Air pollutant exposure evidence review
A review of existing data on pollutant concentrations and demographic features, used to understand the strength of the evidence base on how the Scottish population is being exposed to outdoor concentrations of NO2, PM10, and PM2.5.
Appendix 1: Overview of Exposure Reduction Targets
This section provides a high-level overview of considerations in setting an exposure reduction target as a background to the main section of the report, in particular Research Question 3.
In order to define exposure, and ultimately to measure compliance, there are three broad approaches: based on modelling, monitoring or through calculating population-weighted concentrations. A brief overview of these approaches is provided below.
Monitoring
There is a significant history of using PM2.5 mass concentrations as a proxy for exposure to airborne particulate matter and for reporting against compliance criteria. In particular, the large epidemiological studies linking exposure and poor health outcomes generally rely on measurements of mass concentration, this means that potential differences in toxicity of different components are not fully clear. Therefore, most targets are based on a mass concentration.
Compliance with both the EU Exposure Reduction obligation (covering PM2.5 and NO2) and the Population Exposure Reduction Target (PERT) implemented in England (PM2.5 only) are being assessed through monitoring. Advice given by AQEG to Defra was that the PERT should be based on monitors located predominantly at urban background locations, which would mean that they would (indirectly) represent the population-weighted mean concentration by being representative of significant proportions of the population. Both targets are assessed against an average across a number of urban background monitoring sites (not the average at each site).
The EU Exposure Reduction obligation is based on an ‘average exposure indicator’ (AEI) which is defined as ‘an average level determined on the basis of measurements at urban background locations throughout the average exposure territorial unit or, if there is no urban area located in that territorial unit, at rural background locations, and which reflects population exposure, used to check whether the average exposure reduction obligation and the average exposure concentration objective for that territorial unit have been met’.
Both targets therefore require a minimum number of monitoring sites and details are set out in relevant statutory instruments. Both use multiyear averaging to smooth the year-to-year variability arising from different weather conditions, but it is clear that this approach requires good long term data capture from a relatively static network of monitoring sites. PM2.5 is largely defined by how it is measured. Different measurement approaches produce different results, within a defined uncertainty envelope. It is therefore critical that consistent measurement techniques and quality assurance are deployed to track concentration changes and exposure reduction over the time period of the target. Instruments and techniques may also differ in their sensitivity to the various components of the aerosol particle mixture, affecting their ability to track changes as the particle source mix changes over time. For these reasons, measurement techniques have also been set out within the statutory processes for both the PERT and the EU Exposure Reduction obligation. It should be noted that tracking changes in concentrations also provides challenges for measurement, especially in locations where concentrations are low, and where the likely rate of change is small.
In Wales, the Clean Air Advisory Panel (CAAP) recommend that PM2.5 compliance assessment should be based on fixed monitoring using standard methods deployed at appropriate site locations at a sufficient number of sites. The CAAP suggest that the most suitable exposure reduction target for consideration is the Population Exposure Reduction Target (PERT), again based on fixed monitoring at suitable locations with enough monitoring points for a robust assessment. Other metrics were suggested, including ‘Population (number of people) living in locations above a concentration’, ‘Area above a concentration’ and ‘Population weighted mean exceedance’(PWME). CAAP also concluded that if Welsh Government were going to consider subtracting of natural sources of PM2.5 for compliance, then, in addition to acknowledgement of the concerns expressed above, there would need to be a robust method of assessment and justification for subtracting, with the possibility of revisiting the issue if new evidence or methods became available. This advice has provided evidence for a recent Welsh Government consultation on PM2.5 targets in Wales. Consultation closed on 16th February 2026 and at the time of writing, outcomes have not been published.
Consideration could also be given to whether roadside sites should be included in the assessment of exposure. Their inclusion would be designed to reflect the additional exposure in such situations. If they are to be included, then the number of sites would need to be broadly in proportion to the population living close to busy roads (possibly defined as roads with >10,000 veh/day and population within 50m from the edge of the road). However, roadside concentrations, particularly for NO2 are highly variable from site to site, depending on the traffic flow on the nearby road and on the exact distance from the road and the surrounding built environment (e.g. street canyons). It would therefore be difficult to determine suitably representative roadside sites. It would also be easy to influence the outcome of the Exposure Reduction target by changing the immediate environment of the road alongside the monitoring site, e.g. by pedestrianising the road. Roads are decreasing in terms of their contribution to overall emissions (particularly for NOx) and will continue to decrease further as the vehicle fleet decarbonises.
Modelling
Modelling is often used to support the process of air quality target setting. For example, Integrated Assessment Models can estimate future emissions and concentrations resulting from specified policy measures for different sectors, hence adding to the evidence base for identifying an achievable target that still delivers significant health and environmental benefits. In relation to the exposure reduction target setting process in England, the Air Quality Expert Group (AQEG) were asked to comment on the use of models[51]:
Put simply, the skill of models to predict exposure reduction of PM2.5 may not be uniform across the UK. There is limited data on the ability of models to accurately predict rates of change. Exposure reduction metrics will challenge models to provide predictions in locations where the concentration of PM2.5 is already relatively low, for example in the range 4-8 µg m-3 and where changes year on year may only be small, and potentially less than the measurement uncertainty.
Spatial variability in contributing sources; for example, the balance between local emissions and transboundary contributions, may lead to different uncertainties in model predictions of exposure reduction for different locations.
Modelling is generally not used to specifically assess compliance with exposure reduction targets, but could be used to evaluate progress and/or ascertain whether the policy has been successful more widely (i.e. not just with respect to whether the target concentration has been achieved), for example by checking whether inequalities have reduced or whether the exposure reduction target has had greater benefits for specific regions of Scotland. Progress in reducing PM2.5 is unlikely to follow simple linear trends over time and may vary by location. This may be due to a multitude of factors including variability in meteorology, non-linear aspects to the chemistry, or the timing of interventions. Models are likely to be critical to support the interpretation of ambient monitoring data.
In Wales, the Clean Air Advisory Panel provided an advice note for the Development of New PM2.5 Air Quality Targets for Wales[52]. In relation to how to assess compliance with the targets, the panel acknowledged that ‘Although monitors can provide us with a record of pollution levels at specific locations, they are not without their own drawbacks’, including representativeness, cost, coverage (especially in rural areas) and inconsistencies between monitoring types. The panel went on to say that: ‘atmospheric dispersion models have their own set of limitations which may lead to uncertainties. For the time being, the use of monitoring is potentially considered to be more legally robust than predicted concentrations from modelling. However, it is not possible to monitor everywhere across the country, so a combined approach that uses a combination of monitoring and modelling could add value’.
Modelling has the advantage of providing comprehensive coverage of predicted concentrations, compared to the specific (and relatively sparse) number of monitoring locations. Modelling also means that potential ‘what-if’ type scenarios for future policy development can be explored. This will require capability to adequately predict overall PM2.5 concentrations and to robustly assess the factors contributing to observed trends. Estimating future exposure reduction through modelling may vary depending on the contributing sources at any given location. Estimating exposure reduction, particularly in areas with low concentrations, is likely to come with significant uncertainties.
A modelled approach could also incorporate calculation of population-weighted mean concentrations. Population-weighted means is the calculation of the average pollution exposure (typically) experienced by individuals in a region, rather than simply averaging concentrations across a land area. This metric, crucial for health impact assessments, is dominated by air quality data from areas with higher population densities.