Assessing the role of atmospheric dispersion vs. emission strength in the southern Po Valley (Italy) using dispersion-normalised multi-time receptor modelling
Federica Crova,Alice Corina Forello,Vera Bernardoni,Giulia Calzolai,Silvia Canepari,Stefania Argentini,Francesca Costabile,Maria Agostina Frezzini,Fabio Giardi,Franco Lucarelli,Dario Massabò,Lorenzo Massimi,Silvia Nava,Marco Paglione,Giulia Pazzi,Paolo Prati,Matteo Rinaldi,Mara Russo,Sara Valentini,Gianluigi Valli,Virginia Vernocchi,Roberta Vecchi
DOI: https://doi.org/10.1016/j.atmosenv.2023.120168
IF: 5
2024-01-01
Atmospheric Environment
Abstract:In this paper, we applied the Dispersion Normalised Positive Matrix Factorisation (DN-PMF) approach recently proposed in the literature to provide a more realistic picture of the relative importance of emission strength vs. atmospheric dispersion conditions. The disentanglement of such effects is of great concern in pollution hot spots like the Po Valley (Italy), where particulate matter limit values are exceeded despite the existing abatement measures. To explore the potentiality of the DN-PMF approach – still scarcely applied in the literature – a well-chemically characterised PM1 (atmospheric particles with aerodynamic diameter <1 μm) dataset comprising samples collected at different time resolutions at an urban background site (Bologna) in the southern Po Valley was used. Indeed, it is well known that shallow mixing layers promote pollutant accumulation but this observation is not enough to exclude an enhancement of emission strength which could be tackled by appropriate abatement strategies. The source apportionment of sub-micron sized aerosols having a quite long atmospheric residence time in a complex environment like the Po Valley - which is also strongly impacted by secondary aerosol formation on a basin-scale - is generally quite challenging when using receptor models. Due to the availability of a huge dataset with variables having multiple time resolutions, in this work the DN-PMF was implemented in a multi-time resolution approach (MT) to achieve a better source identification and to gain knowledge about the relative importance of atmospheric dilution vs. emissions. A comparison between results obtained by the application of the regular multi time resolution (REG-MT) vs. the DN-MT approach is presented here for the five factors identified (nitrate-dominated, sulphate-dominated, biomass burning, mineral dust, and urban aerosol). The first interesting outcome is that REG-MT and DN-MT results do not point at significant differences in temporal patterns for aerosol components and sources impacting at the basin-scale (i.e. sulphate- and nitrate-dominated aerosol, biomass burning) thus suggesting that the diel modulation of these PM1 emissions is somehow masked by the stronger variability of the mixing layer. Conversely, contributions from local sources with more pronounced diel variation like traffic are quite well reproduced by DN-MT and the ambient concentrations are enhanced compared to REG-MT. This is an important piece of information highlighting that PM1 concentrations from local sources have been likely underestimated by REG-MT assessments. To our knowledge, this is one of the very few applications of DN-MT and the first one at a European site where the huge effort made to implement air pollution containment measures is still not very much effective in reducing PM levels; moreover, in this paper a detailed discussion about the possible interpretation of the output of DN-MT in terms of temporal patterns is reported.
environmental sciences,meteorology & atmospheric sciences