/FT3 ratio, respectively.

Within-City Variation in Reactive Oxygen Species from Fine Particle Air Pollution and COVID-19

Rationale: Evidence linking outdoor air pollution with coronavirus disease (COVID-19) incidence and mortality is largely based on ecological comparisons between regions that may differ in factors such as access to testing and control measures that may not be independent of air pollution concentrations. Moreover, studies have yet to focus on key mechanisms of air pollution toxicity such as oxidative stress.Objectives: To conduct a within-city analysis of spatial variations in COVID-19 incidence and the estimated generation of reactive oxygen species (ROS) in lung lining fluid attributable to fine particulate matter (particulate matter with an aerodynamic diameter ⩽2.5 μm [PM2.5]).Methods: Sporadic and outbreak-related COVID-19 case counts, testing data, population data, and sociodemographic data for 140 neighborhoods were obtained from the City of Toronto. ROS estimates were based on a mathematical model of ROS generation in lung lining fluid in response to iron and copper in PM2.5. Spatial variations in long-term average ROS were predicted using a land-use regression model derived from measurements of iron and copper in PM2.5. Data were analyzed using negative binomial regression models adjusting for covariates identified using a directed acyclic graph and accounting for spatial autocorrelation.Measurements and Main Results: A significant positive association was observed between neighborhood-level ROS and COVID-19 incidence (incidence rate ratio = 1.07; 95% confidence interval, 1.01–1.15 per interquartile range ROS). Effect modification by neighborhood-level measures of racialized group membership and socioeconomic status was also identified.Conclusions: Examination of neighborhood characteristics associated with COVID-19 incidence can identify inequalities and generate hypotheses for future studies.

The Neural Network Assisted Land Use Regression

Land Use Regression (LUR) is one of the air quality assessment modelling techniques. Its advantages lie mainly in a much simpler mathematical apparatus, quicker and simpler calculations, and a possibility to incorporate more factors affecting pollutant concentration than standard dispersion models. The goal of the study was to perform the LUR model in the Polish-Czech-Slovakian Tritia region, to test two sets of pollution data input factors, i.e., factors based on emission data and pollution dispersion model results, to test regression via neural networks and compare it with standard linear regression. Both input datasets, emission data and pollution dispersion model results, provided a similar quality of results in the case when standard linear regression was used, the R2 of the models was 0.639 and 0.652. Neural network regression provided a significantly higher quality of the models, their R2 was 0.937 and 0.938 for the factors based on emission data and pollution dispersion model results respectively.

Development and intercity transferability of land-use regression models for predicting ambient PM<sub>10</sub>, PM<sub>2.5</sub>, NO<sub>2</sub> and O<sub>3</sub> concentrations in northern Taiwan

Abstract. To provide long-term air pollutant exposure estimates for epidemiological studies, it is essential to test the feasibility of developing land-use regression (LUR) models using only routine air quality measurement data and to evaluate the transferability of LUR models between nearby cities. In this study, we developed and evaluated the intercity transferability of annual-average LUR models for ambient respirable suspended particulates (PM10), fine suspended particulates (PM2.5), nitrogen dioxide (NO2) and ozone (O3) in the Taipei–Keelung metropolitan area of northern Taiwan in 2019. Ambient PM10, PM2.5, NO2 and O3 measurements at 30 fixed-site stations were used as the dependent variables, and a total of 156 potential predictor variables in six categories (i.e., population density, road network, land-use type, normalized difference vegetation index, meteorology and elevation) were extracted using buffer spatial analysis. The LUR models were developed using the supervised forward linear regression approach. The LUR models for ambient PM10, PM2.5, NO2 and O3 achieved relatively high prediction performance, with R2 values of > 0.72 and leave-one-out cross-validation (LOOCV) R2 values of > 0.53. The intercity transferability of LUR models varied among the air pollutants, with transfer-predictive R2 values of > 0.62 for NO2 and < 0.56 for the other three pollutants. The LUR-model-based 500 m × 500 m spatial-distribution maps of these air pollutants illustrated pollution hot spots and the heterogeneity of population exposure, which provide valuable information for policymakers in designing effective air pollution control strategies. The LUR-model-based air pollution exposure estimates captured the spatial variability in exposure for participants in a cohort study. This study highlights that LUR models can be reasonably established upon a routine monitoring network, but there exist uncertainties when transferring LUR models between nearby cities. To the best of our knowledge, this study is the first to evaluate the intercity transferability of LUR models in Asia.

