Atmospheric convection plays a key role in the climate of tidally-locked planets:
insights from high-resolution simulations
NASA GISS journal club | 21 April 2020
Denis Sergeev
University of Exeter, UK
This presentation is available at slides.com/denissergeev/2020-04-21-nasa-giss
Exeter Exoplanet Theory Group
Group leader: Nathan Mayne
Staff: Eric Hebrard, F. Hugo Lambert
Postdocs: Duncan Christie, Denis Sergeev, Maria Zamyatina
PhD students: Michelle Bieger, Jake Eager, Robert Ridgway
IT support: Krisztian Kohary


UK Met Office collaborators: Ian Boutle, James Manners, Benjamin Drummond

Exeter Exoplanet Theory Group
Group leader: Nathan Mayne
Staff: Eric Hebrard, F. Hugo Lambert
Postdocs: Duncan Christie, Denis Sergeev, Maria Zamyatina
PhD students: Michelle Bieger, Jake Eager, Robert Ridgway
IT support: Krisztian Kohary


UK Met Office collaborators: Ian Boutle, James Manners, Benjamin Drummond

Talk outline
- Challenges and importance of simulating convection on exoplanets
- Global simulations of Trappist-1e and Proxima Centauri b
- Regional high-resolution simulations and implications for the global climate

Introduction
Why are we interested in terrestrial tidally locked planets?
- Habitable ~ have surface liquid water
- M-dwarf are relatively cool
- Potentially habitable planets around M-dwarfs have to orbit closer
- Strong tidal forces
- Planet can become tidally locked
- Orbital period = rotation period
- Permanent day-side and night-side

Credit: ESA/Hubble

Credit: Engine House Animation Studio
Climate of Earth-like tidally locked exoplanets

Climate of Earth-like tidally locked exoplanets

Climate of Earth-like tidally locked exoplanets

Challenges in convection modeling
- Wide scale separation between convection and planetary-scale circulation
- Computationally expensive
- Convection is subgrid-scale and parameterized
- All existing parameterizations are tested against observations on Earth and still are a big uncertainty for future climate projections
- No in-situ measurements on exoplanets

Modeling framework
Met Office Unified Model
- Full deep-atmosphere non-hydrostatic Navier-Stokes equations
- Full suite of parameterizations
- Seamless modeling: the same model is used for global climate & local high-res modeling
- Adapted for exoplanet atmospheres
Mayne+ (2014)
Drummond+ (2018)
Lines+ (2019)
Boutle+ (2017)
Adaptation of the dynamical core
Climate modeling of a terrestrial exoplanet
Chemistry on Hot Jupiters
Clouds on Hot Jupiters

- Orbital parameters of Trappist-1e and Proxima Centauri b
- Planets are tidally locked to M-dwarf stars
- N2-dominated atmosphere with traces of CO2 and H2O
- Mean surface pressure of 1 bar
- Aquaplanet regime (a slab ocean)
- Two modes
- Global coarse-resolution model with parameterized convection
- Regional high-resolution model with explicit convection

Credit: ESA

Credit: ESA/Hubble

Credit: NASA/JPL-Caltech
Representation of convection in our global simulations
- Standard operational scheme in the UM
- Most sophisticated class of convection schemes
- Involves a lot of "tuning" parameters
- Convective adjustment to a reference state
- Only 2 free parameters
- Using a setup similar to Way+ (2018)
NoCnvPm
Adjust
MassFlux (Control)
- Convection parameterization switched off
Mean global climate and its sensitivity to the convection parameterization
Global simulation of Trappist-1e
Shown are:
- Surface temperature
- Cloud condensate
- Upper troposphere wind vectors
Surface temperature (Ts)
- Ts day-night contrast larger on Proxima b than on Trappist-1e
MassFlux

Surface temperature (Ts)
- Ts day-night contrast larger on Proxima b than on Trappist-1e
MassFlux

- Ts increases globally, especially in cold traps on the night side
- ΔTs,dn reduces
Adjust
Adjust minus MassFlux
Adjust minus MassFlux
Surface temperature (Ts)
- Ts day-night contrast larger on Proxima b than on Trappist-1e
MassFlux

- Ts increases globally, especially in cold traps on the night side
- ΔTs,dn reduces
NoCnvPm
Adjust
- Small global-mean change
- Different response on Trappist1-e vs Proxima b
NoCnvPm minus MassFlux
NoCnvPm minus MassFlux
Adjust minus MassFlux
Adjust minus MassFlux
Surface temperature (Ts) and free troposphere winds

- Ts day-night contrast larger on Proxima b than on Trappist-1e
- Ts increases globally, especially in cold traps on the night side
- ΔTs,dn reduces
NoCnvPm
Adjust
MassFlux
- Small global-mean change
- Different response on Trappist1-e vs Proxima b
NoCnvPm minus MassFlux
NoCnvPm minus MassFlux
Adjust minus MassFlux
Adjust minus MassFlux
Circulation regime (eddy components)
- Planetary-scale equatorial Rossby wave-like pattern
MassFlux

Circulation regime (eddy components)
- Planetary-scale equatorial Rossby wave-like pattern
MassFlux

Circulation regime (eddy components)
- Planetary-scale equatorial Rossby wave-like pattern
MassFlux

Circulation regime (eddy components)
- Planetary-scale equatorial Rossby wave-like pattern
MassFlux

- Almost a mirror-opposite of the MassFlux
- Dominated by extratropical baroclinic gyres
Adjust

