Exoplanet Atmosphere Forecast: Observers Should Expect Spectroscopic Transmission Features to be Muted to 33%

To ensure robust constraints are placed on exoplanet atmospheric transmission spectra, future observations need to obtain high signal-to-noise ratio (S/N) measurements assuming smaller amplitude molecular signatures than those of clear solar metallicity atmospheres. Analyzing 37 exoplanet transmission spectra we find clear solar molecular features are measured in <7% of cases. Clear solar metallicity atmospheres, free from cloud opacities, represent the highest S/N scenario to measure molecular absorption features in exoplanet transmission spectra and are often the maximum assumed atmospheric signal. Some of the most prominent absorption features expected in close-in giant exoplanet spectra are those of water. The distinct absorption feature of H₂O at 1.4 μm has been observed in multiple exoplanet transmission spectra (e.g., Sing et al. 2016; Wakeford et al. 2018) using the Hubble Space Telescopes (HST) Wide Field Camera 3 (WFC3) G141 grism. However, many show little or no water absorption (e.g., Kreidberg et al. 2014; Sing et al. 2016). In this study we determine the statistical significance on the muting of molecular absorption features as a population. The purpose of this study is to perform a global analysis, not to present detailed analyses of each planet.

We find the amplitude of molecular absorption features, 1.1–1.7 μm, is muted to 33% ± 24% of the expected clear solar atmospheric models (Figure 1, top). To obtain this, we compiled measurements of 37 exoplanets observed with HST WFC3 G141 (across 16 programs) in transmission. The data are all publicly available in MAST at 10.17909/t9-6ty4-4f15. For unpublished data sets we use methods outlined in Wakeford et al. (2016) and Stevenson et al. (2014) to measure the transmission spectrum.

Zoom In Zoom Out Reset image size

We fit each transmission spectrum with a 1D isothermal model from the generic ATMO grid (Goyal et al. 2019), where the model has solar metallicity and C/O ratio with no scattering or uniform opacity sources, to represent the best case scenario for atmospheric absorption. Each model is scaled to the individual planetary parameters based on the stellar radius, planetary radius, equilibrium temperature (T_eq), and planetary gravity (g_p).⁴ We use the T_eq and g_p to select the most appropriate model from the grid, and rescale the model to the specific planetary parameters. To fit the model to the data we use a least-squares minimizer with the model defined as S1 = (S0 × p₀) + p₁, where S1 is the scaled model, S0 the clear solar model, p₀ the model amplitude scale factor, and p₁ a baseline offset; we also calculate the uncertainty ${\sigma }_{{p}_{0}}$. As expected, there is a correlation between the uncertainty on the model amplitude and the mean uncertainty on the transmission spectral data. Conversely, we find no correlation between the S/N of the atmospheric transmission and the uncertainty on the model amplitude. An example model fit is shown in Figure 1, bottom left.

To determine the collective distribution of model amplitudes across all data sets we use p₀ and ${\sigma }_{{p}_{0}}$, fitting as a Gaussian distribution from −60% to 150% (Figure 1, bottom right). From each normal distribution we randomly sample 5000 points to determine the global 16th, 50th, and 84th percentile bounds. We find the median and 1σ bounds on the amplitude of molecular transmission features are 33% ± 24%. The model percentage amplitude forms a one-to-one correlation with atmospheric scale heights (H) with the median distribution equivalent to 0.89 ± 0.77 H, encompassing the more conservative value of 1.4 H presented by Fu et al. (2017). Statistically, 30% of the time the amplitude of the observed feature is below 20% or 0.5H, and an approximate clear solar metallicity atmosphere, ≥70% or ≥2H, is measured only 7% of the time.

In summary, we find a median absorption feature amplitude that is one-third the strength derived assuming a clear solar metallicity atmosphere. Our analysis demonstrates that assumptions of clear solar metallicity atmospheres for exoplanets will overestimate the achievable S/N and potentially underestimate observational time requirements ∼93% of the time. To ensure robust constraints on exoplanet atmospheric properties, future observations should account for the likelihood that atmospheric absorption features, particularly in transmission, may be muted due to a broad range of atmospheric properties and processes.