Agglomeration and infrastructure effects in land use regression models for air pollution – Specification, estimation, and interpretations

Established land use regression (LUR) techniques such as linear regression utilize extensive selection of predictors and functional form to fit a model for every data set on a given pollutant. In this paper, an alternative to established LUR modeling is employed, which uses additive regression smoothers. Predictors and functional form are selected in a data-driven way and ambiguities resulting from specification search are mitigated. The approach is illustrated with nitrogen dioxide (NO2) data from German monitoring sites using the spatial predictors longitude, latitude, altitude and structural predictors; the latter include population density, land use classes, and road traffic intensity measures. The statistical performance of LUR modeling via additive regression smoothers is contrasted with LUR modeling based on parametric polynomials. Model evaluation is based on goodness of fit, predictive performance, and a diagnostic test for remaining spatial autocorrelation in the error terms. Additionally, interpretation and counterfactual analysis for LUR modeling based on additive regression smoothers are discussed.Our results have three main implications for modeling air pollutant concentration levels: First, modeling via additive regression smoothers is supported by a specification test and exhibits superior in- and out-of-sample performance compared to modeling based on parametric polynomials. Second, different levels of prediction errors indicate that NO2 concentration levels observed at background and traffic/industrial monitoring sites stem from different processes. Third, accounting for agglomeration and infrastructure effects is important: NO2 concentration levels tend to increase around major cities, surrounding agglomeration areas, and their connecting road traffic network.

Characterizing long-term NO2 concentration surfaces across a large metropolitan area through spatiotemporal land use regression modelling of mobile measurements

A spatiotemporal land use regression (LUR) model optimized to predict nitrogen dioxide (NO2) concentrations obtained from on-road, mobile measurements collected in 2015–16 was independently evaluated using concentrations observed at multiple sites across Toronto, Canada, obtained more than ten years earlier. This spatiotemporal LUR modelling approach improves upon estimates of historical NO2 concentrations derived from the previously used method of back-extrapolation. The optimal spatiotemporal LUR model (R2 = 0.71 for prediction of NO2 data in 2002 and 2004) uses daily average NO2 concentrations observed at multiple long-term monitoring sites and hourly average wind speed recorded at a single site, along with spatial predictors based on geographical information system data, to estimate NO2 levels for time periods outside of those used for model development. While the model tended to underestimate samplers located close to the roadway, it showed great accuracy when estimating samplers located beyond 100 m which are probably more relevant for exposure at residences. This study shows that spatiotemporal LUR models developed from strategic, multi-day (30 days in 3 different months) mobile measurements can enhance LUR model's ability to estimate long-term, intra-urban NO2 patterns. Furthermore, the mobile sampling strategy enabled this new LUR model to cover a larger domain of Toronto and outlying suburban communities, thereby increasing the potential population for future epidemiological studies.

Outdoor air pollution and hormone-assessed pubertal development in children: Results from the GINIplus and LISA birth cohorts

