Refining simulated mineral dust composition through modified size distributions: dual validation with mineral-specific and elemental observations

Abstract

Two 2022 measurement campaigns in Cape Verde provided a unique opportunity to collect mineral dust aerosols from multiple Saharan source regions and characterize their composition. Mineral dust aerosols comprise a complex assemblage of minerals with distinct physico-chemical properties and differentiated climatic impacts through interactions with radiation, cloud microphysics, and atmospheric chemistry. A crucial physical property governing these interactions is the particle size distribution (PSD), influencing aerosol optical properties, transport, and deposition. Although contemporary atmospheric models have begun integrating mineralogical data into their dust aerosol representations, implementation faces complications due to variations in dust emission parameterizations, making compatibility with soil mineralogical databases model-dependent. This work addresses the challenges encountered when incorporating mineralogical information into the COSMO5.05-MUSCAT atmospheric model, which employs the Marticorena and Bergametti (1995) dust emission scheme. We present an improved approach that refines the translation of mineralogical soil PSDs into emitted aerosol PSDs. This improved approach is evaluated using historical Saharan dust measurements and new mineralogical observations from the JATAC2022 and DUSTRISK2022 campaigns. Model performance is assessed using a dual validation framework considering both mineral-resolved and elemental composition. The elemental validation provides additional constraints on model performance, exposing biases in composition that mineral-only comparisons may obscure due to the aggregate nature of mineral dust particles. Results indicate that the proposed modification substantially improves representation of phyllosilicates (illite, kaolinite, and smectite), quartz, and feldspar, while biases in iron, calcium, and magnesium highlight fundamental challenges in representing the heterogeneous internal structure of natural dust particles.