PM_2.5 in the UK
ER12
December 2010

EXECUTIVE SUMMARY

Background to research

It is recognised that exposure to airborne particulate matter (PM) can give rise to significant health effects and there is no evident safe level. Attention was initially directed towards PM of less than 10 mm diameter (PM₁₀), but epidemiological evidence regarding the health effects of smaller particles has now changed the focus to smaller particles less than 2.5 mm diameter (PM_2.5). As a consequence, new legislation has recently come into effect to control exposure to PM_2.5 (with a continuation of controls on PM10), and the United Kingdom (UK)Government will need to develop approaches to reduce exposure to PM_2.5. To ensure this is carried out in the most cost-effective way it is first necessary to have a good understanding of the sources, pathways and health effects of PM_2.5, and of the legislation that can contribute to its control. The focus is on annual mean concentrations of PM_2.5, as these are the most significant in terms of health outcomes, and on concentrations within urban areas, where most people are exposed.

Objectives of research
This SNIFFER project is designed to provide regulators with an understanding of PM_2.5, including health effects, sources, speciation, pathways, regulatory requirements, monitoring and modelling tools. It will be used to advise regulators on how PM sources need to be regulated to achieve the air quality limit values and objectives.

Key findings and recommendations

• There are many different sources, both natural and anthropogenic, contributing to PM_2.5 concentrations. These can be primary particles, which are emitted directly into the atmosphere, or precursor gases, which form secondary particles through atmospheric reactions.
• Industrial sources and power stations contribute most to national primary anthropogenic emissions (35%), followed by road transport (24%), residential (13%) and shipping (10%).
• The PM_2.5 fraction of PM is removed only slowly from the atmosphere; the dispersion of PM_2.5 in the near-field can thus be treated like that of a gas.
• Primary anthropogenic particles (from all sources) make a relatively small contribution to urban background PM_2.5, probably less than 25% (although there is some uncertainty with regard to this estimate). Primary emissions from combustion sources are made up largely of elemental and organic carbon. These particles will have trace metals and trace organic compounds associated with them.
• In general terms, motor vehicle emissions (both exhaust, and brake and tyre wear) contribute slightly more than industrial sources (including point sources) to urban background PM_2.5.
• More specifically for PM_2.5 from traffic, the overall breakdown is broadly 64% vehicle exhaust, 32% brake and tyre wear, and 4% resuspended road particles. Thus around 36% of traffic emissions are likely to be from sources that are currently uncontrolled.
• Natural sources of PM include sea salt, which accounts for ~5-15% of urban background PM_2.5, with higher contributions found towards the western coastal areas of the UK, while windblown dust contributes ~5-8%.
• The formation of secondary particles happens relatively slowly (hours to days), thus secondary PM_2.5 is found well downwind of the sources of emission of the precursor gases. As a consequence there is a reasonably even distribution of secondary PM_2.5 on a regional scale.
• Secondary particles are mainly formed from emissions of sulphur dioxide (SO₂), nitrogen oxides (NOx), ammonia (NH3) and volatile organic compounds (VOCs). Emissions of these precursor gases have declined significantly over the last three decades, especially for SO₂, followed by NO_x, VOCs and to a limited extent NH₃.
• Secondary particles dominate urban background PM_2.5 in the UK, accounting for some 30-50% of the PM_2.5 in urban areas. They are predominantly ammonium sulphate, ammonium nitrate, and organic particles. A significant proportion of secondary PM_2.5 is imported into the UK, having been formed from precursor emissions in continental Europe.

