Cloud Feedbacks in Limited-Area and Near-Global Cloud-Resolving Simulations of an Aquaplanet in SAM

Abstract

Global climate models (GCMs) with conventional cumulus parameterization produce a wide spread in cloud feedbacks due to uncertainty from low clouds. Cloud-resolving model (CRM) and large-eddy simulations (LES) can be used to explicitly calculate individual clouds and study cloud-climate responses. This study concentrates on feedbacks of different cloud regimes in ocean-only CRMs of fixed sea-surface temperature (SST), both with small-area and near-global domain sizes. The first part of the thesis focuses on improvement of stratocumulus simulations in a coarse-resolution CRM with a grid spacing of 250 x 250 x 20 m. Three approaches were implemented to help the model maintain stratocumulus liquid water path, but only one is found to be helpful: turning off the subgrid scalar diffusivity except at the surface. The model setup has been further tested for its predictability of cloud feedbacks in various dynamical regimes. The CRM shows some biases in simulating the deep convective clouds and shallow cumulus since the domain size is too small. But it is able to simulate stratocumulus and stratocumulus under cumulus well. The second part focuses on cloud feedbacks in a near-global aquaplanet CRM with fixed meridionally-varying SST, under three climate perturbations: a uniform 4 K SST increase, CO2 quadrupling, and both combined. The CRM has a horizontal resolution of 4 km with no cumulus parameterization. Its domain is a zonally periodic 20480 km-long tropical channel, spanning 46◦S-N with rigid walls. It produces plausible mean distributions of clouds, rainfall and winds. After spin-up, 80 days are analyzed for the control and increased-SST simulations, and 40 days for those with quadrupled CO2. The Intertropical Convergence Zone width and tropical cloud cover are not strongly affected by SST warming or CO2 increase, except for the expected upward shift in high clouds with warming, but both perturbations weaken the Hadley circulation. Increased SST induces a statistically significant increase in subtropical low cloud fraction and in-cloud liquid water content but decreases midlatitude cloud, yielding slightly positive domain-mean shortwave cloud feedbacks. CO2 quadrupling causes a slight shallowing and a statistically insignificant reduction of subtropical low cloud fraction. Warming-induced low cloud changes are strongly correlated with changes in estimated inversion strength, which increases modestly in the subtropics but decreases in the midlatitudes due to poleward jet shifts. Enhanced clear-sky boundary-layer radiative cooling in the warmer climate accompanies the robust subtropical low cloud increase. The probability density function of column relative humidity and precipitation rate across the tropics and subtropics is compared between the control and the increased-SST simulations as a measure of convective aggregation and precipitation extreme, respectively. It shows no evidence of increased aggregation in the warmer climate, while precipitation extreme increases approximately at the Clausius-Clapeyron rate.