Evaluating cloud-aerosol interaction parameterizations in the WRF-Chem model against cloud observations during a large volcanic eruption

In 2014/2015, the intense eruption of the Holuhraun/Bárðarbunga volcano in Iceland emitted extreme amounts of SO2, far above the anthropogenic or natural background. This event had major impacts on cloud properties observed by satellite in the Northern Atlantic. Malavelle et al. (2017) showed that many climate models struggled to reproduce these observed impacts on clouds, indicating potential serious issues in the cloud-aerosol interaction frameworks currently used in models. These issues could explain part of the very large uncertainty remaining in current estimates of the radiative effect of aerosol-cloud interactions.Here, we use MODIS observations of cloud properties during the eruption to evaluate 3 different cloud-aerosol interaction approaches of decreasing complexity in the WRF-Chem 4 regional atmospheric model: First, the default model setup, using the Abdul-Razzak and Ghan (2000) parameterization (AR2000), coupling MOSAIC-4bin aerosols to the Morrison-2-moment microphysics. Second, the Thompson & Eidhammer (2014) aerosol-aware microphysics (TE2014), coupled for this study to MOSAIC-4bin aerosols. Third, the default version of TE2014 in WRF 4 using forced offline aerosols, where we replaced the original static aerosol climatology with 3D time-varying aerosols during the eruption. This last simplified approach does not require simulating fully interactive aerosols, and could be used to investigate regional cloud-aerosol processes and radiative forcing at high resolutions and climate time scales at a lower computational cost.In addition, we compare how these 3 cloud-aerosol approaches impact the detailed cloud response during the eruption in terms of cloud microphysical and optical properties, radiative fluxes, and precipitation.