Aerosol Sensitivity in Large-Eddy Simulations of Subtropical Boundary Layer Clouds Observed in MAGIC

The 2012-2013 MAGIC shipborne deployment of the ARM mobile facility sampled a broad range of subtropical marine stratocumulus, cumulus, and transition regimes during cruises between California and Hawaii with a broad suite of observations of cloud structure, aerosol, precipitation and radiative properties taken on the Horizon Spirit container ship. Ship-following large-eddy simulations (LES) made for each of the 13 suitable westbound cruise legs of 4-5 days compare well with the observations without major systematic biases in cloud properties, as we will document in the first part of our presentation. 

This modeling setup (described below) is then used to study the sensitivity of these clouds to imposed changes in cloud droplet number concentration N_d.  Our preliminary analysis suggests that the sensitivity of cloud macrophysical properties (liquid water path, cloud fraction and cloud vertical structure) to N_d is weak, such that the Twomey effect typically dominates ‘cloud lifetime’ effects on downwelling surface and upward top-of-atmosphere shortwave radiation across the range of clouds sampled in MAGIC.

Our simulations use the System for Atmospheric Modeling (SAM 6.10) with Morrison microphysics, RRTMG radiation, and the UM5 scalar advection scheme.  A 6.4x6.4 km doubly-periodic domain and variable vertical resolution are used.  To avoid over-entrainment, a small vertical grid spacing of 5 m is used across the height range 600-2100 m in which sharp inversions capping stratocumulus cloud layers are typically found.  Each simulation is initialized using thermodynamic profiles from the first balloon sounding of each cruise leg.  Sea-surface temperatures are prescribed from observations, and ECMWF analyses are used to derive time-varying geostrophic wind, ship-relative large-scale advective forcing and large-scale vertical velocity. ECMWF vertical velocities are adjusted to relax the lower-tropospheric temperature profile toward observations with a relaxation timescale of 1 day, and temperature and humidity are strongly nudged toward the balloon profiles above 3 km.  Humidity and wind profiles are also gently relaxed toward the observations to prevent substantial mean drifts.

The time-varying cloud droplet number concentration is specified based on the ship-measured accumulation-mode aerosol and an empirical linear regression between this measurement and satellite-retrieved N_d.    The aerosol sensitivity study replaced the default time-varying droplet concentrations for each simulated cruise with fixed values of 50 and 100 cm^-3, consistent with simplifying assumptions used in many operational weather forecast models.