• Rural annual mean concentrations of PM_2.5 range from around 3.5 μg/m³ in Scotland to around 10 μg/m³ in southern England. Urban background concentrations are a few μg/m³ higher, and are highest in central London at around 15 μg/m³. Annual mean concentrations close to busy roads will be a few μg/m³ higher still, although concentrations decline rapidly on moving away from the carriageway, such that they are indistinguishable from the background after about 20-50 m.
• Estimated population-weighted annual mean PM_2.5 concentrations for the UK in 2010 range from 5.5 μg/m³ in Scotland, to 6.4 μg/m³ in Northern Ireland, to 8.3 μg/m³ in Wales, and to 10.6 μg/m³ in England. These values will be determined largely by concentrations within urban areas, where the majority of the population is exposed. The population-weighted annual mean PM_2.5 concentration in inner London in 2010 is estimated to be 14.1 μg/m³.

• Both short- and long-term exposure to PM_2.5 gives rise to a range of health effects. These include hospital admissions and mortality from respiratory and cardiovascular diseases.
• The effects of long-term exposure are more significant than those of short-term exposure, in terms of the overall impact on the nation’s health.
• There is no recognised threshold below which there are no health effects.
• There is no clear evidence as to which of the PM_2.5 components give rise to the toxic effects. Thus, at this stage, all components need to be considered capable of giving rise to these effects.
• Interventions to reduce exposure to PM, such as the ban on coal burning in Dublin, have been shown to be beneficial in health terms, with fewer respiratory and cardiovascular deaths following the interventions.
• Exposure to PM_2.5 reduces life expectancy by around six months averaged over the whole of the UK. For those individuals who are particularly sensitive the reduction in life expectancy could be much greater. For instance, if 10% of the population is affected then the loss of life expectancy for these individuals would rise to an average of around 6 years. Whilst it is not straightforward to compare health risks, it is estimated that eliminating exposure to man-made PM_2.5 would yield greater benefits than eliminating road traffic accidents or exposure to passive smoking.
• Significant health benefits across the UK and European Union (EU) populations have been calculated for a given reduction in exposure to PM_2.5. These translate into financial benefits, which more than offset the costs of mitigation programmes currently in place.

• The focus for PM_2.5 is on limiting long-term exposure through use of annual mean standards.
• There are two strands to the legislation to control exposure to PM_2.5:
• exposure standards to define the level of control required; and
• measures to limit emissions to meet the standards (including emission standards used to control emissions from both transport and industrial sources, and measures such as the National Emissions Ceilings Directive).
• Control strategies to reduce emissions that contribute to PM_2.5 include:
• continued efforts to reduce emissions of precursor gases, to limit the formation of secondary PM, which will require national controls within a European-wide structure; and
• continued efforts to reduce emissions from primary sources giving rise to the urban enhancement of PM_2.5 concentrations, which will need to focus on industrial sources, road transport and to a lesser extent on domestic sources.
• There is a large body of legislation already in place to control exposure to PM, much of it involving the control of emissions. This is often focused on PM in general and has not been specific to PM₁₀ or PM_2.5. However, measures to control total PM and/or PM₁₀ will in general result in lower PM_2.5 concentrations.
• The concentration standards to limit public exposure, which previously were focused on PM₁₀, have, in the last few years, been extended to include PM_2.5.
• A new exposure-reduction approach has been introduced for PM_2.5. This is in recognition of the absence of an identified safe level for exposure to PM, and requires a reduction of PM_2.5background concentrations in urban areas across the UK over the period 2010 to 2020.

