Long-range and local air pollution: what can we learn from chemical speciation of particulate matter at paired sites?

Here we report results of a detailed analysis of the urban and non-urban contributions to particulate matter (PM) concentrations and source contributions in five European cities, namely Schiedam (the Netherlands, NL), Lens (France, FR), Leipzig (Germany, DE), Zurich (Switzerland, CH) and Barcelona (Spain, ES). PM chemically speciated data from 12 European paired monitoring sites (one traffic, five urban, five regional and one continental background) were analysed by positive matrix factorisation (PMF) and Lenschow's approach to assign measured PM and source contributions to the different spatial levels. Five common sources were obtained at the 12 sites: sulfate-rich (SSA) and nitrate-rich (NSA) aerosols, road traffic (RT), mineral matter (MM), and aged sea salt (SS). These sources explained from 55 % to 88 % of PM mass at urban low-traffic-impact sites (UB) depending on the country. Three additional common sources were identified at a subset of sites/countries, namely biomass burning (BB) (FR, CH and DE), explaining an additional 9 %–13 % of PM mass, and residual oil combustion (V–Ni) and primary industrial (IND) (NL and ES), together explaining an additional 11 %–15 % of PM mass. In all countries, the majority of PM measured at UB sites was of a regional+continental (R+C) nature (64 %–74 %). The R+C PM increments due to anthropogenic emissions in DE, NL, CH, ES and FR represented around 66 %, 62 %, 52 %, 32 % and 23 %, respectively, of UB PM mass. Overall, the R+C PM increments due to natural and anthropogenic sources showed opposite seasonal profiles with the former increasing in summer and the latter increasing in winter, even if exceptions were observed. In ES, the anthropogenic R+C PM increment was higher in summer due to high contributions from regional SSA and V–Ni sources, both being mostly related to maritime shipping emissions at the Spanish sites. Conversely, in the other countries, higher anthropogenic R+C PM increments in winter were mostly due to high contributions from NSA and BB regional sources during the cold season. On annual average, the sources showing higher R+C increments were SSA (77 %–91 % of SSA source contribution at the urban level), NSA (51 %–94 %), MM (58 %–80 %), BB (42 %–78 %) and IND (91 % in NL). Other sources showing high R+C increments were photochemistry and coal combustion (97 %–99 %; identified only in DE). The highest regional SSA increment was observed in ES, especially in summer, and was related to ship emissions, enhanced photochemistry and peculiar meteorological patterns of the Western Mediterranean. The highest R+C and urban NSA increments were observed in NL and associated with high availability of precursors such as NOx and NH3. Conversely, on average, the sources showing higher local increments were RT (62 %–90 % at all sites) and V–Ni (65 %–80 % in ES and NL). The relationship between SSA and V–Ni indicated that the contribution of ship emissions to the local sulfate concentrations in NL has strongly decreased since 2007 thanks to the shift from high-sulfur- to low-sulfur-content fuel used by ships. An improvement of air quality in the five cities included here could be achieved by further reducing local (urban) emissions of PM, NOx and NH3 (from both traffic and non-traffic sources) but also SO2 and PM (from maritime ships and ports) and giving high relevance to non-urban contributions by further reducing emissions of SO2 (maritime shipping) and NH3 (agriculture) and those from industry, regional BB sources and coal combustion.