Assessment of schoolchildren's exposure to traffic-related air pollution in the French Six Cities Study using a dispersion model

Abstract

The purpose of this work was to estimate exposure to traffic-related air pollution (TAP), of the 6683 schoolchildren included in a cross-sectional epidemiological study conducted in six French cities to determine the effects of urban air pollution (AP) on respiratory and allergic health.

Annual mean concentrations of benzene, CO, NO₂, NO_x, PM₁₀ and SO₂ were calculated, in front of the 108 schools attended by the children, by the validated STREET 5 software, which combines data on regional and local components of AP. STREET contains a database of emissions estimated by the IMPACT 2.0 software developed by ADEME-France and results of ambient concentrations modelled by the WinMISKAM 4.2 dispersion model. The input data required were background AP, traffic conditions (daily traffic density; average speed; percentage of gridlocks and proportion of each type of vehicle) and dispersion conditions (topography of the street segments modelled and meteorology).

Emissions of air pollutants in front of the 108 schools were considerably scattered. Calculated concentrations (μg m⁻³) also varied considerably at: [1.0–5.1] for benzene, [303.8–988.1] for CO, [17.8–78.9] for NO₂, [23.3–195.2] for NO_x, [10.0–52.0] for PM₁₀ and [2.4–16.4] for SO₂. About 64% (29%, respectively) of the schools had annual mean concentrations of NO₂ (PM₁₀, respectively) exceeding the European quality objectives (40 and 30 μg m⁻³, respectively).

These exposure indicators, capable of identifying small area variations in AP contrary to surrogate measures usually used in epidemiology, will enable better studies on the impact of urban AP on health.

Introduction

Traffic is a major source of air pollution (AP) in most urban areas and is responsible for resultant health effects at the population level (Hoek et al., 2002; Finkelstein et al., 2004; Pénard-Morand and Annesi-Maesano, 2004). Assessment of human exposure to ambient AP in urban areas is difficult, however, due to the existence of a regional component (background AP), evaluated through the monitoring of the local Air Quality Monitoring Networks (Nerriere et al., 2005), and a local component (traffic-related AP) with small-scale spatial variations inadequately described only by few monitoring stations (Hewitt, 1991; Monn et al., 1997). Moreover, individual monitoring of AP exposure is all the more difficult and expensive due to the high number of subjects and the long duration of exposure.

The lack of accurate and reliable direct measurements of AP exposure has led epidemiologists to use alternative approaches to assess individual exposure to traffic exhaust, which has not always avoided exposure misclassifications. Most epidemiological studies have assessed exposure to traffic-related AP by using self-reported truck traffic on the street of residence (Weiland et al., 1994; Duhme et al., 1996; Ciccone et al., 1998; Behrens et al., 2004) and traffic counts in the school district (Wjst et al., 1993) or on the street of residence (Edwards et al., 1994; Wyler et al., 2000; Kramer et al., 2000), by measuring distances from the child's home to the nearest main road (Venn et al., 2001), by combining distances from residences or schools to streets and the corresponding traffic density (van Vliet et al., 1997; Brunekreef et al., 1997; Venn et al., 2000; Janssen et al., 2003). To improve the estimation of traffic-related AP exposure, regression models based on a combination of monitored pollution data and exogenous information (Briggs et al., 1997, Briggs et al., 2000; Brauer et al., 2003; Carr et al., 2002) have been implemented, but have so far scarcely been used in epidemiological studies (Hoek et al., 2001; Brauer et al., 2002; Nicolai et al., 2003). Furthermore, regression models are case- and area-specific and their extrapolation in new areas requires a very dense monitoring network or more often specific measurements (Jerrett et al., 2005). Dispersion models such as CALINE (Benson, 1992), CAR (Eerens et al., 1993), AIRVIRO (SMHI, 1993), the simple parameterised OSPM (Hertel and Berkowicz, 1989; Raaschou-Nielsen et al., 2000) or more recently the three-dimensional (3D) computational fluid dynamics (CFD) MISKAM (Lohmeyer et al., 2002) have been successively implemented to provide better assessments of exposure to traffic-related air pollutants (TAP). Such dispersion models can be extrapolated in new areas more easily than regression models (de Hoogh et al., 2002; Jerrett et al., 2005; Vardoulakis et al., 2003). However, dispersion models have rarely been used in epidemiological studies, (Bellander et al., 2001; Nyberg et al., 2000; Clench-Aas et al., 1999), as the data they require can be difficult to collect and often unavailable. The ExTra index, however, based on the OSPM model (Reungoat et al., 2003) was used in the Vesta case-control study to assess the role of TAP in the occurrence of childhood asthma (Zmirou et al., 2002, Zmirou et al., 2004).

The purpose of the multi-centre cross-sectional epidemiological study, the French Six Cities Study, was to determine the impact of AP on childhood respiratory and allergic health at the population level by taking different AP indicators into account. In all, 5-day measurements of common air pollutants in classrooms and playgrounds were performed to assess short-term effects of indoor and outdoor proximity levels of AP (Annesi-Maesano et al., submitted 2005). Long-term effects of background AP were also considered (Pénard-Morand et al., 2005); the prevalence of asthma, allergic rhinitis and sensitisation to pollen were higher in areas with higher background concentrations of PM₁₀, SO₂ and O₃. Lastly, it was decided to build more accurate indicators of exposure to urban AP by applying the STREET 5 software based on a dispersion model to estimate ambient concentrations of TAP in front of the schools attended by the study children. Exposure near school represents an important component of usual outdoor exposure in most children in urban France. This paper has three objectives: (1) to present how STREET assesses individual exposure to urban AP; (2) to describe the study design to collect the input data STREET requires; (3) to report on the distribution of emissions and annual mean concentrations of TAP calculated by STREET in front of the schools of the French Six Cities Study. Children's exposure to TAP is assessed through these calculated concentrations.