• Little monitoring of PM_2.5 concentrations was carried out in the UK prior to the expansion of the national network (as part of the Automatic Urban and Rural Network (AURN)) in 2009 to meet the requirements of the EU Clean Air for Europe (CAFE) Directive. The AURN uses Filter Dynamic Measurement System (FDMS) analysers, which provide hourly mean concentrations that are equivalent to the EU reference method, i.e. the concentrations from these analysers are reference equivalent. The AURN currently comprises of 67 sites, of which 43 are classified as urban background¹, 17 as roadside/kerbside, 4 as industrial and 3 as rural.
• Additional monitoring is carried out by some local authorities, such as that undertaken within the London Air Quality Network (LAQN).
• Annual mean concentrations have been collated for 36 sites that had >90% data capture in 2009 (mostly AURN sites, but including some local authority sites). Urban background concentrations (27 sites) were mostly between 12-16 μg/m³, with the highest concentration recorded at the London Eltham site in east London, at 17.6 μg/m³.
• Analysis of the 2009 data indicates the following overall pattern for annual mean PM_2.5concentrations:
• rural background concentrations are ~3-10 μg/m³, with the lowest values in Scotland and the highest in southern England;
• urban background concentrations are ~ 3-6 μg/m³ above the rural background in major urban areas;
• roadside concentrations alongside busy roads are ~1-2 μg/m³ above the urban background; and
• kerbside concentrations alongside busy roads are ~7-8 μg/m³ above the urban background.
• The limited evidence on trends in PM_2.5concentrations over the last decade indicates:
• no change in concentrations at rural background sites;
• a very slight reduction in concentrations at urban background sites; and
• a downward trend at roadside sites.
• The 2009 results show that concentrations of PM_2.5:
• are substantially higher during the winter months than in the summer, with the highest monthly mean being around 10-15 μg/m3 above the lowest monthly mean.
• increase during the week, to give the highest levels on a Thursday/Friday, before dropping by around 4 μg/m³ to a Sunday low. There is currently no full explanation available for this pattern; and
• have a diurnal pattern that is consistent across central and southern parts of the UK (essentially south of the Lake District) with the lowest concentrations occurring during the afternoon and the highest during the early part of the night. This may be due, at least in part, to loss of semi-volatile PM_2.5 from the atmosphere during the warmer part of the day, with this material then recondensing during the evening and night time, although there is no clear-cut explanation for this pattern. The afternoon drop in concentrations is not apparent in the results for the northern parts of the UK.
• A detailed analysis taking into account wind direction and wind speed has shown a consistent pattern across the UK, with the highest concentrations being associated with winds from the northeast through to southeast. These higher concentrations are often associated with stronger winds >10 m/s, and occur throughout the day and night. This is evidence of a significant PM_2.5 contribution being imported from continental Europe, probably as secondary PM.

• Annual mean concentrations can normally be modelled with lesser uncertainty than is the case for short-term (hourly or daily mean) concentrations. Thus modelling of PM_2.5 should, in some respects, be more straightforward (than for PM10, for which 24-hour concentrations are important) given that the focus is on the annual mean.
• A wide range of models can be applied to the modelling of PM_2.5, although these generally work at different spatial scales. Defra is currently undertaking studies to evaluate which models are best suited to meet policy needs and to identify those that should be investigated and developed.
• Two modelling suites are commonly used in the UK to estimate near-source concentrations of PM_2.5: ADMS (both ADMS4 for point sources and ADMS-Roads) and AERMOD. These are reasonably well-established and widely used for the assessment of development schemes, industrial permitting etc.
• There is only one national model applied routinely for PM_2.5. This is the semiempirical Pollution Climate Mapping (PCM) model developed by AEA and used by Defra for policy development.

