The role of long-range transport and domestic emissions in determining atmospheric secondary inorganic particle concentrations across the UK

. Surface concentrations of secondary inorganic particle components over the UK have been analysed for 2001–2010 using the EMEP4UK regional atmospheric chemistry transport model and evaluated against measurements. Gas/particle partitioning in the EMEP4UK model simulations used a bulk approach, which may lead to uncertainties in simulated secondary inorganic aerosol. However, model simulations were able to accurately represent both the long-term decadal surface concentrations of particle sulfate and nitrate and an episode in early 2003 of substantially elevated nitrate measured across the UK by the AGANet network. The latter was identiﬁed as consisting of three separate episodes, each of less than 1 month duration, in February, March and April. The primary cause of the elevated nitrate levels across the UK was meteorological: a


M. Vieno et al.: The role of long-range transport and domestic emissions
population-weighted exposure to PM 2.5 by a specified percentage between 2010 and 2020 (Heal et al., 2012).
The complexity of ambient PM composition and formation, combined with the influence of meteorology on chemistry, dispersion and deposition, considerably complicates pinpointing the contributions of different chemical pollutant emission sources to ambient PM at specific locations (AQEG, 2012). Consequently, it is a complicated process to formulate cost-effective policy action to reduce harm caused by PM. The inorganic chemical components of PM -ammonium (NH + 4 ), sulfate (SO 2− 4 ) and nitrate (NO − 3 ) -constitute a major fraction of PM 2.5 (Putaud et al., 2010). The anthropogenic emissions of the gaseous precursors of inorganic PM -ammonia (NH 3 ), sulfur dioxide (SO 2 ) and nitrogen oxides (NO x ) -are also subject to various legislation that seeks to limit and reduce either a country's total emissions or the emissions from individual sources or source sectors (Heal et al., 2012;Reis et al., 2012). For SO 2 and NO x in particular, emissions reductions have been very effective over the past few decades and this is reflected in reductions in ambient concentrations of the gases (RoTAP, 2012). Despite this, PM 10 concentrations across much of western Europe have not fallen significantly since the year 2000 .
The longer lifetime of secondary PM components compared with their gaseous precursors means that transboundary transport from Europe and meteorology are important drivers. Previous studies suggest that transatlantic transport of these secondary inorganic aerosol (SIA) species has a small effect on EU surface SIA concentrations and deposition (Sanderson et al., 2008;Simpson et al., 2012), hence "transboundary" hereafter refers to Europe. This is of particular relevance for the design of air quality policies seeking to reduce PM concentrations, especially as some limit values may be sensitive to a small number of high-concentration episodes rather than long-term average concentrations. This is particularly important for the nitrate component which has been shown to be the dominant component on days when PM 10 exceeds 50 µg m −3 (Yin and . There remains a gap in understanding the extent to which domestic emissions and transboundary import of secondary inorganic PM contribute inter-annually and to episodes of elevated concentrations in the UK (RoTAP). This was the motivation for this work. Ambient concentrations of the inorganic components have been measured since the 1990s on a monthly average basis, as part of the UK Acid Gas and Aerosol Network (AGANet http://uk-air.defra.gov.uk/ networks/network-info?view=aganet, see Tang et al. (2009) for description of the approach), providing a data set against which to compare model output.
In Sect. 2 the modelling approach using the EMEP4UK Eulerian atmospheric chemistry transport model (ACTM) (Vieno et al., 2009(Vieno et al., , 2010 simulations and AGANet measurements are fully described. In Sect. 3, first the model performance is evaluated against these AGANet measurements and then the results of sensitivity simulations to assess the contributions of trans-boundary and domestic emissions to secondary inorganic particle concentrations in the UK and their inter-annual variability are assessed. Section 4 discusses this novel decadal inter-comparison and attribution results and conclusions are presented in Sect. 5.

Model description and setup
The EMEP4UK model used for this work is a nested regional ACTM based on version v3.7 of the main EMEP MSC-W model . A detailed description of the EMEP4UK model framework and setup are given in Vieno et al. (2010) and only brief relevant details are presented here.
