Exploring Biases of Atmospheric Retrievals in Simulated JWST Transmission Spectra of Hot Jupiters

Abstract

With a scheduled launch in October 2018, the James Webb Space Telescope (JWST) is expected to revolutionise the field of atmospheric characterization of exoplanets. The broad wavelength coverage and high sensitivity of its instruments will allow us to extract far more information from exoplanet spectra than what has been possible with current observations. In this paper, we investigate whether current retrieval methods will still be valid in the era of JWST, exploring common approximations used when retrieving transmission spectra of hot Jupiters. To assess biases, we use 1D photochemical models to simulate typical hot Jupiter cloud-free atmospheres and generate synthetic observations for a range of carbon-to-oxygen ratios. Then, we retrieve these spectra using TauREx, a Bayesian retrieval tool, using two methodologies: one assuming an isothermal atmosphere, and one assuming a parametrized temperature profile. Both methods assume constant-with-altitude abundances. We found that the isothermal approximation biases the retrieved parameters considerably, overestimating the abundances by about one order of magnitude. The retrieved abundances using the parametrized profile are usually within one sigma of the true state, and we found the retrieved uncertainties to be generally larger compared to the isothermal approximation. Interestingly, we found that using the parametrized temperature profile we could place tight constraints on the temperature structure. This opens the possibility to characterize the temperature profile of the terminator region of hot Jupiters. Lastly, we found that assuming a constant-with-altitude mixing ratio profile is a good approximation for most of the atmospheres under study.