Cloud regimes as phase transitions

Abstract

Clouds are repeatedly identified as a leading source of uncertainty in future climate predictions. Of particular importance are stratocumulus clouds, which can appear as either (i) closed cells that reflect solar radiation back to space or (ii) open cells that allow solar radiation to reach the Earth's surface. Here we show that these clouds regimes—open versus closed cells—fit the paradigm of a phase transition. In addition, this paradigm characterizes pockets of open cells as the interface between the open‐ and closed‐cell regimes, and it identifies shallow cumulus clouds as a regime of higher variability. This behavior can be understood using an idealized model for the dynamics of atmospheric water as a stochastic diffusion process. With this new conceptual viewpoint, ideas from statistical mechanics could potentially be used for understanding uncertainties related to clouds in the climate system and climate predictions.

A New Stochastic Model for the Boundary Layer Clouds and Stratocumulus Phase Transition Regimes: Open Cells, Closed Cells, and Convective Rolls

“…From Figure , we see that as expected, when the forcing drives the system to extremely dry conditions, we get an equilibrium state of perfectly clear sky, whereas when F is large and positive we obtain a fully CAF saturated solution; we have a phase transition from the zero state to the full coverage state uniquely determined by F variations alone, independently of J 0 , somewhat similar to the results of Stechmann and Hottovy (). However, the transition layer, its width and location, depend strongly on J 0 .…”

“…Using a vertically uniform model for the total moisture forced by a stochastic white noise process, Stechmann and Hottovy () have recently suggested that the open and closed cell cloud regimes can be viewed as the two main states of a dynamical system undergoing phase transitions. More precisely, Stechmann and Hottovy used the following stochastic partial differential equation to simulate the dynamical evolution of shallow clouds (Stechmann & Hottovy, ), …”

“…Following Stechmann and Hottovy (), we consider the bulk PBL equation for the conservation of total water content, where q = q v + l is the sum of water vapor ( q v ) and cloud condensate (liquid or ice, denoted by l ) averaged from the surface to the top of the cloud layer, u is the corresponding horizontal wind, D is the horizontal diffusion of moisture, and τ s and τ t are the surface evaporation and top of cloud layer turbulent mixing time scales, respectively. Here ∇ = ( ∂ x , ∂ y ) T is the horizontal gradient vector, ∇ 2 = ∂ xx + ∂ yy is the Laplacian operator, Δ s q = q s − q and Δ t q = q − q t are, respectively, the discrepancies between the surface water content, which is set to q s = q ∗ ( T s ) (the saturation vapor mixing ratio at the ground), and q , and between q and the top of the cloud layer water content, q t , or equivalently the low‐level free tropospheric vapor, which we assume is obtained from observational data or a large‐scale model.…”

“…More importantly, while Stechmann and Hottovy use an additive noise model (the strength of the sub‐grid stochastic noise is independent of the large scale moisture), here we take into account the feedback effect of the cloud cover onto the dynamics of the large scale moisture by modifying separately both the top of the cloud detrainment and mixing with the upper‐tropospheric air and the surface evaporation. The new approach represents an important improvement to the Stechmann and Hottovy () model mainly because of the built‐in direct feedbacks, which allow dynamical transitions between the various stratocumulus regimes, as demonstrated by numerical simulations. Also, we specifically used the Ising model with the nearest neighbor interaction with the possibility of multiple equilibria resulting in spatially inhomogeneous CAF distributions characterized by elongated regions of closed cells separated by regions of scattered cloudiness resembling the occurrence of convection rolls in nature.…”

Aerosol indirect effect from turbulence-induced broadening of cloud-droplet size distributions

“…In the atmosphere, feedbacks between the microphysical evolution and σs may exist (e.g., due to varying radiative flux divergence). In this context, the results presented here certainly provide support for similar stochastic differential equation approaches for cloud microphysics (9,20,28) and perhaps even for macroscopic cloud properties (36). It will be intriguing to investigate exactly how the current turbulence-induced aerosol effect relates to that observed in closed parcel studies that consider wide ranges of aerosol concentrations (37,38).…”