Atmospheric Retrievals with petitRADTRANS
E. Nasedkin, P. Mollière, Doriann Blain
Abstract
petitRADTRANS (pRT) is a fast radiative transfer code used for computing emission and transmission spectra of exoplanet atmospheres (Mollière et al., 2019), combining a FORTRAN back end with a Python based user interface.It is widely used in the exoplanet community with 222 references in the literature to date, and has been benchmarked against numerous similar tools, including many listed in (MacDonald & Batalha, 2023).The spectra calculated with pRT can be used as a forward model for fitting spectroscopic data using Monte Carlo techniques, commonly referred to as an atmospheric retrieval (Madhusudhan & Seager, 2009).The new retrieval module combines fast forward modelling with nested sampling codes, allowing for atmospheric retrievals on a large range of different types of exoplanet data.Thus it is now possible to use pRT to easily and quickly infer the atmospheric properties of exoplanets in both transmission and thermal emission. Statement of needAtmospheric retrievals are a cornerstone of exoplanet atmospheric characterisation.Previously this required interfacing with a sampler and writing a likelihood function in order to explore a parameter space and fit input data, following the description provided in (Mollière et al., 2019).pRT now provides a powerful and user-friendly tool for researchers to fit exoplanet spectra with a range of built-in or custom atmospheric models.Various thermal structures, chemistry and cloud parameterisations and opacity calculation methods can be combined and used to perform parameter estimation and model comparison for a given atmospheric spectrum.The new retrieval module standardises data input, parameter setup, likelihood calculation and interfaces with Multinest, allowing for both simple 'out-of-the-box' retrievals as well as fully customisable setups.New tutorials and documentation have been published to facilitate the adoption of the retrieval module as a widely used tool in the exoplanet community.With increasing volumes of both ground-and space-based spectra available, it is necessary for exoplanet researchers to have access to a range of characterisation tools.While many other retrieval codes exist, pRT offers a computationally efficient and highly flexible framework for inferring the properties of a wide variety of exoplanets.While most feature of pRT are available in other retrieval codes, few of the implement the diversity of the pRT feature set; the availability of both correlated-k and line-by-line opacities, free, equilibrium, and disequilibrium chemistry, and multiple cloud implementations in a single framework let it stand out as a uniquely flexible approach to atmospheric retrievals.