The Response of Marine Boundary Layer Clouds to Idealized Climate Perturbations

Abstract:

Stratocumulus clouds, as simulated by a number of large eddy simulation (LES) models, are found to become less reflective in response to both (1) quadrupled carbon dioxide (CO2) concentrations without any warming, and (2) a composite climate perturbation that approximates the CMIP3 multi-model mean response to doubled CO2 over subtropical oceans. Further, shallow cumulus clouds are also found to become less reflective (i.e., to have weaker shortwave cloud radiative effect, SWCRE) in response to the composite climate perturbation, though the magnitude of SWCRE changes are smaller than for stratocumulus. While the shallow cumulus simulations reflect changes in both cloud thickness and fraction, the stratocumulus simulations are fully overcast, so that only changes in cloud optical depth (and not in cloud fraction) may be simulated. In the stratocumulus simulations, the cloud-topped boundary layer as a whole also thins in response to both climate perturbations.

This study is part of the second phase of CGILS, the CFMIP/GCSS Intercomparison of Large-Eddy and Single-Column Models. Steady, idealized forcings based on the ECMWF July climatology are applied at three locations in the Northeast Pacific Ocean, where the simulated equilibrium cloud states change from shallow, well-mixed stratocumulus to cumulus under stratocumulus and to shallow cumulus convection as the locations move away from the coast. The aforementioned composite climate perturbation includes doubled CO2, increases in sea surface temperature (2.2-2.5 K) and inversion strength, and decreases in mean subsidence (5%), free tropospheric relative humidity (1.5%) and surface wind speed (1.5%). Six LES modeling groups participated in the study. Their LESs exhibited good agreement in the cloud responses, except for the response of shallow cumulus clouds to quadrupled CO2, even though the models differ in their numerical discretizations and treatments of subgrid turbulence and microphysics.

Results from a single LES model for two CGILS locations suggest that cloud response to climate perturbations is similar between simulations with steady forcing and those with realistically varying forcing with the same time-mean, as long as the forcings for both simulations are identically perturbed.