Receptor modelling of fine particles in southern England using CMB including comparison with AMS-PMF factors

PM2.5 was collected during a winter campaign at two southern England sites, urban background North Kensington (NK) and rural Harwell (HAR), in January-February 2012. Multiple organic and inorganic source tracers were analysed and used in a Chemical Mass Balance (CMB) model, which apportioned seven separate primary sources, that explained on average 53% (NK) and 56% (HAR) of the organic carbon (OC), including traffic, woodsmoke, food cooking, coal combustion, vegetative detritus, natural gas and dust/soil. With the addition of source tracers for secondary biogenic aerosol at the NK site, 79% of organic carbon was accounted for. Secondary biogenic sources were represented by oxidation products of alpha-pinene and isoprene, but only the former made a substantial contribution to OC. Particle source contribution estimates for PM2.5 mass were obtained by the conversion of the OC estimates and combining with inorganic components ammonium nitrate, ammonium sulfate and sea salt. Good mass closure was achieved with 81% (92% with the addition of the secondary biogenic source) and 83% of the PM2.5 mass explained at NK and HAR respectively, with the remainder being secondary organic matter. While the most important sources of OC are vehicle exhaust (21 and 16 %) and woodsmoke (15 and 28 %) at NK and HAR respectively, food cooking emissions are also significant, particularly at the urban NK site (11% of OC), in addition to the secondary biogenic source, only measured at NK, which represented about 26 %. In comparison, the major source components for PM2.5 at NK and HAR are inorganic ammonium salts (51 and 56 %), vehicle exhaust emissions (8 and 6 %), secondary biogenic (10% measured at NK only), woodsmoke (4 and 7 %) and sea salt (7 and 8 %), whereas food cooking (4 and 1 %) showed relatively smaller contributions to PM2.5. Results from the CMB model were compared with source contribution estimates derived from the AMS-PMF method. The overall mass of organic matter accounted for is rather similar for the two methods. However, appreciably different concentrations were calculated for the individual primary organic matter contributions, although for most source categories the CMB and AMS-PMF results were highly correlated (r(2) = 0.69-0.91). In comparison with the CMB model, the AMS appears to overestimate the biomass burning/coal and food cooking sources by a factor of around 1.5 to 2 while estimates of the traffic source are rather similar for each model. The largest divergence is in the primary/secondary organic matter split, with the AMS estimating an appreciably smaller secondary component. Possible reasons for these discrepancies are discussed, but despite these substantial divergences, the strong correlation of the two methods gives some confidence in their application.