This calculation is reasonably straight forward for annual means since it is simply the sum of the two components. The calculation of maps for CO is more complex because the maximum 8-hour concentration at the roadside is unlikely to be the sum of maximum background and roadside enhancements. This was assumed to be the case in the 'worst case' calculations carried out for the review of the NAQS (DETR et al, 1999) but the results of the individual site analyses presented in Table 2 indicate lower concentrations than the worst case analysis.

Table 3. Analysis of maximum running 8-hour CO concentrations at roadside sites in 1998 (ppm)

A road link with zero emissions can reasonably be expected to have a maximum roadside enhancement of zero, so the regression was forced through the origin:

This relationship is reasonably robust for these nine sites but the data for three sites have been excluded. The roadside enhancements at Cromwell Road 2 and London A3 Roadside are much lower than indicated by this relationship. This was expected for the London A3 site, which is close to a busy highway and has a more open aspect than the majority of the other sites. This is less obviously the case for Cromwell Road 2, but the buildings are a considerable distance from the roadside at this site. The roadside enhancement of concentrations at Exeter Roadside is much higher than indicated by the relationship. The reason for this is not known but may reflect particularly poor dispersion conditions at this site, which is in a narrow street canyon.

Annual mean background concentrations of a range of pollutants have been successfully mapped by deriving estimates of concentrations from emission inventory estimates of low level area emissions using an empirical box model method (Stedman, 1998). A map of annual mean background CO concentrations has been published by Stedman (1998) but the relationship between ambient concentrations and emissions estimates is considerably less reliable than for pollutants such as oxides of nitrogen. This was assumed to be due to additional uncertainties in the spatial distribution of CO emissions due to local variations in traffic speeds and congestion. Figure 3 shows that there is no clear relationship between the 99 th percentile of measured running 8-hour mean CO concentrations during 1998 and area emission estimates from the NAEI for the 25 1 km squares surrounding each background monitoring site. For emissions between 1,000 and about 12,000 Tonnes per 25 km² per year the concentration ranges between 1 and 2.8 ppm. At the sites with the highest estimated emissions, which are in inner London, the 99 th percentile concentration is consistently less than 2 ppm. In the absence of a clear relationship between concentrations and emissions, a constant value of 2.9 ppm has been chosen as the maximum likely background contribution to the maximum running 8-hour roadside concentrations in 1998.

The equation for estimating roadside CO concentration is 1998 is:

Estimates derived from this equation are compared with measured maximum running 8-hour CO concentrations in Figure 4. There is reasonably good agreement (mean of measurements = 5.2 ppm, mean of estimates = 5.8 ppm, r² = 0.56, n = 9). This equation was then used to calculate the maps of roadside CO concentrations in 1998 presented in section 5.

Maps of CO in years other than 1998 can be calculated by applying appropriate emission change factors to the background and roadside components:

We have assumed that the roadside enhancement will be directly related to road link emissions; fe_y is therefore derived from the ratio of annual emissions estimates listed in Table 1. The measurements of 99th percentile concentrations at background sites do not show any relationship with the spatial variation of area emissions estimates (Figure 3). Figure 5, however, shows that there is a clear relationship between the year to year changes in measured 99 th percentile of 8-hour CO concentrations at background sites and estimates of annual emissions from low level sources (mostly traffic sources) from the NAEI. Urban road traffic emissions estimates for each year between 1990 and 1998 were taken from Table 1 and the CO concentration is the maximum value measured in that year at any of the three background monitoring sites for which data are available throughout (London Bridge Place, West London and Glasgow City Chambers). The regression equation for the 99 th percentile of running 8-hour mean CO at background sites, for data for years with typical meteorology (all years except 1991) is:

where k_t is 0.0017.

There was an extreme air pollution episode in December 1991 with very high levels of vehicle related pollutants measured in both London and Glasgow (Bower et al, 1994). The unusual meteorological conditions experienced during this episode lead to higher CO concentrations than would be predicted by the regression relationship with annual emissions totals. An alternative estimate of background CO concentrations for years with extreme meteorology can be defined, with k_t = 0.0027.

The equation for calculating projections for year y is therefore:

where k_t = 0.0017 for typical meteorology and k_t = 0.0027 for extreme meteorology.

Chapter 3         Chapter 5