Spatial–temporal and cancer risk assessment of selected hazardous air pollutants in Seattle

https://doi.org/10.1016/j.envint.2010.06.006Get rights and content

Abstract

In the Seattle Air Toxics Monitoring Pilot Program, we measured 15 hazardous air pollutants (HAPs) at 6 sites for more than a year between 2000 and 2002. Spatial–temporal variations were evaluated with random-effects models and principal component analyses. The potential health risks were further estimated based on the monitored data, with the incorporation of the bootstrapping technique for the uncertainty analysis. It is found that the temporal variability was generally higher than the spatial variability for most air toxics. The highest temporal variability was observed for tetrachloroethylene (70% temporal vs. 34% spatial variability). Nevertheless, most air toxics still exhibited significant spatial variations, even after accounting for the temporal effects. These results suggest that it would require operating multiple air toxics monitoring sites over a significant period of time with proper monitoring frequency to better evaluate population exposure to HAPs. The median values of the estimated inhalation cancer risks ranged between 4.3 × 10 5 and 6.0 × 10 5, with the 5th and 95th percentile levels exceeding the 1 in a million level. VOCs as a whole contributed over 80% of the risk among the HAPs measured and arsenic contributed most substantially to the overall risk associated with metals.

Introduction

Hazardous air pollutants (HAPs, or ‘air toxics’) are pollutants that are known or suspected to cause adverse health effects, including cancer, reproductive, immunological, developmental, and neurological effects (USEPA, 2004). There are currently 188 HAPs regulated under the US federal Clean Air Act Amendments of 1990. To identify those air toxics which are of greatest concern in terms of contribution to population risk, the USEPA established the National-scale Air Toxics Assessment (NATA) projects (USEPA, 2002, USEPA, 2006, USEPA, 2009d). Several studies also performed additional risk assessments and modeling evaluation based on the predicted exposure concentrations from NATA (Loh et al., 2007, Ozkaynak et al., 2008, Woodruff et al., 2000). The NATA projects relied heavily on the modeling approach. The accuracy of the concentration predictions and risk estimates depends on the completeness of the National Emissions Inventory and meteorological data. It has been reported that the modeling results for most HAPs were underestimated by a factor of 2 or more (Payne-Sturges et al., 2004, USEPA, 2006).

Several studies have been conducted to evaluate the population exposure to HAPs using limited ambient monitoring data. At the national scale, Touma et al. (2006) summarized the data collected under the National Air Toxics Trends Stations monitoring network. McCarthy et al. (2009) compiled ambient measurements of air toxics collected in the US from 2003 through 2005 and calculated risk-weighted concentrations. They found that concentrations for benzene, 1,3-butadiene, carbon tetrachloride, acetaldehyde, and arsenic were above levels of concern for cancer risks at most monitoring locations. At the regional scale, air quality measurements of volatile organic compounds (VOCs) were collected at 13 urban locations in the eastern US (Mohamed et al., 2002). The results showed that levels of carbonyls were higher than those for other organic compounds groups, especially for formaldehyde and acetaldehyde. Under the Urban Air Toxics Monitoring Program, Bortnick and Stetzer (2002) showed that temporal variability comprises most of the overall variability across 12 cities. In spite of these findings, questions about the spatial and temporal variability within individual cities still remain. Not many studies investigate this issue at a local scale and they tend to focus on analyzing VOCs, but not on particulate matter (PM) elements (Mohamed et al., 2002).

Seattle was one of six cities selected by the U.S. EPA to participate in the National Air Toxics Monitoring Pilot Program. VOCs and PM elements were monitored at six sites within Seattle for a year. This dataset provides unique opportunities to evaluate the intraurban variation of air toxics. In this study, we quantify the relative contribution of the temporal and spatial components to overall data variability. This type of analysis is useful when designing monitoring networks for air toxics. We further estimate the potential health risk based on the monitored data.

Section snippets

Study design

The monitoring sites included neighborhood to urban scale locations in distinctly different sub-regions within the Seattle metropolitan area selected to allow the evaluation of spatial variability. These included Beacon Hill (BH, regional scale), Georgetown (GT, industrial), Lake Samamish (LS, Background), Lake Forest Park (LF, neighborhood, wood smoke impact site), SeaTac (ST, mobile and airport), and Mapleleaf reservoir (ML, neighborhood) (Fig. 1). The VOCs measured included benzene,

Quality control and summary statistics

The method detection limit (MDL) for metals analyzed with XRF ranged between 0.50 ng/m3 for nickel and 4.21 ng/m3 for cadmium. The MDL for ICP-MS (Table 1) was 1.3 (chromium) to 145 (cadmium) times more sensitive than those of XRF. Collocated samples, one using the FRM for PM2.5 on a Teflon filter subject to XRF analysis and the other using the High-Vol sampler for TSP on a quartz filter subject to ICP-MS analysis, were collected between August 2001 and January 2002 at the ML site (N = 22 pairs