BackgroundAir pollution is hypothesized to affect pubertal development. However, the few studies on this topic yielded overall mixed results. These studies did not consider important pollutants like ozone, and none of them involved pubertal development assessed by estradiol and testosterone measurements. We aimed to analyze associations between long-term exposure to four pollutants and pubertal development based on sex hormone concentrations among 10-year-old children. MethodsThese cross-sectional analyses were based on the 10-year follow-up medical examinations of 1945 children from the Munich and Wesel centers of the GINIplus and LISA German birth cohorts. Female and male pubertal development was assessed by dichotomizing the concentration of hormones in serum at 18.4 pmol/L and 0.087 nmol/L using the lower limits of quantification for estradiol and testosterone, respectively. Land-use regression models derived annual average concentrations of particulate matter with an aerodynamic diameter < 2.5 and 10 µm (PM2.5 and PM10), as well as spatial models assessed yearly average concentrations of nitrogen dioxide (NO2) and ozone, were calculated at the 10-year residential addresses. To evaluate associations, we utilized logistic regressions adjusted for potential covariates. The analyses were stratified by area and sex. ResultsAround 73% of the 943 females and 25% of the 1002 males had a high level of hormones and had already started puberty at the age of 10. Overall, we found no statistically significant associations between exposure to particles (PM2.5 or PM10) and pubertal development. Results on NO2 and ozone were not significant as well; for instance, per 10 µg/m3 increase in ozone concentration, odds ratios and 95% confidence intervals were 0.900 (0.605, 1.339) and 0.830 (0.573, 1.203) for females and males, respectively. Stratified by area, the aforementioned results did not reveal any associations either. ConclusionsOur study did not observe the associations between ambient air pollutants and pubertal development determined by estradiol and testosterone levels in children. However, due to the current limited number of studies on this topic, our results should be cautiously interpreted. Future longitudinal studies are needed to assess the association.

The application of land use regression model to investigate spatiotemporal variations of PM2.5 in Guangzhou, China: Implications for the public health benefits of PM2.5 reduction.

Understanding the intra-city variation of PM2.5 is important for air quality management and exposure assessment. In this study, to investigate the spatiotemporal variation of PM2.5 in Guangzhou, we developed land use regression (LUR) models using data from 49 routine air quality monitoring stations. The R2, adjust R2 and 10-fold cross validation R2 for the annual PM2.5 LUR model were 0.78, 0.72 and 0.66, respectively, indicating the robustness of the model. In all the LUR models, traffic variables (e.g., length of main road and the distance to nearest ancillary) were the most common variables in the LUR models, suggesting vehicle emission was the most important contributor to PM2.5 and controlling vehicle emissions would be an effective way to reduce PM2.5. The predicted PM2.5 exhibited significant variations with different land uses, with the highest value for impervious surfaces, followed by green land, cropland, forest and water areas. Guangzhou as the third largest city that PM2.5 concentration has achieved CAAQS Grade II guideline in China, it represents a useful case study city to examine the health and economic benefits of further reduction of PM2.5 to the lower concentration ranges. So, the health and economic benefits of reducing PM2.5 in Guangzhou was further estimated using the BenMAP model, based on the annual PM2.5 concentration predicted by the LUR model. The results showed that the avoided all cause mortalities were 992 cases (95% CI: 221-2140) and the corresponding economic benefits were 1478 million CNY (95% CI: 257-2524) (willingness to pay approach) if the annual PM2.5 concentration can be reduced to the annual CAAQS Grade I guideline value of 15μg/m3. Our results are expected to provide valuable information for further air pollution control strategies in China.

A Population-Based Cohort Study of Respiratory Disease and Long-Term Exposure to Iron and Copper in Fine Particulate Air Pollution and Their Combined Impact on Reactive Oxygen Species Generation in Human Lungs.

Metal components in fine particulate matter (PM2.5) from nontailpipe emissions may play an important role in underlying the adverse respiratory effects of PM2.5. We investigated the associations between long-term exposure to iron (Fe) and copper (Cu) in PM2.5 and their combined impact on reactive oxygen species (ROS) generation in human lungs, and the incidence of asthma, chronic obstructive pulmonary disease (COPD), COPD mortality, pneumonia mortality, and respiratory mortality. We conducted a population-based cohort study of ∼0.8 million adults in Toronto, Canada. Land-use regression models were used to estimate the concentrations of Fe, Cu, and ROS. Outcomes were ascertained using validated health administrative databases. We found positive associations between long-term exposure to Fe, Cu, and ROS and the risks of all five respiratory outcomes. The associations were more robust for COPD, pneumonia mortality, and respiratory mortality than for asthma incidence and COPD mortality. Stronger associations were observed for ROS than for either Fe or Cu. In two-pollutant models, adjustment for nitrogen dioxide somewhat attenuated the associations while adjustment for PM2.5 had little influence. Long-term exposure to Fe and Cu in PM2.5 and estimated ROS concentration in lung fluid was associated with increased incidence of respiratory diseases, suggesting the adverse respiratory effects of nontailpipe emissions.

