Coupled online learning as a way to tackle instabilities and biases in neural network parameterizations: general algorithms and Lorenz 96 case study (v1.0)
<p>Over the last couple of years, machine learning parameterizations have emerged as a potential way to improve the representation of subgrid processes in Earth system models (ESMs). So far, all studies were based on the same three-step approach: first a training dataset was created from a hig...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2020-05-01
|
| Series: | Geoscientific Model Development |
| Online Access: | https://www.geosci-model-dev.net/13/2185/2020/gmd-13-2185-2020.pdf |
| _version_ | 1818512493028311040 |
|---|---|
| author | S. Rasp |
| author_facet | S. Rasp |
| author_sort | S. Rasp |
| collection | DOAJ |
| description | <p>Over the last couple of years, machine learning parameterizations have emerged as a potential way to improve the representation of subgrid processes in Earth system models (ESMs). So far, all studies were based on the same three-step approach: first a training dataset was created from a high-resolution simulation, then a machine learning algorithm was fitted to this dataset, before the trained algorithm was implemented in the ESM. The resulting online simulations were frequently plagued by instabilities and biases. Here, coupled online learning is proposed as a way to combat these issues. Coupled learning can be seen as a second training stage in which the pretrained machine learning parameterization, specifically a neural network, is run in parallel with a high-resolution simulation. The high-resolution simulation is kept in sync with the neural network-driven ESM through constant nudging. This enables the neural network to learn from the tendencies that the high-resolution simulation would produce if it experienced the states the neural network creates. The concept is illustrated using the Lorenz 96 model, where coupled learning is able to recover the “true” parameterizations. Further, detailed algorithms for the implementation of coupled learning in 3D cloud-resolving models and the super parameterization framework are presented. Finally, outstanding challenges and issues not resolved by this approach are discussed.</p> |
| first_indexed | 2024-12-10T23:47:39Z |
| format | Article |
| id | doaj.art-ba9bd953d3f243698325de39eaaad32c |
| institution | Directory Open Access Journal |
| issn | 1991-959X 1991-9603 |
| language | English |
| last_indexed | 2024-12-10T23:47:39Z |
| publishDate | 2020-05-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | Geoscientific Model Development |
| spelling | doaj.art-ba9bd953d3f243698325de39eaaad32c2022-12-22T01:28:53ZengCopernicus PublicationsGeoscientific Model Development1991-959X1991-96032020-05-01132185219610.5194/gmd-13-2185-2020Coupled online learning as a way to tackle instabilities and biases in neural network parameterizations: general algorithms and Lorenz 96 case study (v1.0)S. Rasp<p>Over the last couple of years, machine learning parameterizations have emerged as a potential way to improve the representation of subgrid processes in Earth system models (ESMs). So far, all studies were based on the same three-step approach: first a training dataset was created from a high-resolution simulation, then a machine learning algorithm was fitted to this dataset, before the trained algorithm was implemented in the ESM. The resulting online simulations were frequently plagued by instabilities and biases. Here, coupled online learning is proposed as a way to combat these issues. Coupled learning can be seen as a second training stage in which the pretrained machine learning parameterization, specifically a neural network, is run in parallel with a high-resolution simulation. The high-resolution simulation is kept in sync with the neural network-driven ESM through constant nudging. This enables the neural network to learn from the tendencies that the high-resolution simulation would produce if it experienced the states the neural network creates. The concept is illustrated using the Lorenz 96 model, where coupled learning is able to recover the “true” parameterizations. Further, detailed algorithms for the implementation of coupled learning in 3D cloud-resolving models and the super parameterization framework are presented. Finally, outstanding challenges and issues not resolved by this approach are discussed.</p>https://www.geosci-model-dev.net/13/2185/2020/gmd-13-2185-2020.pdf |
| spellingShingle | S. Rasp Coupled online learning as a way to tackle instabilities and biases in neural network parameterizations: general algorithms and Lorenz 96 case study (v1.0) Geoscientific Model Development |
| title | Coupled online learning as a way to tackle instabilities and biases in neural network parameterizations: general algorithms and Lorenz 96 case study (v1.0) |
| title_full | Coupled online learning as a way to tackle instabilities and biases in neural network parameterizations: general algorithms and Lorenz 96 case study (v1.0) |
| title_fullStr | Coupled online learning as a way to tackle instabilities and biases in neural network parameterizations: general algorithms and Lorenz 96 case study (v1.0) |
| title_full_unstemmed | Coupled online learning as a way to tackle instabilities and biases in neural network parameterizations: general algorithms and Lorenz 96 case study (v1.0) |
| title_short | Coupled online learning as a way to tackle instabilities and biases in neural network parameterizations: general algorithms and Lorenz 96 case study (v1.0) |
| title_sort | coupled online learning as a way to tackle instabilities and biases in neural network parameterizations general algorithms and lorenz 96 case study v1 0 |
| url | https://www.geosci-model-dev.net/13/2185/2020/gmd-13-2185-2020.pdf |
| work_keys_str_mv | AT srasp coupledonlinelearningasawaytotackleinstabilitiesandbiasesinneuralnetworkparameterizationsgeneralalgorithmsandlorenz96casestudyv10 |