3D Modeling of Moist Convective Inhibition in Idealized Sub-Neptune Atmospheres

TL;DR

This work tests the hypothesis that condensation-driven convective inhibition forms stable layers in hydrogen-rich sub-Neptune atmospheres using 3D CM1 simulations that include turbulent mixing and evaporation. By initializing with isothermal and adiabatic states and applying a moist-condensation framework, the authors show that saturated tracers exceeding the critical threshold $q_{ m crit}$ produce a convectively inhibited layer with near-zero vertical motions and transport dominated by latent heat and weak turbulence, while radiative processes slowly steepen the temperature toward radiative equilibrium. Radiative timescales are lengthy, on the order of tens to hundreds of years in their models, implying slow evolution of these inhibited regions. The findings support the existence of condensation-driven layered structures in sub-Neptune atmospheres, with important implications for tracer transport, observables, and the interpretation of 3D convection in $H_2$-rich worlds.

Abstract

Atmospheric convection behaves differently in hydrogen-rich atmospheres compared to higher mean molecular weight atmospheres due to compositional gradients of tracers. Previous 1D studies predict that when a condensible tracer exceeds a critical mixing ratio in H$_2$-rich atmospheres, convection is inhibited leading to the formation of radiative layers where the temperature decreases faster with height than in convective profiles. We use 3D convection-resolving simulations to test whether convection is inhibited in H$_2$-rich atmospheres when the tracer mixing ratio exceeds the critical threshold, while including processes neglected in 1D, e.g. turbulent mixing and evaporation. We run two sets of simulations. First, we perform simulations initialized on saturated isothermal states and find that compositional gradients can destabilize isothermal atmospheres. Second, we perform simulations initialized on adiabatic profiles which show distinct, stable inhibition layers form when the condensable tracer exceeds the critical threshold. Within the inhibition layer, only a small amount of energy is carried by latent heat flux, and turbulent mixing transports a small amount of tracer upwards, but both are generally too weak to sustain substantial tracer or heat transport. The thermal profile gradually relaxes to a steep radiative state, but radiative relaxation timescales are long. Our results suggest stable layers driven by condensation-induced convective inhibition form in H$_2$-rich atmospheres, including those of sub-Neptune exoplanets.

Figures (14)

• Figure 1: Overview of the atmospheric state from the initially isothermal simulations. In each row, the panels show from left to right, profiles of temperature (K), virtual potential temperature (K), vapor mixing ratio (kg/kg), relative humidity and condensate mixing ratio (g/kg). In the temperature panels, the gray dashed and solid lines plot a dry and moist adiabats, and the gray dot-dashed line plots the temperature profile assuming gray gas pure radiative equilibrium. The dashed gray line in the vapor mixing ratio panels plots the critical vapor mixing ratio of the final state. Blue lines show the initial state while black lines show the final state. In the third row, the final state in the temperature profile lies perfectly on top of the gray moist adiabat curve throughout the domain. The ratio $q_{v,\rm init}/{q_{v,\rm crit}}$ is calculated using the surface values of the initial state of the respective simulation. No initial state is shown in the $q_l$ panels as this was set to be zero everywhere in all the simulations.
• Figure 2: Snapshots of vertical velocity (top row), vapor mixing ratio (middle row), and condensate mixing ratio (bottom row) during the early rapid convective mixing phase in the isothermal simulations. All four isothermal simulations undergo a similar rapid convective mixing phase within the first 12-15 hours. Here, we present the early mixing of the $q_{v,s}~<~q_{\rm crit}$ test case. Mixing initiates near the top of the domain, driven by dense parcels sinking. Within the convecting layer, as compensating rising air parcels become supersaturated, water vapor condenses within them. The condensate then falls through lower cells and evaporates in sub-saturated grid cells. The convecting layer moves down through the atmosphere until it reaches the surface or a neutrally stable layer. In the isothermal simulation initialized with $q_{v,s}~<~q_{\rm crit}$ (shown here), the convecting layer stops and stabilizes at the top of the Guillot stable region rather than reaching the surface as in the other three isothermal simulations. After the initial rapid mixing, the isothermal simulations reach a convectively stable state, which is subsequently altered on a much slower timescale by radiative processes.
• Figure 3: Atmospheric state for the CM1 simulations initialized on (a.) the dry then moist adiabat, (b.) the moist adiabat, and (c.) the superadiabatic state. In each row, the panels show from left to right, profiles of temperature (K), temperature difference between the final and initial state (K), vapor mixing ratio (kg/kg), relative humidity, and vertical cross-sections taken in the middle of the y-domain of vertical velocity (m/s) and condensate mixing ratio (g/kg). In the temperature panels, the gray dashed and solid lines plot a dry and moist adiabats, and the gray dot-dashed line plots the gray gas pure radiative equilibrium temperature profile. The dashed gray line in the vapor mixing ratio panels indicates the critical vapor mixing ratio of the final state. Initial states are illustrated by blue lines in the profile plots, whereas the final state from the CM1 simulation is shown by the solid black line. The cross-sections are shown for the final state. In panels (a) and (b), we observe only the initial stages of steeping towards a pure radiative equilibrium superadiabatic state, as highlighted in the temperature difference profiles. The simulations show long radiative timescales and the results presented here are not in a statistical equilibrium state.
• Figure 4: Vertical velocity (top row) and condensate mixing ratio (bottom row) cross-sections in (x,z) taken in the middle of the y-domain for 5 time steps in the CM1 simulation initialized on the dry adiabat to the level of saturation from which point the initial state followed a moist adiabat. The vertical velocity cross-sections show 3 distinct layers; a dry convective region near the surface, a layer where convection is inhibited and velocities are nearly zero, and a layer of weak moist convection in the upper atmosphere. The corresponding condensate mixing ratio plots shows that a cloud deck forms in the moist convection layer and condensates fall through and evaporate in the convective inhibition layer.
• Figure 5: Domain-mean sensible and latent heat fluxes for the CM1 test case initialized on a dry then moist adiabat, averaged over the final 25 days of the simulation. The sharp drop of to zero sensible heat flux within the stable layer indicates suppressed convection. Latent heat flux dominates transport in the condensation-driven convective inhibition layer, and positive $Q_{\rm latent}$ values suggests upward vapor transport driven by evaporation.