학술논문

The implementation of dust mineralogy in COSMO5.05-MUSCAT.
Document Type
Article
Source
Geoscientific Model Development Discussions. 9/11/2023, p1-37. 37p.
Subject
*MINERAL dusts
*DUST
*CHEMICAL models
*MINERALOGY
*DESERTS
*AEROSOLS
Language
ISSN
1991-9611
Abstract
Mineral dust aerosols are composed from a complex assemblage of various minerals depending on the region they originated. Giving the different mineral composition of desert dust aerosols, different physico-chemical properties and therefore varying climate effects are expected. Despite the known regional variations in mineral composition, chemical transport models typically assume that mineral dust 5 aerosols have uniform composition. This study adds, for the first time, mineralogical information to the mineral dust emission scheme used in the chemical transport model COSMO-MUSCAT. We provide a detailed description of the implementation of the mineralogical database, GMINER (Nickovic et al., 2012), together with a specific set of physical parametrizations in the model's mineral dust emission module. These changes lead to a general improvement of the model performance when comparing the simulated mineral dust aerosols with measurements over the Sahara Desert region for January - February 2022 . 10 The simulated mineral dust aerosol vertical distribution is tested by a comparison with aerosol lidar measurements from the lidar system PollyXT, located at Cape Verde. For a lofted mineral dust aerosol layer on the 2 February 5:00 UTC the lidar retrievals yield on a dust mass concentration peak of 156 µg/m³ while the model calculates the mineral dust peak at 136 µg/m³. The results highlight the possibility of using the model with resolved mineral dust composition for interpretation of the lidar measurements since higher absorption the UV-VIS wavelength is correlated to particles having higher hematite 15 content. Additionally, the comparison with in-situ mineralogical measurements of dust aerosol particles show how important they are, but also that more of them are needed for model evaluation. [ABSTRACT FROM AUTHOR]