The EMEP4UK model is driven by the Weather Research Forecast (WRF) model version 3.1.1 (http://www.wrf-model. org). The model horizontal resolution scales down from 50 km × 50 km in the main EMEP "Greater European" domain to 5 km × 5 km for the domain covering the British Isles (Fig. 1). The boundary conditions for the inner domain are derived from the results of the European domain in a one-way nested setup. The EmChem09 chemical scheme was chosen for the present study, as it has been extensively validated at the European scale (Simpson et al., 2012, www.emep.int). The EMEP model is based on Berge and Jakobsen (1998), but extended with photo-oxidant chemistry (Andersson-Skold and Simpson, 1999; Simpson et al. (2012). Gas/aerosol partitioning used the EQSAM formulation (Metzger et al., 2002a, b). The calculated nitrate is then split into coarse and fine mode using a parameterised approach dependent on relative humidity, as described by Simpson et al. (2012). In this version of the EMEP model, nitrate is the only secondary inorganic component present in PM coarse (the difference between PM 10 and PM 2.5 ). This split between PM 2.5 and PM 2.5−10 for nitrate is rather uncertain as discussed in Aas et al. (2012); a more explicit aerosol scheme is under development. The EQSAM scheme used here is equivalent to the EQSAM2 scheme used in the global model TM5 (Karl et al., 2009;Huijnen et al., 2010). Anthropogenic emissions of NO x , NH 3 , SO 2 , primary PM 2.5 , primary PM coarse , CO and non-methane volatile organic compounds are included. PM 10 is the size fraction of particles with an aerodynamic diameter ≤ 10 µm. For the UK, emissions values are taken from the National Atmospheric Emission Inventory (NAEI, http://naei.defra.gov.uk) at 1 km 2 resolution and aggregated to 5 km × 5 km resolution. The underpinning methods by which these emission inventories have been established are reported by Hellsten et al. (2008) and Dore et al. (2008). For the rest of the outer domain, the model uses the EMEP 50 km × 50 km resolution emission estimates provided by the Centre for Emission Inventories and Projections (CEIP, http://www.ceip.at/). Emissions estimates for international shipping (ENTEC, 2010) are aggregated to 5 km × 5 km for those emissions within the inner domain. The EMEP(4UK) model uses a yearly boundary condition for SIA at the edge of the European domain adjusted for each year as described in Simpson et al. (2012).

Model experiments
Hourly surface concentrations of several pollutants, including NO 2 , SO 2 and NH 3 , and particle NO − 3 , SO 2− 4 and NH + 4 , were simulated for the decade 2001-2010. To quantify the influence of long-range (i.e. non-UK, or "transboundary") and short-range (UK, "domestic") emissions on the UK surface concentrations of these components, a perturbation experiment was carried out by setting UK land emissions to zero for the year. This provides an approximate model estimate of the contribution of non-UK emissions to gaseous and PM concentrations in the UK.

Measurement data
Model surface concentrations were compared with observational data from the UK Acid Gases and Aerosols Network (AGANet), which is one of the four component UK Eutrophying and Acidifying Pollutants monitoring networks (Conolly et al., 2011;Tang et al., 2013). The AGANet monitoring sites were established in 1999 for the long-term simul-taneous measurement of the concentrations of SO 2 , HNO 3 and NH 3 gases and particle NO − 3 , SO 2− 4 and NH + 4 , in relation to changes in European emissions of SO 2 , NO x and NH 3 . Measurements are made using DELTA system (DEnuder for Long-Term Atmospheric sampling) and have monthly averaged time resolution (Sutton et al., 2001). The size cut-off of the DELTA sampler has been estimated to be ∼ 4.5 µm , therefore the measured concentrations are between the PM 2.5 and PM 10 size fractions.
The EMEP4UK model assigns all SO 2− 4 and NH + 4 components to PM 2.5 . Modelled NO − 3 is assigned to both PM 2.5 and PM coarse which leads to potential negative bias in modelled versus measured concentrations for NO − 3 . Four sites representing different areas of the UK (marked on Fig. 1) have been selected for the comparisons presented here: Strathvaich Dam (northwest Scotland); Bush (central Scotland); Rothamsted (southeast England); and Yarner Wood (southwest England).

