Limits...
A comparison of exposure metrics for traffic-related air pollutants: application to epidemiology studies in Detroit, Michigan.

Batterman S, Burke J, Isakov V, Lewis T, Mukherjee B, Robins T - Int J Environ Res Public Health (2014)

Bottom Line: While showing some agreement, the simple categorical and proximity classifications (e.g., high diesel/low diesel traffic roads and distance from major roads) do not reflect the range and overlap of exposures seen in the other metrics.Information provided by the traffic density metric, defined as the number of kilometers traveled (VKT) per day within a 300 m buffer around each home, was reasonably consistent with the more sophisticated metrics.Dispersion modeling provided spatially- and temporally-resolved concentrations, along with apportionments that separated concentrations due to traffic emissions and other sources.

View Article: PubMed Central - PubMed

Affiliation: Department of Environmental Health Sciences, School of Public Health, University of Michigan, 1420 Washington Heights, Ann Arbor, MI 48109, USA. stuartb@umich.edu.

ABSTRACT
Vehicles are major sources of air pollutant emissions, and individuals living near large roads endure high exposures and health risks associated with traffic-related air pollutants. Air pollution epidemiology, health risk, environmental justice, and transportation planning studies would all benefit from an improved understanding of the key information and metrics needed to assess exposures, as well as the strengths and limitations of alternate exposure metrics. This study develops and evaluates several metrics for characterizing exposure to traffic-related air pollutants for the 218 residential locations of participants in the NEXUS epidemiology study conducted in Detroit (MI, USA). Exposure metrics included proximity to major roads, traffic volume, vehicle mix, traffic density, vehicle exhaust emissions density, and pollutant concentrations predicted by dispersion models. Results presented for each metric include comparisons of exposure distributions, spatial variability, intraclass correlation, concordance and discordance rates, and overall strengths and limitations. While showing some agreement, the simple categorical and proximity classifications (e.g., high diesel/low diesel traffic roads and distance from major roads) do not reflect the range and overlap of exposures seen in the other metrics. Information provided by the traffic density metric, defined as the number of kilometers traveled (VKT) per day within a 300 m buffer around each home, was reasonably consistent with the more sophisticated metrics. Dispersion modeling provided spatially- and temporally-resolved concentrations, along with apportionments that separated concentrations due to traffic emissions and other sources. While several of the exposure metrics showed broad agreement, including traffic density, emissions density and modeled concentrations, these alternatives still produced exposure classifications that differed for a substantial fraction of study participants, e.g., from 20% to 50% of homes, depending on the metric, would be incorrectly classified into "low", "medium" or "high" traffic exposure classes. These and other results suggest the potential for exposure misclassification and the need for refined and validated exposure metrics. While data and computational demands for dispersion modeling of traffic emissions are non-trivial concerns, once established, dispersion modeling systems can provide exposure information for both on- and near-road environments that would benefit future traffic-related assessments.

Show MeSH

Related in: MedlinePlus

Map of modeled road network in study area, and locations of 218 homes of participants in NEXUS. Shaded area defines city of Detroit and population by Census Block group. Axis scales are Universal Traverse Mercator coordinates (m). AADT is annual average daily traffic (vehicles/day). Highlighted roads are National Functional Class 11, called high diesel/high traffic roads in NEXUS. Windsor, Canada (not shown), is immediately to the south-east.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4199035&req=5

ijerph-11-09553-f001: Map of modeled road network in study area, and locations of 218 homes of participants in NEXUS. Shaded area defines city of Detroit and population by Census Block group. Axis scales are Universal Traverse Mercator coordinates (m). AADT is annual average daily traffic (vehicles/day). Highlighted roads are National Functional Class 11, called high diesel/high traffic roads in NEXUS. Windsor, Canada (not shown), is immediately to the south-east.

Mentions: Ultimately, 139 children were recruited and participated in the study from September 2010 to December 2012. Because a number of children moved during the study, a total of 218 residence locations were considered (Figure 1). The study population had approximately equal distribution across the three traffic categories. The population was primarily minority (non-Hispanic Blacks constituted 82% of the participants, Hispanics 8%, non-Hispanic Whites 4%, and other/multiracial 6%). Many households were poor (about one-third of families reported annual household incomes below $15,000).