Conclusions

This study evaluates the spatial and temporal variations of 15 air toxics in Seattle, WA. It was found that the associated temporal component of variation was generally larger than the spatial component. The principal component analysis also shows that the temporal factor was more dominant. Nevertheless, differences among the locations after accounting for temporal effects cannot be denied. These results suggest that operating multiple air toxics monitoring sites over a significant period of

Acknowledgements

The Seattle monitoring study was conducted as a collaborative effort among the US EPA Region X, the Washington State Department of Ecology (the Ecology), the Puget Sound Clean Air Agency, the Washington State University, and the University of Washington. This study was funded by the Ecology under a cooperative agreement with the Washington Cooperative Fish & Wildlife Research Unit. This study was also partially funded by the U.S. Environmental Protection Agency through its Office of Research

References (35)

  • CalEPA

    Air toxics hot spots program risk assessment guidelines: part II technical support document for describing available cancer potency factors

    (2005)
  • B. Efron et al.

    An introduction to the bootstrap

    (1998)
  • D.E. Johnson

    Applied multivariate methods for data analysts

    (1998)
  • E. Kim et al.

    Estimation of organic carbon blank values and error structures of the speciation trends network data for source apportionment

    J Air Waste Manage Assoc

    (2005)
  • D.G. Kleinbaum et al.

    Applied regression analysis and other multivariable methods

    (1988)
  • L.J.S. Liu et al.

    Exposure assessment of particulate matter for susceptible populations in Seattle

    Environ Health Perspect

    (2003)
  • M.M. Loh et al.

    Ranking cancer risks of organic hazardous air pollutants in the United States

    Environ Health Perspect

    (2007)
  • Cited by (24)

    • Estimating future industrial emissions of hazardous air pollutants in the United States using the National Energy Modeling System (NEMS)

      2021, Resources, Conservation and Recycling
      Citation Excerpt :

      Emissions inventories and air quality modeling related to these emissions have been performed to understand the environmental fate and transport of HAPs (Mohamed et al., 2002; Ohura et al., 2006; Shie and Chan, 2013; Vennam et al., 2015; Wu et al., 2011). Urban and regional-level air quality modeling studies have revealed substantial spatiotemporal variations both in emissions (Tian et al., 2012) and ambient concentrations (Wu et al., 2011). At a national level, the U.S. National Air Toxics Assessment (NATA) is conducted periodically by the EPA (U.S. EPA, 2018).

    • Characterization and health impact assessment of hazardous air pollutants in residential areas near a large iron-steel industrial complex in Korea

      2020, Atmospheric Pollution Research
      Citation Excerpt :

      The data were then used in a risk-based prioritization of HAPs. To identify which HAPs are most likely to cause significant health impacts to the residents, we applied a deterministic risk assessment approach (Bari and Kindzierski, 2017, 2018; Fox et al., 2004; Jia and Foran, 2013; Tam and Neumann, 2004; Wu et al., 2009, 2011). The potential public health risk was evaluated for chronic inhalation exposure to HAPs, including cancer and non-cancer effects.

    • A streamlined risk screening method for managing reutilization of abandoned factories in Taiwan

      2017, Sustainable Environment Research
      Citation Excerpt :

      To facilitate evaluation of the sustainability of contaminated land reutilization, many useful tools and decision-making models have been established [13–16]. Geographic information system (GIS) has been recognized as valuable for land management, health risk, and environmental quality assessments to address regional spatial variability [17–20]. Additional research has classified spatial variability to characterize the vulnerable zone of potential contamination resulting in environmental conditions combing the decision tree and spatial statistics [21–23].

    • Physiochemical properties of carbonaceous aerosol from agricultural residue burning: Density, volatility, and hygroscopicity

      2016, Atmospheric Environment
      Citation Excerpt :

      8 quantified mineral elements account 0.1–0.9 wt% of smoke particles, and the mineral elements reside more in PM2.5 (details in Table S1-S2, SI). As is the second largest metallic emission followed by Cr, Ni, Zn from agricultural residue burning, which present potential carcinogenic risk and impose a long term burden on biogeochemical cycling (Duan et al., 2013; Wu et al., 2011). The calculated ensemble density of cotton, soybean, corn, rice, and wheat residue burning smoke is 1.317 ± 0.013, 1.194 ± 0.016, 1.346 ± 0.015, 1.301 ± 0.035, and 1.286 ± 0.033 g cm−3, respectively.

    View all citing articles on Scopus
    1

    Contributed equally to this work.

    View full text