Evaluation of modelled LOTOS-EUROS with observational based PM10 source attribution. (April 2022)

• Record Type: Journal Article
• Authors: Timmermans, R.; van Pinxteren, D.; Kranenburg, R.; Hendriks, C.; Fomba, K.W.; Herrmann, H.; Schaap, M.
• Is Part Of: Atmospheric environment. Volume 14(2022)
• Publication Date: 2022-04
• Subjects: Source attribution -- Air pollution -- Chemistry transport model -- Germany -- Particulate matter -- PMF
• DOI: 10.1016/j.aeaoa.2022.100173
• Languages: English
• ISSNs: 2590-1621

Abstract

Abstract: Due to its serious health impact particulate matter is one of the air pollutants subject to abatement policies. Information on the main sources responsible for high concentrations of pollutants is therefore crucial to enable effective policy measures. In this study we compared two different methods for attribution of particulate matter concentrations to different sources: A tagging approach within the regional chemistry transport model LOTOS-EUROS and an observational method using speciated particulate matter observations and Positive Matrix Factorisation (PMF). The methods have been applied for winter 2016/2017 over Eastern Germany where in wintertime high woodburning emissions, cold temperatures and regular easterly winds can lead to a build-up of pollutant concentrations. The comparison allows the validation of the modelled source attribution for a selection of source categories. The contributions for biomass and total combustion compare well between both methods providing trust in the determined contributions, applied emissions including their timing. The total contribution from combustion is estimated between 3.3-7.7 μg/m 3 (PMF) and 3.3-7.2 μg/m 3 (LOTOS-EUROS) for the 9 stations incorporated in the study. The temporal Pearson correlation coefficient ranges between 0.3-0.64 for total combustion and 0.34 and 0.7 for biomass combustion. The mean absolute contributions for traffic at background stations also compare well with most values between 1.5-2.0 μg/m 3 for PMF and 1–1.6 μg/m 3 for LOTOS-EUROS. A lack of correlation for this contribution however suggests that the model has difficulty in representing the source category traffic in a time consistent manner and developments are needed to improve the temporal distribution of the traffic emissions within the model. The modelled particulate matter concentrations displayed a 20–40% underestimation of the observed concentrations with an increasing bias during high pollution events. The underestimation showed a high correlation with the observed contribution from combustion and secondary particulate matter including ammonium sulfate and organic carbon suggesting that at least a part of the missing mass in LOTOS-EUROS is related to transformation of volatile combustion emissions, likely from solid fuels, to secondary particle mass and missing enhanced formation of sulfate. Implementation of these missing processes would help to improve the source attribution of particulate matter with the LOTOS-EUROS model. Highlights: Observed combustion source contribution mainly represents primary emissions. Evaluation against receptor models allows identification of missing sources in a CTM. Total and biomass combustion contributions compare well. Development of dynamical temporal distributions of traffic emissions is essential.