Introduction

        An Air Quality Monitoring Network has been developed in the Lower Fraser Valley (LFV) area, collecting data for specific locations. Monitoring allows Metro Vancouver to compare concentrations of key air contaminants with air quality objectives, which are intended to minimize health risks¹.  However, the monitoring network might not capture all areas with NO₂ concentration exceeding the objective due to its limited density. Maps generated from the LUR model, with a much higher resolution, may facilitate detecting the non-compliance. In this section, maps based on LUR models in 2003 and 2009/10 were used to identify areas where the objective was not met. Further, the most vulnerable population living at the identified areas was mapped to picture the magnitude of potential risk

 

Methods

Study area and the vulnerable population

Three cities, the city of Vancouver, Burnaby and Richmond, were selected as the study area. Kids under 5-year old were identified as the most vulnerable population for analysis. Population data was obtained from Statistics Canada, Census 2006

Identify areas with NO₂ concentration exceeding annual objective

Firstly, the study area was clipped from the original maps, in raster dataset with 10m*10m pixels. Secondly, raster dataset was converted to points (Conversion Tools - From Raster - Raster to Point), so that each point was assigned a value that was estimated for the specific grid. To reduce processing time, the points were resampled at 25m*25m resolution.  Those points with NO₂ value exceeding annual objective (21ppb) were then selected using “selecting by attributes”. Finally, the selected points were converted back to raster for displaying purpose. The same process was done for both 2003 and 2009/10. Results were compared between the two years.

Magnitude of potential risk at dissemination area (DA) level (linkage to census units and population data)

Average concentrations for each DA were calculated by spatial join (points to polygons). DAs with NO₂ value exceeding annual objective was identified by “selecting by attributes” and linked to population data by table join.

 

Results

Areas exceeding annual objectiveFig. 1 and Fig.2 show the modeled surfaces for the three cities in 2003 and 2009/10. Areas exceeding annual objective (21ppb) were identified in blue colors.  They all appear to be at intersections or along roads. Most of the along-road pollutions, however, disappeared in the 2009/10 map.

jiohoiphg

Vulnerable population living at areas with pollution level above objective

DAs with average concentration exceeding the objective level were identified for both years (Fig.3 and Fig.4).  26 DAs exceeded the objective in 2003, but the number reduced to 4 in 2009/10. Yaletown was identified as the most at-risk community, but the situation is much better in 2009/10.The maps also show number of kids under 5-year old living in those highly polluted DAs.

(for a line)

Discussion

        This section demonstrated that the LUR model could be used to investigate outdoor air quality compliance, to supplement monitoring network. Based on the 20 monitoring stations in the network that are currently collecting data for NO₂, all average levels met Metro Vancouver’s annual object in 2009 except for Vancouver-Downtown¹.  The 2009/10 map, however, revealed other areas with pollutant level above the 21ppb objective, which was not detected by the monitoring network. In addition, the results from the LUR model might also facilitate the sitting of monitoring stations. For example, a new monitoring station might be built in an area with increasing pollution levels, or a station might be relocated to another place depending on the pollution pattern and specific purpose of sitting a station.
        The maps could also provide evidence for evaluating regional air quality programs. As illustrated in this case, areas with pollution level above the objective has been greatly reduced from 2003 to 2009/10. Those reductions might be attributed to on-road emission control, improved traffic patterns etc. To understand what is likely to be the cause, more information needs to be acquired.
        Another use of the maps is to inform public health actions. There is substantial evidence linking outdoor air pollution and health. Relationships between the environment and health are intrinsically geographical, where two geographies intersect: the agents of risk and the population at risk.²  In the second part of this section, those areas with air pollution concerns were mapped with vulnerable population sizes. Compared with pollution-only maps, it provides a better picture of the magnitude of potential risk at the population level, which might be used for public health decision. For example, it may guide allocation of limited resources to the most at-risk area, to achieve best possible benefits.  GIS offered ways to analyze the two geographies, pollution and population, to bring them together, which has great potential in exposure and health risk assessment.
        A major limitation of linking the pollution and population data arises from areal interpolation. Pollution estimates within one DA were averaged to calculate the overall level for the DA block so that pollution data can be linked to population data (available at DA level for the chosen population at risk in this example). And analysis was done at the DA framework. However, in doing so, the underlying assumption of evenly distributed population within one DA polygon might not be true. A better estimate of people’s exposure levels within a DA can be obtained if people’s location within the DA is available, especially for continuous data (in this example, air pollution).

___________________________

1. Metro Vancouver. 2009 Lower Fraser Valley Air Quality Report. June, 2010

2. Hoogh et al.Using Modeled Outdoor Air Pollution Data for Health Surveillance. P117-144. GIS in public health practice . CRC Press.  2005