Circulation regime (eddy components)
- Planetary-scale equatorial Rossby wave-like pattern
- Almost a mirror-opposite of the MassFlux
- Dominated by extratropical baroclinic gyres
NoCnvPm
Adjust
MassFlux
- Broadly similar response as in Adjust

Circulation regime (eddy components)
- Planetary-scale equatorial Rossby wave-like pattern
- Almost a mirror-opposite of the MassFlux
- Dominated by extratropical baroclinic gyres
NoCnvPm
Adjust
MassFlux
- Broadly similar response as in Adjust

Clouds and precipitation
MassFlux

Clouds and precipitation
MassFlux

Clouds and precipitation
MassFlux

Clouds and precipitation
- Large reduction of clouds on the day side
- More clouds in polar regions of Trappist-1e
Adjust
MassFlux

Clouds and precipitation
- Large reduction of clouds on the day side
- More clouds in polar regions of Trappist-1e
NoCnvPm
Adjust
- Day side: closer to MassFlux
- Night side: closer to Adjust
MassFlux
Summary of global simulations: the effects of convection parameterization
- The global climate of TL exoplanets is altered (and not just the day side!)
- The effect is planet-dependent
- Caution should be taken when interpreting results of GCMs with one parameterization or another.
Extra

Circulation regime

Differences in circulation regimes
- Leconte+ (2013)
- Haqq-Misra+ (2017)

Energy redistribution
- Using the horizontal divergence of the MSE flux, ∫0ztopρ∇⋅(u(cpT+gz+Lvq))dz
- The largest difference is in the moist component (Lq)
- Adjust results in a more intense energy transport

Vertical profiles of temperature and humidity

Vertical cross-sections of heat fluxes at 90∘E
High-resolution simulations

"global" (MassFlux)
"HighRes"
Nested grid setup
Spatial variability: top-of-atmosphere outgoing LW radiation


Spatial variability: top-of-atmosphere outgoing LW radiation

Spatial variability: top-of-atmosphere outgoing LW radiation

Spatial variability: top-of-atmosphere outgoing LW radiation
Histograms of instantaneous TOA OLR


Spatial variability: top-of-atmosphere outgoing LW radiation

Histograms of instantaneous TOA OLR



Vertical cloud structure (averaged profiles)


Vertical cloud structure (averaged profiles)


Vertical cloud structure (averaged profiles)

- Biases in liquid water clouds
- Thicker layer of mixed-phase clouds
- Smaller bias in ice clouds
Latent heating (averaged profiles)

- Latent heating is stronger in the global model, compared to HighRes
- Impact on the upper-level flow divergence?

Extra

Global impact of resolved convection
- Caveat: one-way nesting of the model
- How to assess the impact of HighRes simulations on the global climate?

- Caveat: one-way nesting of the model
- How to assess the impact of HighRes simulations on the global climate?
- Run an ensemble of global experiments with perturbed convection parameters

- Run an ensemble of global experiments with perturbed convection parameters
- Integrate a metric of convection strength over the substellar region

- Caveat: one-way nesting of the model
- How to assess the impact of HighRes simulations on the global climate?
- Run an ensemble of global experiments with perturbed convection parameters
- Integrate a metric of convection strength over the substellar region
- Correlate with a metric of day-night heat redistribution

- Caveat: one-way nesting of the model
- How to assess the impact of HighRes simulations on the global climate?
- Run an ensemble of global experiments with perturbed convection parameters
- Integrate a metric of convection strength over the substellar region
- Correlate with a metric of day-night heat redistribution
- Extrapolate for the convection metric from HighRes simulations

- Caveat: one-way nesting of the model
- How to assess the impact of HighRes simulations on the global climate?
- Run an ensemble of global experiments with perturbed convection parameters
- Integrate a metric of convection strength over the substellar region
- Correlate with a metric of day-night heat redistribution
- Extrapolate for the convection metric from HighRes simulations

- Caveat: one-way nesting of the model
- How to assess the impact of HighRes simulations on the global climate?

- Run an ensemble of global experiments with perturbed convection parameters
- Integrate a metric of convection strength over the substellar region
- Correlate with a metric of day-night heat redistribution
- Extrapolate for the convection metric from HighRes simulations
- Caveat: one-way nesting of the model
- How to assess the impact of HighRes simulations on the global climate?
Summary
- Representation of convection is important for the global climate of exoplanets
- Using a simpler adjustment scheme leads to
- on the day side: reduction of clouds, decrease in albedo (by ~60%)
- on the night side: warming of 20-30 K
- reduction in the day-night temperature contrast
- change in the global circulation
- The sensitivity to convection scheme is planet-dependent
- The choice of convection scheme does not push the simulated climate out of habitable state
- A potential global convection-permitting simulation may enhance the day-night temperature contrast





Future work
- Extending this analysis to different planetary parameters and stellar spectra
- Exploring the sensitivity to the lower boundary condition, e.g. a presence of a continent on the day side
- Coupling the high-resolution simulation with chemical kinetics and space weather
- Global resolved convection (LFRic project at the Met Office)
Thank you!

My website: dennissergeev.github.io
Code for figures: github.com/dennissergeev/exoconvection-apj-2020
My twitter: meteodenny

This presentation: slides.com/denissergeev/2020-04-21-nasa-giss

The paper: arxiv.org/abs/2004.03007
NASA GISS Lunch time seminar talk
By Denis Sergeev
NASA GISS Lunch time seminar talk
- 473