Results
The time series of the modelled and observed monthly mean surface concentrations of particle NO − 3 and SO 2− 4 at the four selected AGANet sites are shown in Figs Figures 2 and 3 show generally good agreement between the two data sets for three of the four sites included here, as illustrated quantitatively by the correlation and linear regression statistics for these particlephase components and for the gas-phase species HNO 3 and SO 2 ( Table 1). The model-measurement comparison at the Strathvaich Dam site is adversely impacted by two extreme measurement values in 2006 and 2007 not present in the simulations. No anomaly is present in the meteorology at this location for these two months. From an analysis of the mass balance of the aerosol components, the two anomalous data points appear to be outliers which may be attributed to sampling or analytical contamination in determination of NO − 3 and SO 2− 4 . The anomaly could be potentially due to local influence or an unusually high positive artefact on the HNO 3 . Recent investigations indicate that the AGANet HNO 3 observation using the DELTA methodology includes a positive bias from other NO y chemical species, which could include HONO (heterogeneously oxidised), N 2 O 5 and PANs (Peroxyacetyl nitrate). However, there is nothing to indicate that Strathvaich Dam should be affected more from this than other sites.
The spatial pattern across the British Isles and the interyear variability of modelled annual mean surface concentrations are shown in Fig. 4 for NO 2 , SO 2 and NH 3 and in    (Fig. 5b). The spatial distribution of NH 3 shows a very different pattern to the other modelled components, with highest modelled concentrations in Brittany and northwest France and northwest England, reflecting the distribution of modelled NH 3 emissions which mainly arise from agricultural sources. The concentrations of the particle components NO − 3 , SO 2− 4 and NH + 4 are spatially smoother across the UK than the gaseous precursors (Figs. 4 and 5). The modelled annual surface concentrations of NO 2 and SO 2 ( Fig. 4a and b) show that the concentrations of these gaseous components decline   To highlight the role of UK sources, the differences between the base simulations and the simulations with zero UK emissions are shown in the lower panels of Fig. 6a and b, with the data expressed as the percentage of the modelled concentrations that are directly attributable to UK domestic emissions (i.e. 100 × (Base Run − Experiment)/Base), again as monthly averages. While the lower maps clearly show the dominating contribution of UK domestic sources to NO 2 and SO 2 concentrations over mainland UK, a smaller contribution in the vicinity of major shipping channels reflects the fact that the scenario treated international shipping as part of the non-UK emissions. Figure 7a and b show similar model results to Fig. 6 but for surface concentrations of particle NO − 3 and SO 2− 4 , respectively. For these components, there is a smaller percentage contribution from UK sources than for SO 2 and NO x concentrations.
The highest concentrations of NO 2 and SO 2 occurred during February and March (Fig. 6), with highest concentrations for NO − 3 and SO 2− 4 occurring during February, March and April (Fig. 7). Figure 7 shows that, for February, up to 40 % of the monthly average NO − 3 concentrations over the UK are attributable to UK emissions. In March and April, the UK contribution to NO − 3 concentrations rises to up to 80 %. The characteristic differences between these three periods are illustrated in Fig. 9. Here the 12:00 wind vector is superimposed to the mean modelled surface concentration of PM NO − 3 for selected days during the three component episodes. It is seen that 12-15 February (episode F) and 17-20 March (episode M) were associated with stagnant air masses allowing NO − 3 PM concentrations to build up, while the period 11-14 April (episode A) was associated with a highlight polluted air mass arriving from the east.   The UK February episode was associated with an easterly light wind advecting PM NO − 3 produced in the area of the north of France, Holland, north of Germany, and Denmark, where the centre of the high pressure was located (Fig. 9). During the March episode, the centre of the high pressure was over the UK with an associated light wind, clear sky, and cooler conditions leading to the accumulation of NO − 3 from UK emissions with little import of NO − 3 or its precursors from outside the UK. The April episode was a mixture of conditions described for February and March.
The model sensitivity analyses of the proportions of UK nitrate and sulfate derived from UK emissions of anthropogenic precursors was extended over the whole period 2001-2010, and the results for the locations of the four study sites, Strathvaich Dam, Bush, Rothamsted and Yarner Wood (highlighted in Fig. 1) are shown in Fig. 10. The 10 years analysed here shows that the monthly averaged UK emissions contributions to SO 2− 4 and NO − 3 at these sites range from 10 to 80 %. Yarner Wood and Strathvaich Dam are closer than the other selected sites to areas of shipping emissions, therefore on average the SO 2− 4 concentration at this site is less influenced by UK emissions compared with the other two sites.