• Analysis of concentrations in relation to the EU limit values and UK objectives has shown that the key driver for action to control exposure to PM_2.5 will be the exposure-reduction targets that have been established. The UK target is for a 15% reduction in annual mean PM_2.5 at background locations across the major urban areas, while the EU target will be 10-15% (still to be determined, once monitoring results for the three years 2009-11 are available). These represent reductions in annual mean concentrations of around 1.5-2 μg/m3 over the ten years between 2010 and 2020. If these reductions are to be achieved from the local sources that give rise to the urban background enhancement of around 3-6 μg/m3, then these urban source contributions would need to be reduced by some 25-67%. On the other hand, if the whole of the reduction were to be achieved by reducing the secondary PM contribution, which accounts for ~30-50% of urban background PM_2.5 (around 4-6 μg/m3), then the required reduction of this secondary PM would be some 25-50%. In either case the reductions required to meet what appears to be a small target reduction, are substantial. If both are tackled equally the percentage reductions would essentially be halved, but would still remain challenging.
• EU limit values established to ensure there are no hot-spots with excessive concentrations of PM_2.5 are unlikely to be exceeded in the UK. The PM10 objectives and limit values have been shown to be more stringent than the PM_2.5 objectives and limit values, thus the former will drive policies to reduce exposure in hot spots; this will, in general, help drive down PM_2.5 concentrations at these locations.
• The response of secondary PM concentrations to changes in precursor gas emissions is less than proportional. This implies that greater reductions in precursor gas emissions are required than might at first sight be the case. The chemistry of secondary inorganic PM formation is such that reductions in ammonia emissions may be more effective at reducing PM concentrations than equivalent reductions in sulphur dioxide and nitrogen oxides, although this requires further investigation. The chemistry of secondary organic PM formation is poorly understood and it is not clear which sources of non-methane VOCs (NMVOCs) should be targeted to reduce PM concentrations.

• The following recommendations are proposed for further work to provide a better understanding of PM_2.5 sources and concentrations, so as to allow appropriate control strategies to be developed. Their justification is described in Section 9:
1. Modelling studies should be carried out to establish the contributions of UK and other EU emissions of precursor gases to annual mean PM_2.5 concentrations, and how these contributions will respond to changes in emissions. This will help guide the development of the most cost-effective control programme;
2. A programme of field, chamber and modelling studies should be carried out to establish the contributions from anthropogenic and ‘natural’ sources of organic carbon to PM_2.5, and how these will respond to changes in emissions;
3. Further work should be carried out on quantifying emissions, in particular from the road transport sector under ‘real-world’ driving conditions, and from industry. The contributions to urban background concentrations from other sources should be investigated using PM speciation studies. Such studies should then be used to inform the source contributions via modelling studies;
4. Improved deterministic modelling approaches should be developed at the national and urban scale, with robust treatments of atmospheric dynamics, chemistry and aerosol processes, so as to give size and composition-resolved information on airborne concentrations. This will enable reliable projections to be made of future concentrations, to complement the findings of more empirical models.
5. Seven additional PM_2.5 monitors should be set up at rural background sites, as a minimum. These could usefully be collocated with existing ozone monitors;
6. Sufficient resources should be provided to ensure that the results of the PM monitoring programmes are subject to thorough analysis on an on-going basis;
7. A programme of work should be established to support source apportionment of annual mean PM_2.5 at urban background locations in different areas of the UK; and
8. Consideration should be given to ways of using the planning system to require reductions in PM_2.5 emissions to be incorporated into the planning of all new developments.

• To deliver the additional work and subsequent control strategies:
• The UK Government and devolved administrations will need to establish a work programme to define the sources contributing to urban background PM_2.5 in those urban agglomerations where compliance monitoring is being carried out and then predict the changes that will take place between 2010 and 2020, before developing measures to fill any short-fall in the predicted concentration reduction;
• The Environment Agency, the Scottish Environment Protection Agency and the Environment Agency Northern Ireland should contribute to the understanding of primary and secondary emissions from industrial, agricultural and waste sources that they regulate. They should also assist in modelling the contribution of primary PM_2.5 emissions to urban background concentrations and identifying measures to limit primary and secondary emissions; and
• Local Authorities could be expected to help implement national measures to reduce primary PM_2.5 emissions, as well as consider ways to control PM_2.5 from new developments. They could also help develop amendments to the Clean Air Act to control emissions from biomass boilers.

Key words: PM_2.5, health effects, ambient, emissions, sources, concentrations, legislation, monitoring, modelling.

Copies of this report are available from the Foundation, in electronic format on CDRom at £20.00 + VAT or hard copy at £50 00, less 20% to FWR members.

N.B. The report is available for download from the SNIFFER Website