Convection modeling of pure-steam atmospheres

TL;DR

This study uses high-resolution 3D simulations to explore convection in pure-steam planetary atmospheres, revealing distinct condensing and noncondensing regions and confirming the limitations of existing parameterization schemes.

Contribution

It provides the first detailed 3D analysis of moist convection in non-dilute, pure-steam atmospheres, highlighting the roles of gravity waves and convective overturning.

Findings

Atmospheres have upper condensing and lower noncondensing regions.

Velocities are smaller in the condensing region, with homogeneous temperature variations.

Gravity waves trigger patchy condensation near the boundary.

Abstract

Condensable species are crucial in shaping planetary climate. A wide range of planetary climate systems involve understanding non-dilute condensable substances and their influence on climate dynamics. There has been progress on large-scale dynamical effects and on 1D convection parameterization, but resolved 3D moist convection remains unexplored in non-dilute conditions, though it can have a profound impact on temperature/humidity profiles and cloud structure. We tackle this problem for pure-steam atmospheres using three-dimensional, high-resolution numerical simulations of convection in post-runaway atmospheres where the water reservoir at the surface has been exhausted. We show that the atmosphere is comprised of two characteristic regions, an upper condensing region dominated by gravity waves and a lower noncondensing region characterized by convective overturning cells. Velocities in the condensing region are much smaller than those in the lower noncondensing region, and the horizontal temperature variation is small overall. Condensation in the thermal photosphere is largely driven by radiative cooling and tends to be statistically homogeneous. Some condensation also happens deeper, near the boundary of the condensing region, due to triggering by gravity waves and convective penetrations and exhibit random patchiness. This qualitative structure is insensitive to varying model parameters, but quantitative details may differ. Our results confirm theoretical expectations that atmospheres close to the pure-steam limit do not have organized deep convective plumes in the condensing region. The generalized convective parameterization scheme discussed in Ding & Pierrehumbert (2016) is appropriate to handle the basic structure of atmospheres near the pure-steam limit but is difficult to capture gravity waves and their mixing that appear in 3D convection-resolving models.