Intercomparison of Brown Dwarf Model Grids and Atmospheric Retrieval Using Machine Learning
Understanding differences between substellar spectral data and models has proven to be a major challenge, especially for self-consistent model grids that are necessary for a thorough investigation of brown dwarf atmospheres. Using the random forest supervised machine-learning method, we study the in...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IOP Publishing
2023-01-01
|
| Series: | The Astrophysical Journal |
| Subjects: | |
| Online Access: | https://doi.org/10.3847/1538-4357/ace530 |
| _version_ | 1797697942442737664 |
|---|---|
| author | Anna Lueber Daniel Kitzmann Chloe E. Fisher Brendan P. Bowler Adam J. Burgasser Mark Marley Kevin Heng |
| author_facet | Anna Lueber Daniel Kitzmann Chloe E. Fisher Brendan P. Bowler Adam J. Burgasser Mark Marley Kevin Heng |
| author_sort | Anna Lueber |
| collection | DOAJ |
| description | Understanding differences between substellar spectral data and models has proven to be a major challenge, especially for self-consistent model grids that are necessary for a thorough investigation of brown dwarf atmospheres. Using the random forest supervised machine-learning method, we study the information content of 14 previously published model grids of brown dwarfs (from 1997 to 2021). The random forest method allows us to analyze the predictive power of these model grids, as well as interpret data within the framework of Approximate Bayesian Computation. Our curated data set includes three benchmark brown dwarfs (Gl 570D, ϵ Indi Ba, and Bb) as well as a sample of 19 L and T dwarfs; this sample was previously analyzed by Lueber et al. using traditional Bayesian methods (nested sampling). We find that the effective temperature of a brown dwarf can be robustly predicted independent of the model grid chosen for the interpretation. However, inference of the surface gravity is model-dependent. Specifically, the BT-Settl , Sonora Bobcat , and Sonora Cholla model grids tend to predict $\mathrm{log}g\sim 3$ –4 (cgs units) even after data blueward of 1.2 μ m have been disregarded to mitigate for our incomplete knowledge of the shapes of alkali lines. Two major, longstanding challenges associated with understanding the influence of clouds in brown dwarf atmospheres remain: our inability to model them from first principles and also to robustly validate these models. |
| first_indexed | 2024-03-12T03:47:11Z |
| format | Article |
| id | doaj.art-1ac6034ddf7a4d979420be62b06c2a32 |
| institution | Directory Open Access Journal |
| issn | 1538-4357 |
| language | English |
| last_indexed | 2024-03-12T03:47:11Z |
| publishDate | 2023-01-01 |
| publisher | IOP Publishing |
| record_format | Article |
| series | The Astrophysical Journal |
| spelling | doaj.art-1ac6034ddf7a4d979420be62b06c2a322023-09-03T12:41:49ZengIOP PublishingThe Astrophysical Journal1538-43572023-01-0195412210.3847/1538-4357/ace530Intercomparison of Brown Dwarf Model Grids and Atmospheric Retrieval Using Machine LearningAnna Lueber0https://orcid.org/0000-0001-6960-0256Daniel Kitzmann1https://orcid.org/0000-0003-4269-3311Chloe E. Fisher2https://orcid.org/0000-0003-0652-2902Brendan P. Bowler3https://orcid.org/0000-0003-2649-2288Adam J. Burgasser4https://orcid.org/0000-0002-6523-9536Mark Marley5https://orcid.org/0000-0002-5251-2943Kevin Heng6https://orcid.org/0000-0003-1907-5910University Observatory Munich, Ludwig Maximilian University , Scheinerstrasse 1, Munich D-81679, Germany ; anna.lueber@physik.uni-muenchen.de; Center for Space and Habitability, University of Bern , Gesellschaftsstrasse 6, CH-3012, Bern, SwitzerlandCenter for Space and Habitability, University of Bern , Gesellschaftsstrasse 6, CH-3012, Bern, SwitzerlandDepartment of Physics, Denys Wilkinson Building, University of Oxford , Oxford, OX1 3RH, UKDepartment of Astronomy, The University of Texas at Austin , 2515 Speedway, Stop C1400, Austin, TX 