2021MNRAS.505.2675C

We present the open-source PYTHON Radiative-transfer in a Bayesian framework (PYRAT BAY) framework for exoplanet atmospheric modelling, spectral synthesis, and Bayesian retrieval. The modular design of the code allows the users to generate atmospheric 1D parametric models of the temperature, abundances (in thermochemical equilibrium or constant-with-altitude), and altitude profiles (in hydrostatic equilibrium); sample ExoMol and HITRAN line-by-line cross-sections with custom resolving power and line-wing cutoff values; compute emission or transmission spectra considering cross-sections from molecular line transitions, collision-induced absorption, Rayleigh scattering, grey clouds, and alkali resonance lines; and perform Markov chain Monte Carlo atmospheric retrievals for a given transit or eclipse data set. We benchmarked the PYRAT BAY framework by reproducing line-by-line sampling of ExoMol cross-sections, producing transmission and emission spectra consistent with PETITRADTRANS models, accurately retrieving the atmospheric properties of simulated transmission and emission observations generated with TAUREX models, and closely reproducing AURA retrieval analyses of the space-based transmission spectrum of HD 209458 b. Finally, we present a retrieval analysis of a population of transiting exoplanets, focusing on those observed in transmission with the HST WFC3/G141 grism. We found that this instrument alone can confidently identify when a data set shows H₂O-absorption features; however, it cannot distinguish whether a muted H₂O feature is caused by clouds, high atmospheric metallicity, or low H₂O abundance. Our results are consistent with previous retrieval analyses. The PYRAT BAY code is available at PyPI (pip install pyratbay) and conda. The code is heavily documented (https://pyratbay.readthedocs.io) and tested to provide maximum accessibility to the community and long-term development stability.