Development of land-use regression models to estimate particle mass and number concentrations in Taichung, Taiwan

Land-use regression (LUR) models have been used to estimate particle mass concentration (PMC), but few studies apply it to predict particle number concentration (PNC) at different sizes. This study aimed to determine both PMC and PNC throughout one year to establish predictive models in Taichung, Taiwan. The annual averages of PM10, PM2.5, and PM1 were 71 ± 46 μg/m3, 44 ± 35 μg/m3, and 32 ± 28 μg/m3, respectively. The PNC at size ranges of <0.5 μm, 0.5–1 μm, 1–2.5 μm, 2.5–10 μm, and ≥10 μm were 715098 ± 664879 counts/L, 29053 ± 30615 counts/L, 1009 ± 659 counts/L, 647 ± 347 counts/L, and 3 ± 3 counts/L, respectively. The model-explained variance (R2) values of PM10, PM2.5, and PM1 were 0.42, 0.53, and 0.51, respectively. The magnitude of the R2 values ranged from 0.31 to 0.50 for the PNC with the highest R2 between 0.5 and 1 μm. The differences between the model R2 and the leave-one-out cross-validation R2 ranged from 4% to 8% for PMC and from 3% to 10% for PNC. This study developed LUR models with moderate performance to estimate PMC and PNC at different sizes in an Asian metropolis. The built LUR models may be improved by combining with other open data to increase the predictive capacity.

Using a land use regression model with machine learning to estimate ground level PM2.5.

Ambient fine particulate matter (PM2.5) has been ranked as the sixth leading risk factor globally for death and disability. Modelling methods based on having access to a limited number of monitor stations are required for capturing PM2.5 spatial and temporal continuous variations with a sufficient resolution. This study utilized a land use regression (LUR) model with machine learning to assess the spatial-temporal variability of PM2.5. Daily average PM2.5 data was collected from 73 fixed air quality monitoring stations that belonged to the Taiwan EPA on the main island of Taiwan. Nearly 280,000 observations from 2006 to 2016 were used for the analysis. Several datasets were collected to determine spatial predictor variables, including the EPA environmental resources dataset, a meteorological dataset, a land-use inventory, a landmark dataset, a digital road network map, a digital terrain model, MODIS Normalized Difference Vegetation Index (NDVI) database, and a power plant distribution dataset. First, conventional LUR and Hybrid Kriging-LUR were utilized to identify the important predictor variables. Then, deep neural network, random forest, and XGBoost algorithms were used to fit the prediction model based on the variables selected by the LUR models. Data splitting, 10-fold cross validation, external data verification, and seasonal-based and county-based validation methods were used to verify the robustness of the developed models. The results demonstrated that the proposed conventional LUR and Hybrid Kriging-LUR models captured 58% and 89% of PM2.5 variations, respectively. When XGBoost algorithm was incorporated, the explanatory power of the models increased to 73% and 94%, respectively. The Hybrid Kriging-LUR with XGBoost algorithm outperformed the other integrated methods. This study demonstrates the value of combining Hybrid Kriging-LUR model and an XGBoost algorithm for estimating the spatial-temporal variability of PM2.5 exposures.

Assessing the Distribution of Air Pollution Health Risks within Cities: A Neighborhood-Scale Analysis Leveraging High-Resolution Data Sets in the Bay Area, California.