A comparison of exposure metrics for traffic-related air pollutants: application to epidemiology studies in Detroit, Michigan.

Batterman S, Burke J, Isakov V, Lewis T, Mukherjee B, Robins T - Int J Environ Res Public Health (2014)

Map of modeled road network in study area, and locations of 218 homes of participants in NEXUS. Shaded area defines city of Detroit and population by Census Block group. Axis scales are Universal Traverse Mercator coordinates (m). AADT is annual average daily traffic (vehicles/day). Highlighted roads are National Functional Class 11, called high diesel/high traffic roads in NEXUS. Windsor, Canada (not shown), is immediately to the south-east.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4199035&req=5

ijerph-11-09553-f001: Map of modeled road network in study area, and locations of 218 homes of participants in NEXUS. Shaded area defines city of Detroit and population by Census Block group. Axis scales are Universal Traverse Mercator coordinates (m). AADT is annual average daily traffic (vehicles/day). Highlighted roads are National Functional Class 11, called high diesel/high traffic roads in NEXUS. Windsor, Canada (not shown), is immediately to the south-east.
Mentions: Ultimately, 139 children were recruited and participated in the study from September 2010 to December 2012. Because a number of children moved during the study, a total of 218 residence locations were considered (Figure 1). The study population had approximately equal distribution across the three traffic categories. The population was primarily minority (non-Hispanic Blacks constituted 82% of the participants, Hispanics 8%, non-Hispanic Whites 4%, and other/multiracial 6%). Many households were poor (about one-third of families reported annual household incomes below $15,000).

Bottom Line: While showing some agreement, the simple categorical and proximity classifications (e.g., high diesel/low diesel traffic roads and distance from major roads) do not reflect the range and overlap of exposures seen in the other metrics.Information provided by the traffic density metric, defined as the number of kilometers traveled (VKT) per day within a 300 m buffer around each home, was reasonably consistent with the more sophisticated metrics.Dispersion modeling provided spatially- and temporally-resolved concentrations, along with apportionments that separated concentrations due to traffic emissions and other sources.

View Article: PubMed Central - PubMed

Affiliation: Department of Environmental Health Sciences, School of Public Health, University of Michigan, 1420 Washington Heights, Ann Arbor, MI 48109, USA. stuartb@umich.edu.

ABSTRACT
Vehicles are major sources of air pollutant emissions, and individuals living near large roads endure high exposures and health risks associated with traffic-related air pollutants. Air pollution epidemiology, health risk, environmental justice, and transportation planning studies would all benefit from an improved understanding of the key information and metrics needed to assess exposures, as well as the strengths and limitations of alternate exposure metrics. This study develops and evaluates several metrics for characterizing exposure to traffic-related air pollutants for the 218 residential locations of participants in the NEXUS epidemiology study conducted in Detroit (MI, USA). Exposure metrics included proximity to major roads, traffic volume, vehicle mix, traffic density, vehicle exhaust emissions density, and pollutant concentrations predicted by dispersion models. Results presented for each metric include comparisons of exposure distributions, spatial variability, intraclass correlation, concordance and discordance rates, and overall strengths and limitations. While showing some agreement, the simple categorical and proximity classifications (e.g., high diesel/low diesel traffic roads and distance from major roads) do not reflect the range and overlap of exposures seen in the other metrics. Information provided by the traffic density metric, defined as the number of kilometers traveled (VKT) per day within a 300 m buffer around each home, was reasonably consistent with the more sophisticated metrics. Dispersion modeling provided spatially- and temporally-resolved concentrations, along with apportionments that separated concentrations due to traffic emissions and other sources. While several of the exposure metrics showed broad agreement, including traffic density, emissions density and modeled concentrations, these alternatives still produced exposure classifications that differed for a substantial fraction of study participants, e.g., from 20% to 50% of homes, depending on the metric, would be incorrectly classified into "low", "medium" or "high" traffic exposure classes. These and other results suggest the potential for exposure misclassification and the need for refined and validated exposure metrics. While data and computational demands for dispersion modeling of traffic emissions are non-trivial concerns, once established, dispersion modeling systems can provide exposure information for both on- and near-road environments that would benefit future traffic-related assessments.

Show MeSH
Related in: MedlinePlus