Characterization of aerosol sources in Leon (Spain) using Positive Matrix Factorization and weather types

Y A one-year aerosol sampling campaign, between 2016 and 2017, was conducted in a suburban area of Leon city, Spain. An association between the Positive Matrix Factorization (PMF) results and air masses through circulation weather types was carried out, through the construction of linear models from the PM10 concentrations and its chemical composition. The aerosol sources, identified by PMF six-factor solution, were: traffic (29%), aged sea salt (26%), secondary aerosols (16%), dust (13%), marine aerosol (7%) and biomass burning (3%). Traffic and secondary factors showed the highest PM10 contribution in the hybrid cyclonic types with wind component from the first and second quadrant. Anticyclonic types with wind component from the first quadrant exhibited high values of secondary, aged sea salt and dust factors. The highest contributions of the dust factor were also associated with northerly types. The linear models built for estimating the source apportionment of PM10, from aerosol chemical composition and geostrophic flow, showed positive coefficients for: westerly flows (WF) in marine factor, southerly flows (SF) in secondary and traffic factors, and shear southerly vorticities (ZS) in dust factor. Negative dependences were observed for ZS in aged sea salt factor and for SF in dust factor. The PM10 mass concentration calculated by the linear models and by the PMF model were strongly correlated. This can be very useful to determine the contribution of a specific source to PM10 in Leon, only by knowing some meteorological and chemical variables. (C) 2020 Elsevier B.V. All rights reserved.

Automated summary

This study conducted a one-year aerosol sampling campaign in a suburban area of Leon city, Spain from 2016 to 2017. The research found that traffic and secondary aerosols had the highest contributions to PM10 in hybrid cyclonic types with wind components from the first and second quadrant. Additionally, the study discovered that the dust factor had the highest contributions in northerly types, and linear models built for estimating the source apportionment of PM10 were strongly correlated with PMF models.