Based on the simulations it is possible to estimate the annual contribution of non-UK emissions to the different components of PM 10 at the four study sites. This is summarised in Fig. 11 for the year 2003, also including the contribution of primary particulate matter (emitted PM). Pollution import for PM 2.5 from non-UK sources ranges from an estimated 41 % for Bush 1, up to 63 % for Yarner Wood, highlighting the importance of transboundary pollution import on UK PM 2.5 concentrations. The same model results for 2003 can be expressed in terms of the contribution of non-UK emissions to the current European Commission (EC, 2013) limit value for PM 2.5 and to the World Health Organization (WHO, 2005) guideline value for PM 2.5 at each of the four sites (Table 2). For these example sites, up to 18 and 45 %  Figure 11. Mean composition of PM 10 components as estimated by the EMEP4UK model for four sites across the UK, averaged for the whole of 2003. The model base run (including all national and international emissions) is compared with the results from a simulation excluding UK emissions (*). The difference in magnitudes between the pairs of adjacent bars indicates the PM derived from emissions within the UK. As well as the SIA components, the total modelled PM 10 includes the contribution from emitted primary fine PM (PM 2.5 ) and primary PM coarse (i.e. PM 2.5−10 ), and fine and coarse sea salt.
of the limit and guideline values, respectively, is provided by non-UK emissions.

Discussion
Inorganic particle components were simulated over the period 2001-2010. This is the first time that high spatial resolution (5 km) and temporal resolution (1 h) simulations of inorganic atmospheric species have been undertaken across the whole UK for a multi-year period, and the first time that the EMEP(4UK) simulations have been compared with the UK-wide AGANet monitoring network. Two inorganic aerosol schemes were available for the EMEP and EMEP4UK model: the EQSAM (used in this work) and the MARS scheme . As discussed in Sect. 2.1, both schemes use a bulk approach for particle formation. The EQSAM aerosol scheme was used here as it has demonstrated good performance in the TM5 atmospheric chemistry transport model (Karl et al., 2009;Huijnen et al., 2010). However, the bulk approach may lead to uncertainties in the simulated SIA, as shown in Hu et al. (2008), as the particle sizes are not explicitly resolved in the model. The current aerosols scheme and size partitioning in the EMEP model has been validated and compared with observations across Europe as shown in Fagerli and Aas (2008) and in Simpson et al. (2006). In addition, in a recent model intercomparison (Carslaw, 2011a, b) SIA and its gaseous precursors simulated by EMEP4UK showed good agreement with observations. The smoother distribution of particle components (Figs. 5 and 7) as compared with their gaseous precursors (Figs. 4 and 6) reflects the longer timescales for forming these secondary pollutants, as compared with the emissions-driven patterns for the primary pollutant gases (AQEG, 2012). The lifetime for oxidation of NO x and SO 2 to HNO 3 and H 2 SO 4 is up to a few days and comparable to transnational air-mass transport times. Hence the lifetime of formation plays an important role in determining the influence of non-UK emissions on SIA concentrations in the UK.
The highest modelled concentrations over this period are in 2003, particularly for PM NO − 3 and NH + 4 , and to a lesser extent for SO 2− 4 , whilst lowest concentrations for each of these components are in 2008-2010.The notably high PM NO − 3 concentrations in February to April 2003 were observed at AGANet stations across the UK and could be well reproduced by the model (Fig. 2, Table 1). Concentrations of PM SO 2− 4 were also elevated during this period, although by a smaller amount, and were also well captured by the model (Fig. 3, Table 1). The magnitude of this elevation in annual average PM NO − 3 concentration in 2003 is greater than the decline in annual average concentration across the whole decade to 2010 of 0.1-0.2 µg N m −3 (Fig. 5). The August 2003 heatwave (Vieno et al., 2010) was not associated with high nitrate as the higher temperature limits the partitioning to the condensed phase. However, a secondary peak in sulfate is noted during summer 2003, which is directly attributed to the 2003 August heatwave, whereby elevated temperatures lead to faster SO 2 oxidation to sulfate (Dawson et al., 2007;Jacob and Winner, 2009).
