Matthew West — University of Illinois at Urbana-Champaign
Jeffrey Henry Curtis — University of Illinois at Urbana-Champaign
Nicole Riemer — University of Illinois at Urbana-Champaign
Category
General topics – Aerosols
Description
Level of detail that we can gain from WRF-PartMC simulations. We know the full mixing state of the aerosol in each modeling grid cell. The relative contributions of different source sectors are not only known for each grid cell, but on a per-particle level.
Observational data from ASR field campaigns paints a detailed picture of complex aerosol particle compositions and morphologies (including particle shape, phase, and internal structure). These microscale details are crucial in determining the macroscale aerosol impact on climate but are challenging to mesh with highly simplified aerosol representations in large-scale regional and global climate models. Indeed, many global climate model studies have demonstrated that assumptions regarding aerosol microphysics significantly impact model results. While these studies are useful for quantifying sensitivities to “known unknowns” (i.e., within-framework assumptions such as parameter values), they cannot address “unknown unknowns” (i.e., structural uncertainties) of current aerosol models. For the latter, a higher-detail model is needed, and in our case the particle-resolved model WRF-PartMC fulfills this role.
Here we quantify the structural uncertainties of the widely-used modal aerosol models MAM3 and MAM7 due to their simplifying representation of aerosol mixing state, with cloud condensation nuclei concentrations as target quantity. Our model domain is northern California during the CARES campaign in June 2010. We use WRF-Chem with MAM3 and MAM7 as aerosol modules, and compare these results with WRF-PartMC, which uses the particle-resolved aerosol model PartMC-MOSAIC. While MAM3 and MAM7 assume internal mixtures within each of their aerosol modes, WRF-PartMC explicitly resolves the chemical aerosol mixing state. All other model components are kept the same between simulations, and so the differences in cloud condensation nuclei concentrations are strictly due to the differences in aerosol mixing state representation. The uncertainty quantification includes the contributions due to the dynamic evolution of the aerosol within the modal modeling frameworks.