Background:Air pollution-attributable disease burdens reported at global, country, state, or county levels mask potential smaller-scale geographic heterogeneity driven by variation in pollution levels and disease rates. Capturing within-city variation in air pollution health impacts is now possible with high-resolution pollutant concentrations.Objectives:We quantified neighborhood-level variation in air pollution health risks, comparing results from highly spatially resolved pollutant and disease rate data sets available for the Bay Area, California.Methods:We estimated mortality and morbidity attributable to nitrogen dioxide (), black carbon (BC), and fine particulate matter [PM in aerodynamic diameter ()] using epidemiologically derived health impact functions. We compared geographic distributions of pollution-attributable risk estimates using concentrations from a) mobile monitoring of and BC; and b) models predicting annual , BC and concentrations from land-use variables and satellite observations. We also compared results using county vs. census block group (CBG) disease rates.Results:Estimated pollution-attributable deaths per 100,000 people at the grid-cell level ranged across the Bay Area by a factor of 38, 4, and 5 for [ (95% CI: 9, 50)], BC [ (95% CI: 1, 2)], and , [ (95% CI: 33, 64)]. Applying concentrations from mobile monitoring and land-use regression (LUR) models in Oakland neighborhoods yielded similar spatial patterns of estimated grid-cell–level mortality rates. Mobile monitoring concentrations captured more heterogeneity [mobile monitoring (95% CI: 19, 107) deaths per 100,000 people; (95% CI: 30, 167)]. Using CBG-level disease rates instead of county-level disease rates resulted in 15% larger attributable mortality rates for both and , with more spatial heterogeneity at the grid-cell–level [ CBG deaths per 100,000 people (95% CI: 12, 68); (95% CI: 11, 64); (95% CI: 40, 77); and (95% CI: 37, 71)].Discussion:Air pollutant-attributable health burdens varied substantially between neighborhoods, driven by spatial variation in pollutant concentrations and disease rates. https://doi.org/10.1289/EHP7679

Evaluating the multi-task learning approach for land use regression modelling of air pollution

Air pollution has been linked to several health problems including heart disease, stroke and lung cancer. Modelling and analyzing this dependency requires reliable and accurate air pollutant measurements collected by stationary air monitoring stations. However, usually only a low number of such stations are present within a single city. To retrieve pollution concentrations for unmeasured locations, researchers rely on land use regression (LUR) models. Those models are typically developed for one pollutant only. However, as results in different areas have shown, modelling several related output variables through multi-task learning can improve the prediction results of the models significantly. In this work, we compared prediction results from singletask and multi-task learning multilayer perceptron models on measurements taken from the OpenSense dataset and the London Atmospheric Emissions Inventory dataset. LUR features were generated from OpenStreetMap using OpenLUR and used to train hard parameter sharing multilayer perceptron models. The results show multi-task learning with sufficient data significantly improves the performance of a LUR model.

Incorporating land-use regression into machine learning algorithms in estimating the spatial-temporal variation of carbon monoxide in Taiwan

This paper is the first of its kind to use machine learning algorithms in conjunction with a Land-use Regression (LUR) model for predicting the spatiotemporal variation of CO concentrations in Taiwan. We used daily CO concentration from 2000 to 2016 to develop model and data from 2017 to 2018 as external data to verify the model reliability. Location of temples was used as a predictor to account for Asian culturally specific sources. With the ability to capture nonlinear relationship between observations and predictions, three LUR-based machine learning algorithms were used to estimate CO concentrations, including deep neural network (DNN), random forest (RF), and extreme gradient boosting (XGBoost). The results showed that LUR-based machine-learning model (LUR-XGBoost) has the best computation efficiency and improved adjusted R2 from 0.69 to 0.85. Our studies demonstrate the ability of the LUR-based machine learning algorithms to estimate long-term spatiotemporal CO concentration variations in fine resolution.

The mediating role of lung function on air pollution-induced cardiopulmonary mortality in elderly women: The SALIA cohort study with 22-year mortality follow-up.

Air pollution exposure is associated with reduced lung function and increased cardio-pulmonary mortality (CPM). We analyzed the potential mediating effect of reduced lung function on the association between air pollution exposure and CPM. We used data from the German SALIA cohort including 2527 elderly women (aged 51-56yearsat baseline 1985-1994) with 22-year follow-up to CPM. Exposures to PM10, PM2.5, PM2.5 absorbance, NO2 and NOx were assessed by land-use regression modelling and back-extrapolated to estimate exposures at baseline. Lung function (FVC, FEV1) was measured by spirometry and transformed to GLI z-scores. Adjusted Cox proportional hazards and causal proportional hazards mediation analysis models were fitted. The survival analysis showed that reduced lung function (z-scores of FVC or FEV1 below 5% predicted) reflected significantly lower survival probability from CPM (p<0.0001). Longterm exposures to NOx and NO2 were associated with increased risks of CPM (eg. HR=1.215; 95%CI: 1.017-1.452 for IQR increase in NOx and HR=1.209; 95%CI: 1.011-1.445 for IQR increase in NO2) after adjusting for reduced lung function and additional covariates. The associations of PM2.5 absorbance and CPM remained significant in models adjusted for FEV1/FVC, but the associations with PM10 and PM2.5 were not significant. The mediation analysis showed significant indirect effects of NO2 and NOx on CPM mediated through reduced FEV1 and FVC. The largest indirect effects were found for exposures to NO2 (HR=1.037; 95%CI: 1.005-1.070) and NOx (HR=1.028; 95%CI: 1.004-1.052) mediated through reduced FVC. The mediated proportion effect ranged from 13.9% to 19.6% in fully adjusted models. This study provides insights into the mechanism of reduced lung function in association between long-term air pollution exposure and CPM. The mediated effect was substantial for exposure to nitrogen oxides (NOx and NO2), but less pronounced for PM10 and PM2.5.