Although the magnitude of monthly/daily elevated NO − 3 is similar for the three months of February, March and April 2003, each month has a different characteristic. A distinctive meteorological feature for the three months was a persistent high pressure over the UK and Europe (unusual for this season) with an associated relatively cool temperature and little rainfall (not shown). The location and persistency of the high pressure strongly influenced the production and transport of NO − 3 . Although emissions of NO − 3 precursors are controlled, the model analysis shows the substantial influence of meteorology underpinning the high concentrations of NO − 3 observed in the UK during the first part of 2003. Wang et al. (2014) examined the drivers of PM concentrations in the Shanghai region. Similar to our results for the UK they showed that meteorology determined whether the dominant contributor to PM concentrations was local emissions or regional transport. The authors suggest that particular attention should be given to emissions controls in the upwind adjacent provinces, as well as in local areas, for developing effective strategies to reduce PM 2.5 pollution in Shanghai, again consistent with our conclusions. Zhang et al. (2014) also found that PM concentrations in central China have a clear link with long-range transport. A recent study in the USA by Mwaniki et al. (2014) showed nitrate to have a large variation in winter time, contributing substantially to elevated PM events.
The geographic origins of the PM episodes have been investigated in the model perturbation experiment. The monthly average surface concentrations for the zero UK emissions experiment show that surface concentrations of SO 2 and NO 2 are mainly driven by UK emissions (Fig. 6) and by similar proportions of UK emissions throughout the period of high surface concentrations of NO − 3 . However, the proportions of the NO − 3 that are derived from UK and non-UK emissions changes between months (Fig. 7). The model results show that for February 2003 trans-boundary emissions had a small influence on NO − 3 , whereas for March and April the trans-boundary transport of NO − 3 and/or its precursors was substantial. Abdalmogith et al. (2006) suggest that the annual average import of NO − 3 aerosol to the UK from Europe (as an average of 2002 and 2003) is between 35 and 65 % of the UK total NO − 3 concentration. Our study has found that, for 2003 ( Fig. 7), the import to the UK from Europe was in the range 20-60 % of UK total NO − 3 concentrations, with this proportion varying between the three episodes (labelled F, M and A in Fig. 8). Abdalmogith et al. (2006) concluded that the 2003 NO − 3 spring event was not well represented by their model, and the low emissions resolution (10 km × 10 km grid) was suggested as a possible cause. In the present study the elevated NO − 3 concentrations are well represented by the EMEP4UK model at 5 km × 5 km resolution. However, we find that simulation at 50 km × 50 km horizontal spatial resolution of the EMEP4UK model outer domain also represented these features (results not included here), indicating that transport and dispersion were the main drivers of the pollution events. As shown in Fig. 10, over the full 10-year period there was a substantial variation (10 to 80 %) in the contribution of UK emissions to SIA concentrations in the UK.
The simulated changes in the gaseous precursors for 2001-2010 follow the reductions in UK emissions over that period especially for NO 2 and SO 2 (MacCarthy et al., 2012). The change of SO 2 annual surface concentration especially after 2007 over the North Sea (Fig. 4b) is a direct response to the introduction of a sulfur emission control area (SECA) in the North Sea, including the English Channel, by the 2007 MARPOL convention on marine pollution (Dore et al., 2007). Under the convention the sulfur content of bunker fuel was restricted to 1.5 % by mass in 2007 (and will be further reduced to 0.1 % in SECAs by 2020). This has resulted in a substantial reduction of emissions of SO 2 from the shipping sector.
The results in Figs. 4 and 5 illustrate the non-linear relationship between changes over time in SO 2 and NO 2 surface concentrations over the 2001-2010 decade and changes in the respective PM SO 2− 4 and NO − 3 concentrations. The sensitivity of PM SO 2− 4 to changes in its precursors is, however, considerably greater than for NO − 3 . The small decline in NO − 3 and low sensitivity to UK NO x emission found in this work was supported by the results in Harrison et al. (2013). The formation of both NO − 3 and SO 2− 4 requires NH + 4 as a counter-ion and there appear to be sufficient NH 3 emissions not to be a limiting factor to SO 2− 4 formation. Conversely, UK NO x emissions are still relatively high, especially in urban areas, so with an abundance of NO x available for formation of ammonium nitrate available NH 3 eventually may be consumed. Consequently, in areas of high NO x emissions, NO − 3 formation appears to be more sensitive to NH 3 emissions than is the case for SO 2− 4 formation. This is consistent with Redington et al. (2009) whose modelling showed that SO 2− 4 formation in the UK was less sensitive to a 30 % NH 3 emissions reduction than NO − 3 formation. The modelled annual average NH + 4 shows a change between 2001 and 2010 over the UK which is intermediate between that of NO − 3 and SO 2− 4 (Fig. 5c). By 2010, NH + 4 concentrations decreased by 0.3-0.4 µg N m −3 over most of England, but, as was the case for NO − 3 concentrations, annual average NH + 4 concentrations in 2003 were elevated by 0.2-0.3 µg N m −3 compared with preceding and subsequent years. This confirms that the episodes of elevated NO − 3 in 2003 were driven by ammonium nitrate specifically. The modelled decrease in PM NH + 4 concentrations as compared with minimal decrease (and some increase) in NH 3 concentrations over the period 2001-2010 is consistent with the conclusions of Bleeker et al. (2009) andHorvath et al. (2009) for other parts of Europe that reducing SO 2 emissions have contributed to maintaining or even increasing gaseous NH 3 concentrations.
