6 Ozone

6.1 Introduction
The health benefits calculations that been carried out within the review of the NAQS for ozone are less complete than for the other pollutants because there are a number of complicating factors which are unique to this pollutant.

• The health outcome calculations in the COMEAP quantification report (COMEAP, 1998) were calculated using two different approaches:
  • a threshold for health impact of 50 ppb
    
  • no threshold
• peak ozone concentrations during photochemical episodes are expected to reduce in future years due to reductions in emissions of the precursors of ozone on European scale.
  
• long term mean ozone concentrations in cities are expected to increase in future years as urban NO_x emissions reduce.
  

The methods that have been used previously to map daily ozone concentrations and calculate the resulting health impacts for the summer of 1995 are described in COMEAP (1998) and in some detail in Stedman et al (1997). Maps of maximum running 8-hour mean were calculated for each day during the summer of 1995 and health impacts for ozone concentrations greater than a 50 ppb threshold for the GB population were then calculated. A map of the summer mean of the daily maximum of running 8-hour mean ozone concentration for 1995 was also calculated and used for the no-threshold health impact calculation (this is mathematically the same as carrying out the no-threshold calculation on a daily basis and summing the results). The estimated health impacts due to ozone during the summer of 1995 calculated for the review of the NAQS used the same maps of ozone concentrations as used by COMEAP (1998) and Stedman et al (1997), the baseline rates listed in Table 2.2 and 1991 UK census data. The resulting numbers of deaths and hospital admissions in 1995 due to ozone are therefore different from those published previously.

The modelling work that has been undertaken to quantify the likely effects of emission reductions on peak and long term mean ozone concentrations is briefly described in sections 6.2 and 6.3. The results of the modelling work for peak ozone concentrations has enabled 50 ppb threshold health benefit calculations to be completed for a range of different scenarios and these results are listed in section 6.4. Health benefit calculations for the no-threshold case have not been completed and the reasons why this has not been done are given in section 6.5.


6.2 Modelling peak concentrations in 2010
Modelling of ozone concentrations on a European scale is currently being carried out in the UNECE, as part of the preliminary stages of the negotiations on the so-called Multi-Protocol, and within the EU as part of the development of the combined ozone and acidification strategies and the draft ozone directive. The emissions from each Member State which deliver the environmental goals of the strategies will be embodied in the National Emissions Ceilings Directive, a draft of which the Commission intends to produce in February 1999. The modelling is being carried out by IIASA in both fora, using the integrated assessment model, RAINS, which incorporates the EMEP atmospheric models transfer matrices, abatement cost and efficiency information, and critical loads and levels data for the ecosystems in the countries concerned. As well as this European scale modelling, more detailed modelling of ozone is being carried out for DETR by the Meteorological Office and the Universities of Edinburgh and Lancaster. This latter modelling is concentrating on ozone in the UK at a more detailed spatial level than the RAINS model is currently capable of handling. It deals with regional scale ozone concentrations in the UK, at a spatial resolution of 100 km.

The results of the European and UK scale modelling studies have been combined to produce Table 6.1, which lists the ozone reduction factors that each scenario is likely to deliver on days with photochemical ozone episodes (current ozone concentrations greater 40 ppb). The reference scenario refers to the expected emissions in 2010 resulting from the implementation of current policies. The F1 scenario is a set of emission ceilings for the EU-15 following the cost optimal attainment of environmental targets for acidification and ozone across the EU.

Table 6.1 Photochemical ozone reduction factors relative to 1995 concentrations.

scenario	description	factor
1995		1.000
2010 reference	current policies	0.799
2010 F1		0.774
2010 MFR	Maximum feasible reductions	0.662

6.3 Modelling long term urban ozone concentrations in 2010
Unlike the European scale modelling referred to above, UK modelling has begun the more difficult task of estimating ozone concentrations in urban areas, at a spatial resolution of 1 km. Preliminary results for London (Derwent, pers comm, 1998) suggest that for reductions in NO_x emissions of 50-60%, annual mean ozone concentrations will increase by about 3-4 ppb from current levels of around 10-15 ppb. The functions listed in Table 6.2 have been suggested for predicting the change in annual mean ozone concentration for 60% NOx control (which is roughly business as usual for 2010, relative to 1995).