Using Street View Imagery to Predict Street-Level Particulate Air Pollution.

Land-use regression (LUR) models are frequently applied to estimate spatial patterns of air pollution. Traditional LUR often relies on fixed-site measurements and GIS-derived variables with limited spatial resolution. We present an approach that leverages Google Street View (GSV) imagery to predict street-level particulate air pollution (i.e., black carbon [BC] and particle number [PN] concentrations). We developed empirical models based on mobile monitoring data and features extracted from ∼52 500 GSV images using a deep learning model. We tested theory- and data-driven feature selection methods as well as models using images within varying buffer sizes (50-2000 m). Compared to LUR models with traditional variables, our models achieved similar model performance using the street-level predictors while also identifying additional potential hotspots. Adjusted R2 (10-fold CV R2) with integrated feature selection was 0.57-0.64 (0.50-0.57) and 0.65-0.73 (0.61-0.66) for BC and PN models, respectively. Models using only features near the measurement locations (i.e., GSV images within 250 m) explained ∼50% of air pollution variability, indicating PN and BC are strongly affected by the street-level built environment. Our results suggest that GSV imagery, processed with computer vision techniques, is a promising data source to develop LUR models with high spatial resolution and consistent predictor variables across administrative boundaries.

Long-term trends in local and transported PM2.5 pollution in New York City

Previous source apportionment analyses of New York City (NYC) PM2.5 have found that transported pollution can account for up to half of total fine particulate matter (PM2.5) mass. In recent years, several local and federal emissions regulations on both the transportation and energy sector have been introduced, such as the Clean Heavy-Duty Bus and Truck Rule, and policies related to the NYC Clean Heat Program. Since the implementation of these regulations, there have been limited studies on regional transport of particulate air pollution into NYC. This paper aims to quantify long-term changes in local and regional contributions to PM2.5 in NYC. The composition of PM2.5 in NYC over the period 2002–2018 was evaluated using data from the US Environmental Protection Agency (EPA) Chemical Speciation Network (CSN) and the New York City Community Air Survey (NYCCAS). Background site data obtained from EPA CSN was used to assess regional transport. Two estimates of transported PM2.5 were calculated, one assuming that transported PM2.5 consists solely of ammonium sulfate and ammonium nitrate as a lower bound and the other taking into account organic matter (OM) as an upper bound. Using this method, regional transport was estimated to account for 25–46% of total PM2.5 measured in NYC in 2018, down from 46 to 57% in 2002. Secondary PM2.5 output from the Community MultiScale Air Quality (CMAQ) model for the years 2008, 2011, and 2017 was used as an additional upper-bound estimate of regional transport (56%, 35%, and 30%, respectively). In addition to CMAQ model estimates of sulfate, nitrate, and primary emitted PM2.5, land-use regression (LUR) models for sulfur, vanadium, copper, nickel, and PM2.5 measurements at 60 locations across NYC during the winter and summer seasons of 2008–2015 were assessed for changes in local emissions. The results indicate decreased emissions from residual oil burning, traffic, and ships. As a consequence of regional and local changes, PM2.5 in NYC had fallen to approximately 55% of its 2002 levels by 2018. The observed temporal and spatial trends of PM2.5 and its components point to the effect of regulations as well as economic influences on fuel usage, such as the decline of coal in favor of natural gas, in reducing air pollution over the last two decades. While other major PM2.5 constituents have declined, OM is becoming an increasingly dominant component with potential implications for PM2.5 toxicity. Further improvements in NYC air quality could be achieved by targeting sources of ambient OM, such as commercial cooking, traffic, and biomass burning for residential heating.