Current EU legislation has established a limit value of 25 µg m −3 for annual mean PM 2.5 for the protection of human health; at the same time, the World Health Organization (WHO) publishes a guideline value of 10 µg m −3 annual mean PM 2.5 for the protection of human health. As Fig. 11 illustrates, determining the contribution of transboundary and regional transport to local PM concentrations is vital to inform policy development, as local measures can only address the local contribution. For the four sites analysed for 2003, Fig. 11 shows the share of non-UK contribution to modelled PM 2.5 concentrations ranging from 63 % (Yarner Wood) to 41 % (Bush 1). It is also clear that PM 10 at these locations is dominated by sea salt. As these stations are representative of rural or background levels, it is likely that the relative longrange contribution to PM 2.5 concentrations at urban hotspots is smaller, but still substantial. Table 2 expresses the non-UK contribution to modelled annual mean PM 2.5 relative to the EC limit value and WHO guideline value for PM 2.5 (for the protection of human health). The non-UK contribution ranges from 5 % at Strathvaich to 18 % at Rothamsted for the limit value at 25 µg m −3 (or 14 to 45 % for the same sites with respect to the guideline value of 10 µg m −3 ). This indicates a clear gradient of non-UK contribution from greatest in the southeast and least in the north; this is likewise visible in Fig. 5.
The results presented here clearly demonstrate the need for international agreements to address the transboundary component of air pollution. If, for instance, an overall limit value of 10 µg m −3 were to be established following the WHO guideline, a substantial number of UK monitoring sites (Fig. 2) in particular in the south and southeast of the country may be close to or exceed annual mean limit values due to import of inorganic particle components from continental Europe under specific conditions.
In the view of these results, the rather moderate further reductions agreed by parties to the Convention on Long-range Transboundary Air Pollution in the revision of the Gothenburg Protocol (Reis et al., 2012) for the period between 2010 and 2020 would result in a substantial remaining contribution of transboundary aerosol transport to UK particulate matter concentrations for the next decade.
The results further illustrate how the inter-annual variability of surface concentrations of nitrate for the 2001-2010 decade as a response to changes in meteorological conditions is larger than the effect of changes in anthropogenic emissions. This suggests that for compliance assessment, an average over several years would provide a more robust basis than individual years, where a few short episodes can have a major influence.

Conclusions
For the first time the EMEP4UK model has been operated at high resolution for a multi-year period (2001-2010) and simulated secondary inorganic component concentrations compared with observations from the AGANet network. The drivers of three remarkably high secondary inorganic aerosol episodes across the UK have been investigated in detail, revealing contrasting causes for different periods. Whilst it has been documented that the bulk gas/particle partitioning approach used in these simulations (EQSAM formulation) may lead to uncertainties in simulated secondary inorganic aerosol, the EMEP4UK model was able to accurately represent both the long-term decadal (2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010) surface concentrations of particulate matter (PM) and specific episodes of elevated PM NO − 3 in 2003. The latter was identified as consisting of three separate episodes, each of less than 1 month duration, in February, March and April. The primary cause of the elevated nitrate levels across the UK was meteorological, related to a persistent high-pressure system, with the contribution of imported pollution differing markedly between these events.
The findings emphasise the importance of employing multiple year simulations in the assessment of emissions reduction scenarios on PM concentrations. The inter-annual variability of surface concentrations of nitrate for the 2001-2010 decade as a response to changes in meteorological conditions is larger than the effect of changes in anthropogenic emissions. For instance, up to 60 % of NO − 3 may be imported from outside the UK under specific conditions.
Our results highlight how inter-annual variability can profoundly affect the sensitivity to the attainment of limit values for ambient PM concentrations as a result of non-domestic contributions from transboundary air pollution transport.