9 Daily mapping of PM₁₀ concentrations

9.1 Introduction
Maps of estimated annual mean background PM₁₀ concentrations have been calculated for 1996 (Stedman 1998). Projections of the annual mean concentrations that current policies are likely to deliver in 2005 can also be calculated by making certain assumptions about the likely changes in primary and secondary particle concentrations.

Background PM₁₀ concentrations are made up from contributions from primary, secondary and coarse particles, and the relative proportions vary from day to day. This is why the relationships between numbers of days above thresholds, high percentiles and annual means are not very reliable, particularly if it is applied to future years, such as 2005, when the relative contributions from primary, secondary and coarse particles to PM₁₀ are likely to have changed.

Maps of both current and projected PM₁₀ concentrations are however required for the quantification of health impacts similar to those presented by COMEAP (1998). Whilst maps of annual mean concentrations are required for an assessment of the health benefits of reducing concentrations of PM₁₀, the calculation of these maps from a series of daily maps enables maps of other statistics of interest to be calculated, without having to make assumptions about the relationship between means and numbers of days above thresholds. This report provides the first maps of annual means and numbers of exceedance days for comparison with the PM10 objective and limit values. The use of these mapping methods does, however, increase the level of uncertainty of the projections compared with the projections presented in this report for individual sites.

9.2 Method
The following data have been used:

• daily mean particulate sulphate concentrations at eight rural sites;
  
• daily PM₁₀, black smoke and rural particulate sulphate concentrations for eight different cities have been analysed in order to provide estimates of primary, secondary and 'other' particle concentrations for each day for the following sites:
  
  • London Bloomsbury
    
  • Edinburgh Centre
    
  • Belfast Centre
    
  • Birmingham Centre
    
  • Newcastle Centre
    
  • Liverpool Centre
    
  • Bristol Centre
    
  • Manchester Piccadilly;
    
• a 1 km x 1 km grid square map of combustion related emissions of PM₁₀ from the National Atmospheric Emissions Inventory.
  

The daily PM₁₀ map for each day is calculated as follows:

PM₁₀ map = primary map + secondary map + 'other' map

where

The secondary map is calculated by multiplying a map of particulate sulphate by 2.46, which is both a reasonable theoretical factor and consistent with the daily individual site analysis. The map of particulate sulphate concentrations is interpolated from the daily sulphate measurement data at eight rural sites.

The 'other' map is interpolated from the daily estimates of 'other' particle (largely coarse particle) concentrations derived from the daily analysis of concentrations at urban sites. This is likely to overestimate coarse particle concentrations in rural areas but probably not to a large extent.

The primary map consists of two components:

• long range transported primary particle concentration
  
• local primary particle concentration
  

The long range transported component is expected to be relatively small and show a similar spatial variation to secondary particle concentrations. Daily concentrations have been estimated by multiplying the daily secondary concentration by 0.08333.

The local primary particle concentration has been calculated by multiplying the map of annual combustion related emissions by a factor k. This factor represents the efficiency of the dispersion of these local emissions within a 5 km x 5 km box and is allowed to vary both spatially and from day to day. The k factor is interpolated for each day and is allowed to vary by dividing the estimates of the daily primary contribution to PM₁₀ from the analysis of daily concentrations at individual urban sites by the estimated contribution that primary particles make to annual mean PM₁₀ concentration.

k = ka [p]/[p]a

where
k is the coefficient used to derive daily mean local primary PM₁₀ concentration from annual emissions estimates;
ka is the coefficient between annual emissions and annual mean primary particle concentration;
[p] is the daily mean primary particle concentration derived from the daily analysis of PM₁₀, black smoke and sulphate measurements;
[p]a is the annual mean of primary particle concentration aggregated from the daily analysis of PM₁₀, black smoke and sulphate measurement.

9.3 Results for 1995 and 1996
The general pattern of annual mean concentrations shown in Figure 9.1 are similar to that derived from an analysis of annual concentrations (Stedman 1998). Rural concentrations are highest in the south and east due to the varying magnitude of the secondary particle contribution. Urban concentrations are higher due to the local primary emissions. Rural concentrations are estimated to be rather higher in 1996 than in 1995, while urban centre concentrations were not much higher in 1996 than in 1995. A comparison of the mapped estimates with measured annual mean concentrations has shown reasonably good agreement. Concentrations in central London are, however, overestimated. The high concentrations in North Wales in 1995 are probably caused by the mapping method and are unlikely to be representative.

In common with the individual site receptor modelling results, we have not scaled the maps of number of days with concentrations greater than 50 mgm-3 to enable direct comparison with the NAQS objective because of the uncertainties associated with the large scaling factor that would be required. The maps presented in Figure 9.2 represent the number of days with fixed 24-hour mean PM₁₀ concentrations greater than 50 mgm-3. There is a striking difference between the two years, with the secondary PM₁₀ episodes during the early part of 1996 leading to many more days above 50 mgm-3 in both rural and urban areas.


9.4 Projections for 2005
Projections for 2005 have been calculated from the 1995 and 1996 maps using the same business as usual scenario as for the individual site receptor modelling. Annual mean maps for 2005 are shown in Figure 9.1. The projected annual mean PM₁₀ concentrations for 2005 are significantly lower than those for 1995 and 1996 and the values calculated from the 1996 are not very different from those calculated from the 1995 map.

Maps of the estimated number of days with PM₁₀ concentrations greater than 50 mgm-3 for 2005 based on either 1995 or 1996 meteorology are also shown in Figure 9.2. Large areas of the country have more than four days with projected concentrations greater than 50 mgm-3 , even for projections based on 1995 monitoring data. This is also likely to be a significant underestimate of the number of days with running 24-hour mean concentrations greater than 50 mgm-3, as discussed above.


9.5 Projections for comparison with EU limit values
Figure 9.3 shows maps of the number of days with estimated PM₁₀ concentrations greater than 50 mgm-3 for comparison with the EU 24-hour limit values. The Stage 1 limit value is that 50 mgm-3 should not to be exceeded more than 35 times in 2004. This mapping analysis indicates that central London may be at risk of exceeding this limit value in years such as 1995 and the whole of the London area may be at risk in a year such as 1996. It is likely, however, that concentrations in the London area have been overestimated by the mapping method because the individual site analysis indicated exceedance for 1996 only, and this was confined to central London.

The indicative Stage 2 24-hour limit value of 50mgm-3, not to be exceeded more than 7 times in 2009 is likely to be exceeded over most of England and Wales for the business as usual scenario.