Towards a regional dust modeling system in the central Middle East: Evaluation, uncertainties and recommendations

• Published in: Atmospheric environment (1994) 2021-02, Vol.246, p.118160, Article 118160
• Main Authors: Kontos, Serafim, Kakosimos, Konstantinos, Liora, Natalia, Poupkou, Anastasia, Melas, Dimitrios
• Format: Article
• Language: English
• Subjects: CAMx; Dust; Middle east; Natural emissions; NEMO; WRF
• ISSN: 1352-2310; 1873-2844
• DOI: 10.1016/j.atmosenv.2020.118160

This study aims towards an accurate regional dust modeling system in the central Middle East area (CME), through the implementation of the state-of-art dust parameterizations. The modeling system consists of the natural emission model NEMO, the meteorological model WRF and the chemistry transport model CAMx. An extensive evaluation of 16 different configurations has been realized, incorporating all the combinations of the components utilized in the state-of-art dust modeling approaches, namely the drag partition, the sandblasting efficiency, the horizontal mass flux, as well two commonly used soil particle size distributions. Daily mean PM10 measurements in Doha, as well the satellite AOD products of MODIS have been used for the quantitative and the qualitative assessment of the simulations. Noteworthy, each of these assessments did not yield to the exact same ranking of the configurations (e.g. best five) but they assisted on identifying clear patterns. For example, a consistent overestimation of the daily mean PM10, when the MacKinnon's drag partition scheme is utilized, was found. On the other hand, the assessments led to best three performing configurations, with common components the Raupach's drag partition scheme and Alfaro and Gomes sandblasting efficiency. Although their overall performance is good, several issues were found i.e. on individual dust events and a mean underestimation during the studied period, ranging from −49 to −75 μg/m3. One cause of this underestimation could be the aerodynamic entrainment, a mechanism usually neglected from the dust modeling approaches. Another cause could be missed regional and/or local sources or underestimation of their emission rates. •MacKinnon's drag partition scheme leads to consistent PM10 overestimation.•Raupach drag partition and Alfaro-Gomes sandblasting schemes combined perform best.•PM10 differences because of White's and Owen's equations are the least important.•Best performing configuration underestimates by ~50 μg/m3.