Three-dimensional Simulations of Photochemical Hazes in the Atmosphere of Hot Jupiter HD 189733b

Motivation

The transmission spectra of many hot Jupiters show signatures of high-altitude aerosols [e.g., 1,2]. One hypothesized formation mechanism for these aerosols is that photochemical processes generate hazes on the dayside. All previous studies of photochemical hazes on tidally locked giant planets used one-dimensional models [e.g., 3,4]. However, one-dimensional models have to make strongly simplifying assumptions about the strength of vertical mixing. Furthermore, they ignore that the strong day-night contrast on hot Jupiters and atmospheric circulation can lead to inhomogeneous aerosol distributions. For condensate clouds, it has been shown that inhomogeneous aerosol distributions are to be expected [5-7] and can lead to biases in the interpretation of observations [e.g., 8]. The same is likely to be true for photochemical hazes. Further, it has been suggested differences between the morning and evening terminator in ingress and egress transmission spectra could provide a diagnostic for distinguishing between condensate clouds and photochemical hazes [9]. Three-dimensional general circulation models (GCMs) are needed to study how atmospheric circulation shapes the distribution of photochemical hazes to guide future observations and models.

Methods

We present simulations of hot Jupiter HD 189733b using the MITgcm [10]. We use passive tracers representing photochemical hazes to study how hazes are transported by atmospheric circulation. Haze particles in our model have a constant size and are spherical.

Results

The results show that the haze mass mixing ratio varies horizontally by at least an order of magnitude for all particle sizes considered and over the entire simulated pressure range. Depending on the particle size, the resulting 3D haze distribution falls into one of two regimes: small (<30 nm) and large (>30 nm) particles.

For small particles (< 30 nm), the timescale for gravitational settling is longer than the timescale for horizontal and vertical advection. The 3D distribution is thus controlled by advection and looks similar for all particle sizes in this regime. At low pressures, small particle hazes accumulate on the night side in two large midlatitude vortices centered east of the antistellar point (Fig. 1). Because the night side vortices span across the morning terminator, there are higher mass mixing ratios at the morning terminator compared to the evening terminator.

For large particles (>30 nm), the 3D haze distribution is strongly influenced by settling. At very low pressures, where the settling timescale is much shorter than the advection timescale, the horizontal pattern closely mirrors the haze production function. At somewhat higher pressures, where both timescales are within an order of magnitude from each other, hazes are concentrated on the dayside and the hemisphere east of the substellar point (Fig. 2). This results in higher mass mixing ratios at the evening terminator compared to the morning terminator. Because the distribution of hazes is dependent on where in the atmosphere these timescales become equal, the 3D size distributions look much less similar between different particle sizes within this regime compared to the small particle regime.

At pressures > 1 mbar, the advection time scale is shorter than the settling time scale for all particle sizes considered in our simulations. In this region, the equatorial jet dominates the atmospheric circulation and hazes of all sizes develop a more banded pattern (Fig. 3). Differences between morning and evening terminator become smaller in this pressure range.

Our model does not include particle growth and thus does not make predictions about the particle size distribution. To estimate whether terminator differences could be observable, we tried using a constant particle size as well as a size distribution from a 1D microphysics model [3]. In either case, one obtains a relatively small difference in transit depth between leading and trailing limb (Fig. 4). We note that the simulated spectral slope at short wavelengths is too flat to match the observational data. There could be multiple factors explaining why our model does not reproduce the slope of the data, including that a fully coupled three-dimensional microphysics and circulation model might be needed to reproduce the observational data. Furthermore, part of the slope could arise from unaccounted star spots. Another possibility is that mixing from small-scale turbulence not resolved by the GCM could be much more important than expected. In that case, the haze mass mixing ratio would decline more rapidly with increasing pressure, resulting in a steeper spectral slope.

References:

[1] Pont, F., Sing, D. K., Gibson, N. P., et al.: The prevalence of dust on the exoplanet HD 189733b from Hubble and Spitzer observations, Monthly Notices of the Royal Astronomical Society, 432, 4, 2917-2944, 2013.

[2] Sing, D. K., Fortney, J. J., Nikolov, N., et al.: A continuum from clear to cloudy hot-Jupiter exoplanets without primordial water depletion, Nature, 529, 7584, 59-62, 2016.

[3] Lavvas, P. and Koskinen, T.: Aerosol Properties of the Atmospheres of Extrasolar Giant Planets, The Astrophysical Journal, 847, 1, 32, 2017.

[4] Ohno, K. and Kawashima, Y.: Super-Rayleigh Slopes in Transmission Spectra of Exoplanets Generated by Photochemical Haze, The Astrophysical Journal Letters, 895, 2, L47, 2020.

[5] Parmentier, V., Showman, A. P. and Lian, Y.: 3D mixing in hot Jupiters atmospheres I . Application to the day / night cold trap in HD 209458b, Astronomy & Astrophysics, 558, A91, 2013.

[6] Lee, G., Dobbs-Dixon, I., Helling, Ch., et al.: Dynamic mineral clouds on HD 189733b. I. 3D RHD with kinetic, non-equilibrium cloud formation, Astronomy & Astrophysics, 594, A48, 2016.

[7] Lines, S., Mayne, N. J., Boutle, I. A., et al.: Simulating the cloudy atmospheres of HD 209458 b and HD 189733 b with the 3D Met Office Unified Model, Astronomy & Astrophysics, 615, A97, 2018.

[8] Line, M. and Parmentier, V.: The Influence of Nonuniform Cloud Cover on Transit Transmission Spectra, The Astrophysical Journal, 820, 1, 78, 2016.

[9] Kempton E. M. R., Bean J. L., Parmentier V.: An Observational Diagnostic for Distinguishing between Clouds and Haze in Hot Exoplanet Atmospheres, The Astrophysical Journal, 845, 2, L20, 2017.

[10] Adcroft A., Campin J.-M., Hill C., Marshall J.: Implementation of an Atmosphere Ocean General Circulation Model on the Expanded Spherical Cube, Monthly WeatherReview, 132, 2845, 2004.