78712, USADepartment of Astronomy & Astrophysics, University of California San Diego , La Jolla, CA 92093, USADepartment of Planetary Sciences and Lunar and Planetary Laboratory, University of Arizona , Tucson, AZ 85721-0092, USAUniversity Observatory Munich, Ludwig Maximilian University , Scheinerstrasse 1, Munich D-81679, Germany ; anna.lueber@physik.uni-muenchen.de; Department of Physics, Astronomy & Astrophysics Group, University of Warwick, Coventry, CV4 7AL, UK; ARTORG Center for Biomedical Engineering Research, University of Bern , Murtenstrasse 50, CH-3008, Bern, SwitzerlandUnderstanding differences between substellar spectral data and models has proven to be a major challenge, especially for self-consistent model grids that are necessary for a thorough investigation of brown dwarf atmospheres. Using the random forest supervised machine-learning method, we study the information content of 14 previously published model grids of brown dwarfs (from 1997 to 2021). The random forest method allows us to analyze the predictive power of these model grids, as well as interpret data within the framework of Approximate Bayesian Computation. Our curated data set includes three benchmark brown dwarfs (Gl 570D, ϵ Indi Ba, and Bb) as well as a sample of 19 L and T dwarfs; this sample was previously analyzed by Lueber et al. using traditional Bayesian methods (nested sampling). We find that the effective temperature of a brown dwarf can be robustly predicted independent of the model grid chosen for the interpretation. However, inference of the surface gravity is model-dependent. Specifically, the BT-Settl , Sonora Bobcat , and Sonora Cholla model grids tend to predict $\mathrm{log}g\sim 3$ –4 (cgs units) even after data blueward of 1.2 μ m have been disregarded to mitigate for our incomplete knowledge of the shapes of alkali lines. Two major, longstanding challenges associated with understanding the influence of clouds in brown dwarf atmospheres remain: our inability to model them from first principles and also to robustly validate these models.https://doi.org/10.3847/1538-4357/ace530Brown dwarfsAtmospheric cloudsAstrostatistics techniques |
| spellingShingle | Anna Lueber Daniel Kitzmann Chloe E. Fisher Brendan P. Bowler Adam J. Burgasser Mark Marley Kevin Heng Intercomparison of Brown Dwarf Model Grids and Atmospheric Retrieval Using Machine Learning The Astrophysical Journal Brown dwarfs Atmospheric clouds Astrostatistics techniques |
| title | Intercomparison of Brown Dwarf Model Grids and Atmospheric Retrieval Using Machine Learning |
| title_full | Intercomparison of Brown Dwarf Model Grids and Atmospheric Retrieval Using Machine Learning |
| title_fullStr | Intercomparison of Brown Dwarf Model Grids and Atmospheric Retrieval Using Machine Learning |
| title_full_unstemmed | Intercomparison of Brown Dwarf Model Grids and Atmospheric Retrieval Using Machine Learning |
| title_short | Intercomparison of Brown Dwarf Model Grids and Atmospheric Retrieval Using Machine Learning |
| title_sort | intercomparison of brown dwarf model grids and atmospheric retrieval using machine learning |
| topic | Brown dwarfs Atmospheric clouds Astrostatistics techniques |
| url | https://doi.org/10.3847/1538-4357/ace530 |
| work_keys_str_mv | AT annalueber intercomparisonofbrowndwarfmodelgridsandatmosphericretrievalusingmachinelearning AT danielkitzmann intercomparisonofbrowndwarfmodelgridsandatmosphericretrievalusingmachinelearning AT chloeefisher intercomparisonofbrowndwarfmodelgridsandatmosphericretrievalusingmachinelearning AT brendanpbowler intercomparisonofbrowndwarfmodelgridsandatmosphericretrievalusingmachinelearning AT adamjburgasser intercomparisonofbrowndwarfmodelgridsandatmosphericretrievalusingmachinelearning AT markmarley intercomparisonofbrowndwarfmodelgridsandatmosphericretrievalusingmachinelearning AT kevinheng intercomparisonofbrowndwarfmodelgridsandatmosphericretrievalusingmachinelearning |