2.4 Uncertainties in the receptor modelling of the 'other' particle concentration
One of the limitations of the method that has been used to assign the contributions to daily PM₁₀ concentrations from different sources is that the concentration of particles that are neither primary combustion related or secondary has been calculated as the difference between the measured PM₁₀ concentration and the estimated total of primary and secondary. Any contribution to measured PM₁₀ concentrations that does not exhibit a similar temporal variation to either black smoke or sulphate is therefore included as 'other' particles.

It was noted in the APEG report that all of the 8 cities for which projections were calculated had several days with estimated daily mean 'other' particle concentrations in the range of 25 - 30 mg m-3 in both 1995 and 1996. It is impossible to know if these days are coarse particle episodes or are due to an underestimate of the primary or secondary contribution. For projections to 2005 or 2010 these days can have a significant influence on the 99th percentile. This is important because the other particle concentrations are assumed to remain unchanged between now and 2005 and 2010, while the contributions from the other sources are expected to reduce.
Table 2.5 lists the maximum and 99th percentile of fixed daily mean 'other' particle concentrations estimated within the model.

Table 2.5 Maximum and 99th percentile of fixed daily mean 'other' particle concentrations within the model (mgm-3)

	London Bloomsbury	Birmingham Centre	Bristol Centre	Manchester Piccadilly	Newcastle Centre	Belfast Centre	Edinburgh Centre	Liverpool Centre
1996 max	47	50	66	54	43	30	36	50
1996 99%ile	28	29	31	26	20	26	21	29
1995 max	44	45	29	-	41	44	44	44
1995 99%ile	37	33	33	-	23	35	34	42

These can be compared with the coarse particle concentration derived from the difference between co-located PM₁₀ and PM2.5 measurements (Table 2.6).

Table 2.6 Maximum and 99th percentile of fixed daily mean coarse particle concentrations from co-located PM10 and PM2.5 measurements (mgm-3)

	London Bloomsbury 19/6/97 - 27/8/98	Rochester 21/5/97 - 27/8/98	Marylebone Road 17/7/97 - 27/8/98	Birmingham Hodge Hill 1/1/95 - 31/12/95	Birmingham Hodge Hill 1/1/96 - 30/9/96
max	39	44	29	57	20
99%ile	21	20	27	27	19

The measured coarse particle 99th percentile concentrations indicate that levels are approximately 10 mgm-3 lower than the 99th percentile 'other' particle concentration estimated by the model, although it is not possible to make a direct comparison.
This uncertainty in the daily variation of the 'other' particle concentration within the receptor modelling applies to all of the projections presented in the APEG report and in this current report.


2.5 Alternative model of daily PM10 concentrations using NOX measurements
An alternative regression model of daily PM₁₀ concentrations, which utilises NOx measurements instead of black smoke measurements as an indicator for the primary particles was also presented in the APEG report. This has the advantage that the NOx measurements are directly co-located with the PM₁₀ measurements but an implicit assumption in this model is that NOx and primary PM₁₀ have common sources. While the black smoke monitoring method includes greater uncertainties than NOx measurements, it has the advantage of more directly sampling the atmospheric particle concentration.

The regression coefficients for this NOx-based model are compared with the black smoke based model in Tables 2.7 and 2.8. These are the coefficients obtained from a multiple linear regression of measured daily mean PM₁₀ concentrations, rural sulphate measurements and either black smoke or NOx measurements. The correlation coefficients are very similar. The intercept, C, is generally several mg m-3 lower than for NOx than for the black smoke-based regressions; the difference being PM₁₀ that has now been assigned to either primary or secondary instead of 'other'. Estimates of PM₁₀ concentrations for 2005 and 2010 have also been calculated using this NOx based model and were found to be very similar to those listed in Table 2.2. The similarity of the predictions based on the black smoke and NOx based models increases our confidence in the projections and also enables the calculation of projections based on either black smoke or NOx as the indicator for combustion primary PM₁₀.

Table 2.7 Regression coefficients for Black Smoke and Sulphate receptor modelling of 1996 PM₁₀ data

	Smoke coefficient, A	SO4 coefficient, B	Intercept, C	r2
London Bloomsbury	0.64	2.26	10.96	0.78
Birmingham Centre	0.59	2.41	8.30	0.71
Bristol Centre	1.03	2.35	10.83	0.70
Manchester Piccadilly	0.60	2.46	9.77	0.74
Newcastle Centre	0.66	3.13	7.73	0.84
Belfast Centre	0.71	2.30	9.21	0.79
Edinburgh Centre	0.59	2.46	9.85	0.61
Liverpool Centre	0.92	2.46	9.79	0.76

Table 2.8 Regression coefficients for NOx and Sulphate receptor modelling of 1996 PM₁₀ data (NOx concentration expressed in ppb)

	NOx coefficient, A	SO4 coefficient, B	Intercept, C	r2
London Bloomsbury	0.112	2.55	9.29	0.74
Birmingham Centre	0.207	2.29	6.52	0.72
Bristol Centre	0.145	2.62	7.97	0.76
Manchester Piccadilly	0.160	2.95	7.69	0.74
Newcastle Centre	0.155	3.48	5.66	0.83
Belfast Centre	0.302	2.65	5.91	0.85
Edinburgh Centre	0.122	2.99	6.42	0.67
Liverpool Centre	0.187	2.56	7.57	0.76