From PM2.5 exposure to PM2.5 risks of inhaled dose in daily activities: Empirical evidence during workdays from guangzhou, China

PM2.5 pollution poses a serious health risk to residents. In recent years, individual exposure to PM2.5 in daily activities has received much attention, but the impact of different activities at different times and spaces on PM2.5-triggered individual health risks has been ignored. In this paper, the residents' activity diaries and the data on the spatial-temporal changes in urban PM2.5 concentration obtained through the land-use regression model (LUR) were coupled to dynamically evaluate and analyze the characteristics of Guangzhou residents' PM2.5 inhaled dose in 24 h during workdays. The study showed that the residents' PM2.5 inhaled dose in daily activities is about 940.2 ± 94.5 μg (95% CI). This can better reflect the individual differences than the exposure to PM2.5 in daily activities. For every increase of 1 μg/day in the exposure to PM2.5, the inhaled dose will increase by 1.5 μg/day. The obvious spatial-temporal characteristics of PM2.5 inhaled dose are closely related to people's daily activities. The inhaled dose is high during the day and low at night, and the individual differences in inhaled dose in the commuting and leisure period is significant but is small during the working and sleeping period; if the individual inhaled dose exceeds the population's average for accumulated 7 h in a day, the probability of a high risk of the all-day inhaled dose increases significantly. The inhaled dose at the workplace and residence as well as in light physical activities accounts for a large proportion of the total. However, although the inhaled dose on roads and in other places, and moderate physical activities account for only a small proportion, it has a great impact on the risk level of the individual total inhaled dose. Therefore, the characteristics of PM2.5 inhaled dose and the individual activity should be considered, to propose targeted policies and suggestions from the perspective of both the government and individuals to reduce the health risk caused by PM2.5 pollution.

Association of long-term exposure to ambient air pollution with the incidence of sleep disorders: A cohort study in China

BackgroundSleep disorders have been verified to be associated with adverse health outcomes. Recent studies have linked ambient air pollution to sleep disorders. However, evidence with large sample size and especially prospective studies is very limited. MethodsWe used data from a prospective cohort study established from 2015 to 2018 in Ningbo, China. Participants were followed-up after baseline survey through linkage to the regional Health Information System (HIS). Sleep disorders were defined based on International Classification of Disease 10 (ICD-10). Spatial-temporal land-use regression (LUR) models were used to estimate the annual exposure to particulate matter with diameter ≤ 2.5 µm (PM2.5), ≤ 10 µm (PM10) and nitrogen dioxides (NO2). The associations between long-term exposure to air pollutants and prevalence of sleep disorders were examined using logistic regression models, and Cox regression models for the effects of air pollution on the incidence of sleep disorders. A generalized weighted quantile sum (gWQS) regression was used in the multipollutant analysis. ResultsA total of 38,775 participants were included in the final analysis. Based on baseline data, we observed significant positive associations between air pollution exposure and increased odds of prevalent sleep disorders (Odds Ratio (OR)= 1.48, 95% confidence interval (CI): 1.41–1.55 for PM2.5; OR= 1.47,95%CI:1.38–1.57 for PM10; OR= 1.38, 95%CI:1.31–1.46 for NO2). In the longitudinal analysis, hazard ratios for incident sleep disorders associated with per interquartile range (IQR) increase in PM2.5, PM10 and NO2 were 1.14 (1.03, 1.25), 1.13 (1.01, 1.27) and 1.13 (1.04, 1.23), respectively. A gWQS regression analysis showed significant association between air pollution mixture and incident sleep disorders (OR=1.11, 95%CI: 1.03–1.20). ConclusionsLong-term exposure to PM2.5, PM10 and NO2 were associated with increased risk of sleep disorders in a Chinese population. Our findings could provide evidence for a more general role in the adverse health impact of air pollution.