Golden-Retriever: unveiling planetary atmospheres using a unified model framework and advanced sampling techniques
Golden-Retriever: unveiling planetary atmospheres using a unified model framework and advanced sampling techniques
Golden-Retriever is part of an open-source platform that combines multiple observational datasets to retrieve physical information about the atmospheres of planets. The rapid growth of observational data available from ground- and space-based telescopes requires combining the data available in a unified framework to maximise the physical constraints in our data interpretation.
To combine different types of observational data, including high-resolution spectroscopy, we need to build atmospheric models that are fast enough to interact with nested sampling algorithms. Recent advances in GPU computing allow us to achieve this computational challenge without the need for a large computer cluster. The new software we have developed combines well-established models to calculate gas absorption cross-sections (HELIOS-K), equilibrium chemistry (FastChem), and nested sampling algorithms (UltraNest) with new atmospheric models accelerated by GPU computations. Our atmospheric codes have different levels of complexity, from simple 1D atmospheric codes to fully 3D Planetary Climate Models, and are part of our OASIS lab platform (software-oasis.com). We are currently implementing all the tools necessary to allow easy code use, which will soon be made open-source.
We will present the first test results from the Golden-Retriever software. The first tests include retrieval simulations on high-resolution spectroscopy data in emission and transmission of hot Jupiter planets. Our code can robustly constrain physical quantities such as atmospheric chemical abundances, temperature, clouds, planet bulk densities, and orbital parameters.
Golden-Retriever is intended to explore any planetary atmosphere, and our goal is to make it a standard software in the exoplanet community....
Mendonca, Joao M.
cb29fe08-eb94-4fad-8eba-eac1c5de491b
20 June 2022
Mendonca, Joao M.
cb29fe08-eb94-4fad-8eba-eac1c5de491b
Mendonca, Joao M.
(2022)
Golden-Retriever: unveiling planetary atmospheres using a unified model framework and advanced sampling techniques.
Bulletin of the AAS, 54 (5).
Record type:
Meeting abstract
Abstract
Golden-Retriever is part of an open-source platform that combines multiple observational datasets to retrieve physical information about the atmospheres of planets. The rapid growth of observational data available from ground- and space-based telescopes requires combining the data available in a unified framework to maximise the physical constraints in our data interpretation.
To combine different types of observational data, including high-resolution spectroscopy, we need to build atmospheric models that are fast enough to interact with nested sampling algorithms. Recent advances in GPU computing allow us to achieve this computational challenge without the need for a large computer cluster. The new software we have developed combines well-established models to calculate gas absorption cross-sections (HELIOS-K), equilibrium chemistry (FastChem), and nested sampling algorithms (UltraNest) with new atmospheric models accelerated by GPU computations. Our atmospheric codes have different levels of complexity, from simple 1D atmospheric codes to fully 3D Planetary Climate Models, and are part of our OASIS lab platform (software-oasis.com). We are currently implementing all the tools necessary to allow easy code use, which will soon be made open-source.
We will present the first test results from the Golden-Retriever software. The first tests include retrieval simulations on high-resolution spectroscopy data in emission and transmission of hot Jupiter planets. Our code can robustly constrain physical quantities such as atmospheric chemical abundances, temperature, clouds, planet bulk densities, and orbital parameters.
Golden-Retriever is intended to explore any planetary atmosphere, and our goal is to make it a standard software in the exoplanet community....
Text
faf4d823-207c-4c69-89bc-201ebf75b645
- Version of Record
More information
Published date: 20 June 2022
Identifiers
Local EPrints ID: 497553
URI: http://eprints.soton.ac.uk/id/eprint/497553
ISSN: 0002-7537
PURE UUID: a6c7787c-d913-4a0a-9c5a-ab6dae5911ea
Catalogue record
Date deposited: 27 Jan 2025 17:54
Last modified: 22 Aug 2025 02:46
Export record
Contributors
Author:
Joao M. Mendonca
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics