Comment on acp-2022-623

Abstract

A Lagrangian framework is used to evaluate aerosol-cloud interactions in the U.S. Department of Energy's Energy Exascale Earth System Model (E3SM) version 1 (E3SMv1) for measurements taken at Graciosa Island in the Azores where a U.S. Department of Energy Atmosphere Radiation Measurement (ARM) site is located. This framework uses direct measurements of CCN (instead of relying on satellite retrievals of aerosol optical depth) and incorporates a suite of ground-based ARM measurements, satellite retrievals, and meteorological reanalysis products that when applied to over a 1,500 trajectories provides key insights into the evolution of low-level clouds and aerosol radiative forcing that is not feasible from a traditional Eulerian analysis framework. Significantly lower concentrations (40 %) of surface cloud condensation nuclei (CCN) are measured when precipitation rates in 48-hour back trajectories average above 1.2 mm/d in the Integrated Multi-satellitE Retrievalsfor GPM (IMERG) product. The depletion of CCN when precipitation rates are elevated is nearly twice as large in the ARM observations compared to E3SMv1 simulations. The model CCN bias remains significant despite modifying the autoconversion and accretion rates in warm clouds. As the clouds in trajectories associated with larger surface-based CCN advect away from Graciosa Island they maintain higher values of droplet number concentrations (Nd) over multiple days in observations and E3SM simulations compared to trajectories that start with lower CCN concentrations. The response remains robust even after controlling for meteorological factors such as lower troposphere stability, the degree of cloud coupling with the surface, and island wake effects. E3SMv1 simulates a multi-day aerosol effect on clouds and a Twomey radiative effect that is within 30 % of the ARM and satellite observations. However, the mean cloud droplet concentration is more than 2–3 times larger than in the observations. While Twomey radiative effects are similar amongst autoconversion and accretion sensitivity experiments the liquid water path and cloud fraction adjustments are positive when using a regression model as opposed to negative when using the present-day minus pre-industrial day aerosol emissions approach. This result suggests that tuning the autoconversion and accretion alone are unlikely to produce the desired aerosol susceptibilities in E3SMv1.