Bayesian learning of coupled biogeochemical–physical models
Predictive dynamical models for marine ecosystems are used for a variety of needs. Due to sparse measurements and limited understanding of the myriad of ocean processes, there is however significant uncertainty. There is model uncertainty in the parameter values, functional forms with diverse parame...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Elsevier BV
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/1721.1/153817 |
| _version_ | 1826191406157266944 |
|---|---|
| author | Gupta, Abhinav Lermusiaux, Pierre F.J. |
| author2 | Massachusetts Institute of Technology. Department of Mechanical Engineering |
| author_facet | Massachusetts Institute of Technology. Department of Mechanical Engineering Gupta, Abhinav Lermusiaux, Pierre F.J. |
| author_sort | Gupta, Abhinav |
| collection | MIT |
| description | Predictive dynamical models for marine ecosystems are used for a variety of needs. Due to sparse measurements and limited understanding of the myriad of ocean processes, there is however significant uncertainty. There is model uncertainty in the parameter values, functional forms with diverse parameterizations, level of complexity needed, and thus in the state fields. We develop a Bayesian model learning methodology that allows interpolation in the space of candidate models and discovery of new models from noisy, sparse, and indirect observations, all while estimating state fields and parameter values, as well as the joint PDFs of all learned quantities. We address the challenges of high-dimensional and multidisciplinary dynamics governed by PDEs by using state augmentation and the computationally efficient GMM-DO filter. Our innovations include stochastic formulation and complexity parameters to unify candidate models into a single general model as well as stochastic expansion parameters within piecewise function approximations to generate dense candidate model spaces. These innovations allow handling many compatible and embedded candidate models, possibly none of which are accurate, and learning elusive unknown functional forms. Our new methodology is generalizable, interpretable, and extrapolates out of the space of models to discover new ones. We perform a series of twin experiments based on flows past a ridge coupled with three-to-five component ecosystem models, including flows with chaotic advection. The probabilities of known, uncertain, and unknown model formulations, and of state fields and parameters, are updated jointly using Bayes' law. Non-Gaussian statistics, ambiguity, and biases are captured. The parameter values and model formulations that best explain the data are identified. When observations are sufficiently informative, model complexity and functions are discovered. |
| first_indexed | 2024-09-23T08:55:22Z |
| format | Article |
| id | mit-1721.1/153817 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2025-02-19T04:17:34Z |
| publishDate | 2024 |
| publisher | Elsevier BV |
| record_format | dspace |
| spelling | mit-1721.1/1538172025-02-04T01:53:02Z Bayesian learning of coupled biogeochemical–physical models Gupta, Abhinav Lermusiaux, Pierre F.J. Massachusetts Institute of Technology. Department of Mechanical Engineering Massachusetts Institute of Technology. Center for Computational Science and Engineering Geology Aquatic Science Predictive dynamical models for marine ecosystems are used for a variety of needs. Due to sparse measurements and limited understanding of the myriad of ocean processes, there is however significant uncertainty. There is model uncertainty in the parameter values, functional forms with diverse parameterizations, level of complexity needed, and thus in the state fields. We develop a Bayesian model learning methodology that allows interpolation in the space of candidate models and discovery of new models from noisy, sparse, and indirect observations, all while estimating state fields and parameter values, as well as the joint PDFs of all learned quantities. We address the challenges of high-dimensional and multidisciplinary dynamics governed by PDEs by using state augmentation and the computationally efficient GMM-DO filter. Our innovations include stochastic formulation and complexity parameters to unify candidate models into a single general model as well as stochastic expansion parameters within piecewise function approximations to generate dense candidate model spaces. These innovations allow handling many compatible and embedded candidate models, possibly none of which are accurate, and learning elusive unknown functional forms. Our new methodology is generalizable, interpretable, and extrapolates out of the space of models to discover new ones. We perform a series of twin experiments based on flows past a ridge coupled with three-to-five component ecosystem models, including flows with chaotic advection. The probabilities of known, uncertain, and unknown model formulations, and of state fields and parameters, are updated jointly using Bayes' law. Non-Gaussian statistics, ambiguity, and biases are captured. The parameter values and model formulations that best explain the data are identified. When observations are sufficiently informative, model complexity and functions are discovered. 2024-03-18T18:01:06Z 2024-03-18T18:01:06Z 2023-08 2024-03-18T17:54:13Z Article http://purl.org/eprint/type/JournalArticle 0079-6611 https://hdl.handle.net/1721.1/153817 Gupta, Abhinav and Lermusiaux, Pierre F.J. 2023. "Bayesian learning of coupled biogeochemical–physical models." Progress in Oceanography, 216. en 10.1016/j.pocean.2023.103050 Progress in Oceanography Creative Commons Attribution-Noncommercial-ShareAlike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Elsevier BV arxiv |
| spellingShingle | Geology Aquatic Science Gupta, Abhinav Lermusiaux, Pierre F.J. Bayesian learning of coupled biogeochemical–physical models |
| title | Bayesian learning of coupled biogeochemical–physical models |
| title_full | Bayesian learning of coupled biogeochemical–physical models |
| title_fullStr | Bayesian learning of coupled biogeochemical–physical models |
| title_full_unstemmed | Bayesian learning of coupled biogeochemical–physical models |
| title_short | Bayesian learning of coupled biogeochemical–physical models |
| title_sort | bayesian learning of coupled biogeochemical physical models |
| topic | Geology Aquatic Science |
| url | https://hdl.handle.net/1721.1/153817 |
| work_keys_str_mv | AT guptaabhinav bayesianlearningofcoupledbiogeochemicalphysicalmodels AT lermusiauxpierrefj bayesianlearningofcoupledbiogeochemicalphysicalmodels |