Microphysics and Southern Ocean Cloud Feedback

Abstract

Global climate models (GCMs) change their cloud properties in the Southern Ocean (SO) with warming in a qualitatively consistent fashion. Cloud albedo increases in the mid-latitudes and cloud fraction decreases in the subtropics. This creates a distinctive ‘dipole’ structure in the SW cloud feedback. However, the shape of the dipole varies from model to model. In this thesis we discuss the microphysical mechanisms underlying the SW cloud feedback over the mid-latitude SO. We will focus on the negative lobe of the dipole. The negative SW cloud feedback in the mid-latitudes is created by transitions from ice to liquid in models. If ice transitions to liquid in mixed-phase clouds the cloud albedo increases because ice crystals are larger than liquid droplets and therefore more reflective for a constant mass of water. Decreases in precipitation efficiency further enhance this effect by decreasing sinks of cloud water. This transition is dependent on the mixed-phase cloud parameterization. Parameterizations vary wildly between models and GCMs disagree by up to 35 K on the temperature where ice and liquid are equally prevalent. This results in a wide spread in the model predictions of the increase in liquid water path (LWP, where the path is the vertically integrated mass of water) with warming that drives the negative optical depth cloud feedback. It is found that this disagreement also results in a wide array of climate mean-states as models that create liquid at lower temperatures have a higher mean-state LWP, lower ice water path (IWP), and higher condensed (ice and liquid) water path (CWP). This presents a problem in climate models. GCMs need to have a reasonable planetary albedo in their climate mean-state. We show evidence that GCMs have tuned cloud fraction to compensate for the variation in mid-latitude cloud albedo driven by the mixed-phase cloud parameterization. This tuning results in mid-latitude clouds that are both too few and too bright as well as a strong indirect control of global cloud fraction by the mixed-phase cloud parameterization. As discussed above, ice crystals are so much larger than liquid droplets that a transition from ice to liquid results in a robust increase in albedo, but this effect is modulated by variations in the size of cloud droplets. Cloud droplet size is determined by the prevalence and efficacy of cloud condensation nuclei (CCN). We present observational and modeling data showing that the sources of CCN in the SO are natural and that biogenic sources account for half of the cloud droplet number concentration in summer when biological productivity and sunlight are strongest. This makes it important to accurately represent biogenic CCN sources, especially their depletion as ocean acidification destroys the calcareous marine organisms that generate the majority of CCN. Despite confirming a natural and substantially biogenic source of CCN, both the source terms of CCN and interaction of CCN with liquid clouds are still uncertain. To help validate the cloud-aerosol indirect effect in GCMs we present a recent natural experiment that occurred when the Bárðarbunga-Veiðivötn fissure erupted suddenly releasing several times the total sulfur emission from Europe into the Atlantic. Substantial cloud aerosol indirect effects were observed during the eruption. This natural experiment offers a scenario that may be used in GCMs to validate their modeled cloud-aerosol indirect effect. Overall, accurate representations of liquid and mixed-phase cloud microphysics in the SO are required if we want to model the Earth’s climate sensitivity. Further, efforts to tune around unreasonable portrayals of SO clouds result in long-ranging biases in global cloud properties and feedbacks.