Table 6.2 Function for predicting change in annual mean ozone concentration (Derwent, pers comm, 1998)

ozone concentration in 1995	ozone concentration in 2010
< 10 ppb	add 4.2 ppb
10 - 26 ppb	add 4.2 - (0.25*(ozone - 10))
> 26 ppb	no change

6.4 Health benefits for a 50 ppb threshold calculation
The ozone reduction factors listed in Table 6.1 have been used to scale the daily maps of ozone concentrations for the summer of 1995 and the resulting health benefits for each scenario are listed in Table 6.3. A linear reduction in deaths and admissions between 1995 and 2010 has been assumed in the calculations of the cumulative benefits.

Table 6.3 Numbers of deaths and respiratory hospital admissions due to ozone for a range of scenarios UK Urban and Rural population, 50 ppb threshold).

Deaths Brought Forward		Emergency Respiratory Hospital Admissions (additional or brought forward)
In named year	Total reduction (from 1995 baseline)	In named year	Total reduction (from 1995 baseline)
720		615
235	3890	200	3320
190	4250	160	3625
25	5550	20	4745
Notes Values have been rounded to the nearest 5.

6.5 Health benefits for a no-threshold calculation
The calculation of the numbers of deaths and hospital admissions due to ozone for the no-threshold case is relatively straight forward for 1995 and the results are listed in Table 6.4.

Table 6.4 Numbers of deaths and respiratory hospital admissions due to ozone for a no-threshold calculation (UK Urban and Rural population).

Pollutant	Deaths Brought Forward		Emergency Respiratory Hospital Admissions (additional or brought forward)
ozone	In named year	difference relative to 1995	In named year	difference relative to 1995
1995	12240		10455
2010 effect of reduction in peak concentration	11455	-790	9780	-675
2010 effect if increase in mean concentrations in urban areas	12600	+360	10760	+305
Notes Values have been rounded to the nearest 5. Italics indicate the greater uncertainty associated with these values

For 2010 it is expected that peak concentrations will reduce relative to 1995 but non-episode concentrations in urban areas are likely to increase. Current modelling methods are not sufficiently advanced to enable a full no-threshold calculation of the health outcomes in 2010. The influences of the changes in peak and changes in urban ozone concentrations between 1995 and 2010 have been calculated separately.

For the reduction in peak values we have applied an ozone reduction factor for the reference scenario listed in Table 6.1 on a daily basis to 8-hour ozone concentrations greater than 40 ppb (values above 50 ppb were multiplied by 0.799 and values between 40 and 50 were set to 40 ppb). This results in the reductions in numbers of deaths and admissions which are listed in the second row of Table 6.4.

In order to apply the function for predicting changes in annual mean ozone to the daily maximum of running 8-hour ozone concentrations we calculated a map of the summer mean of the daily maximum of running 8-hour ozone for 1995 and examined the relationship between this statistic and annual mean:

summer mean of daily maximum of running 8-hour ozone = 0.884 * annual mean + 16.328
(r² = 0.78).

This relationship was then used to calculate a map of predicted summer mean of the daily maximum of running 8-hour ozone for 2010 in which the concentrations have been adjusted to take the likely changes in long term mean ozone concentrations in urban areas into account. This results in the increases in numbers of deaths and admissions as listed in the third row of Table 6.4.

Thus the likely reduction in admissions and deaths due to reducing peak levels of ozone is larger than the likely increase due to increases in urban areas. The results of this preliminary calculation are not very robust and firm conclusions will need to await more detailed analysis of the combined effects the emissions changes on future ozone concentrations. It is reasonable to note, however, that the likely reductions in health effects due to the reduction in peak ozone may be offset by the increase in urban ozone for a no-threshold calculation.

7 Acknowledgement
This work was funded by the UK Department of the Environment, Transport and the Regions as part of their Air Quality Research Programme.

---