Generative Data‐Driven Approaches for Stochastic Subgrid Parameterizations in an Idealized Ocean Model
Abstract Subgrid parameterizations of mesoscale eddies continue to be in demand for climate simulations. These subgrid parameterizations can be powerfully designed using physics and/or data‐driven methods, with uncertainty quantification. For example, Guillaumin and Zanna (2021, https://doi.org/10.1...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
American Geophysical Union (AGU)
2023-10-01
|
| Series: | Journal of Advances in Modeling Earth Systems |
| Subjects: | |
| Online Access: | https://doi.org/10.1029/2023MS003681 |
| _version_ | 1797635986855821312 |
|---|---|
| author | Pavel Perezhogin Laure Zanna Carlos Fernandez‐Granda |
| author_facet | Pavel Perezhogin Laure Zanna Carlos Fernandez‐Granda |
| author_sort | Pavel Perezhogin |
| collection | DOAJ |
| description | Abstract Subgrid parameterizations of mesoscale eddies continue to be in demand for climate simulations. These subgrid parameterizations can be powerfully designed using physics and/or data‐driven methods, with uncertainty quantification. For example, Guillaumin and Zanna (2021, https://doi.org/10.1029/2021ms002534) proposed a Machine Learning (ML) model that predicts subgrid forcing and its local uncertainty. The major assumption and potential drawback of this model is the statistical independence of stochastic residuals between grid points. Here, we aim to improve the simulation of stochastic forcing with generative models of ML, such as Generative adversarial network (GAN) and Variational autoencoder (VAE). Generative models learn the distribution of subgrid forcing conditioned on the resolved flow directly from data and they can produce new samples from this distribution. Generative models can potentially capture not only the spatial correlation but any statistically significant property of subgrid forcing. We test the proposed stochastic parameterizations offline and online in an idealized ocean model. We show that generative models are able to predict subgrid forcing and its uncertainty with spatially correlated stochastic forcing. Online simulations for a range of resolutions demonstrated that generative models are superior to the baseline ML model at the coarsest resolution. |
| first_indexed | 2024-03-11T12:27:43Z |
| format | Article |
| id | doaj.art-ab24e754d5fd4e2c9bd283afe4d7f57a |
| institution | Directory Open Access Journal |
| issn | 1942-2466 |
| language | English |
| last_indexed | 2024-03-11T12:27:43Z |
| publishDate | 2023-10-01 |
| publisher | American Geophysical Union (AGU) |
| record_format | Article |
| series | Journal of Advances in Modeling Earth Systems |
| spelling | doaj.art-ab24e754d5fd4e2c9bd283afe4d7f57a2023-11-06T06:42:18ZengAmerican Geophysical Union (AGU)Journal of Advances in Modeling Earth Systems1942-24662023-10-011510n/an/a10.1029/2023MS003681Generative Data‐Driven Approaches for Stochastic Subgrid Parameterizations in an Idealized Ocean ModelPavel Perezhogin0Laure Zanna1Carlos Fernandez‐Granda2Courant Institute of Mathematical Sciences New York University New York NY USACourant Institute of Mathematical Sciences New York University New York NY USACourant Institute of Mathematical Sciences New York University New York NY USAAbstract Subgrid parameterizations of mesoscale eddies continue to be in demand for climate simulations. These subgrid parameterizations can be powerfully designed using physics and/or data‐driven methods, with uncertainty quantification. For example, Guillaumin and Zanna (2021, https://doi.org/10.1029/2021ms002534) proposed a Machine Learning (ML) model that predicts subgrid forcing and its local uncertainty. The major assumption and potential drawback of this model is the statistical independence of stochastic residuals between grid points. Here, we aim to improve the simulation of stochastic forcing with generative models of ML, such as Generative adversarial network (GAN) and Variational autoencoder (VAE). Generative models learn the distribution of subgrid forcing conditioned on the resolved flow directly from data and they can produce new samples from this distribution. Generative models can potentially capture not only the spatial correlation but any statistically significant property of subgrid forcing. We test the proposed stochastic parameterizations offline and online in an idealized ocean model. We show that generative models are able to predict subgrid forcing and its uncertainty with spatially correlated stochastic forcing. Online simulations for a range of resolutions demonstrated that generative models are superior to the baseline ML model at the coarsest resolution.https://doi.org/10.1029/2023MS003681deep learninggenerative modelstochastic parameterizationturbulenceocean |
| spellingShingle | Pavel Perezhogin Laure Zanna Carlos Fernandez‐Granda Generative Data‐Driven Approaches for Stochastic Subgrid Parameterizations in an Idealized Ocean Model Journal of Advances in Modeling Earth Systems deep learning generative model stochastic parameterization turbulence ocean |
| title | Generative Data‐Driven Approaches for Stochastic Subgrid Parameterizations in an Idealized Ocean Model |
| title_full | Generative Data‐Driven Approaches for Stochastic Subgrid Parameterizations in an Idealized Ocean Model |
| title_fullStr | Generative Data‐Driven Approaches for Stochastic Subgrid Parameterizations in an Idealized Ocean Model |
| title_full_unstemmed | Generative Data‐Driven Approaches for Stochastic Subgrid Parameterizations in an Idealized Ocean Model |
| title_short | Generative Data‐Driven Approaches for Stochastic Subgrid Parameterizations in an Idealized Ocean Model |
| title_sort | generative data driven approaches for stochastic subgrid parameterizations in an idealized ocean model |
| topic | deep learning generative model stochastic parameterization turbulence ocean |
| url | https://doi.org/10.1029/2023MS003681 |
| work_keys_str_mv | AT pavelperezhogin generativedatadrivenapproachesforstochasticsubgridparameterizationsinanidealizedoceanmodel AT laurezanna generativedatadrivenapproachesforstochasticsubgridparameterizationsinanidealizedoceanmodel AT carlosfernandezgranda generativedatadrivenapproachesforstochasticsubgridparameterizationsinanidealizedoceanmodel |