Deep learning subgrid-scale parametrisations for short-term forecasting of sea-ice dynamics with a Maxwell  elasto-brittle rheology

Finn, Tobias Sebastian; Durand, Charlotte; Farchi, Alban; Bocquet, Marc; Chen, Yumeng; Carrassi, Alberto; Dansereau, Véronique

doi:10.5194/tc-17-2965-2023

Cited by 6 publications

(7 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…ML models have been shown to be successful at parameterizing subgrid‐scale processes within dynamical models, including ocean mesoscale eddies (Guillaumin & Zanna, 2021), atmospheric convection (Yuval & O’Gorman, 2020), and sea ice dynamics (Finn et al., 2023). Common to each of these studies is that the ML models target specific physical processes, with the aim of replacing pre‐existing knowledge‐based parameterizations, or deriving new parameterizations for physical processes which are not currently represented.…”

Section: Discussionmentioning

confidence: 99%

“…In the context of dynamical climate models, DL algorithms have proven effective tools for deriving model parameterizations directly from numerical simulations. For example, many past studies have focused on learning subgrid parameterizations from high resolution experiments and/or observations of the ocean (Bolton & Zanna, 2019; Sane et al., 2023; Zanna & Bolton, 2020; Zhu et al., 2022), atmosphere (Brenowitz & Bretherton, 2018; Gentine et al., 2018; O’Gorman & Dwyer, 2018; Rasp et al., 2018; P. Wang et al., 2022; Yuval & O’Gorman, 2020), and sea ice (Finn et al., 2023). In the context of DA‐based approaches, some recent studies have relied on iterative sequences of DA and ML to infer unresolved scale parameterizations from sparse and noisy observations (Brajard et al., 2021), or to learn state‐dependent model error from analysis increments (Farchi et al., 2021) and nudging tendencies (Bretherton et al., 2022; Watt‐Meyer et al., 2021), while others have combined DA with equation discovery to extract interpretable structural model errors (Mojgani et al., 2022).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Deep Learning of Systematic Sea Ice Model Errors From Data Assimilation Increments

Gregory,

Bushuk,

Adcroft

et al. 2023

J Adv Model Earth Syst

View full text Add to dashboard Cite

Data assimilation is often viewed as a framework for correcting short‐term error growth in dynamical climate model forecasts. When viewed on the time scales of climate however, these short‐term corrections, or analysis increments, can closely mirror the systematic bias patterns of the dynamical model. In this study, we use convolutional neural networks (CNNs) to learn a mapping from model state variables to analysis increments, in order to showcase the feasibility of a data‐driven model parameterization which can predict state‐dependent model errors. We undertake this problem using an ice‐ocean data assimilation system within the Seamless system for Prediction and EArth system Research (SPEAR) model, developed at the Geophysical Fluid Dynamics Laboratory, which assimilates satellite observations of sea ice concentration every 5 days between 1982 and 2017. The CNN then takes inputs of data assimilation forecast states and tendencies, and makes predictions of the corresponding sea ice concentration increments. Specifically, the inputs are states and tendencies of sea ice concentration, sea‐surface temperature, ice velocities, ice thickness, net shortwave radiation, ice‐surface skin temperature, sea‐surface salinity, as well as a land‐sea mask. We find the CNN is able to make skillful predictions of the increments in both the Arctic and Antarctic and across all seasons, with skill that consistently exceeds that of a climatological increment prediction. This suggests that the CNN could be used to reduce sea ice biases in free‐running SPEAR simulations, either as a sea ice parameterization or an online bias correction tool for numerical sea ice forecasts.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Deep Learning of Systematic Sea Ice Model Errors From Data Assimilation Increments

Gregory,

Bushuk,

Adcroft

et al. 2023

J Adv Model Earth Syst

View full text Add to dashboard Cite

show abstract

“…Meanwhile, only recently have studies begun to investigate ML-based subgrid sea ice parameterizations. For example, Finn et al (2023) successfully demonstrated an ML-based sea ice parameterization which learns short timescale errors associated with subgrid sea ice dynamics in a low resolution sea ice model. Furthermore, Driscoll et al (2023) showed success in emulating a sea ice melt pond parameterization using neural networks.…”

Section: Introductionmentioning

confidence: 99%

“…For example, Finn et al. (2023) successfully demonstrated an ML‐based sea ice parameterization which learns short timescale errors associated with subgrid sea ice dynamics in a low resolution sea ice model. Furthermore, Driscoll et al.…”

Section: Introductionmentioning

confidence: 99%

Machine Learning for Online Sea Ice Bias Correction Within Global Ice‐Ocean Simulations

Gregory,

Bushuk,

Zhang

et al. 2024

Geophysical Research Letters

View full text Add to dashboard Cite

In this study, we perform online sea ice bias correction within a Geophysical Fluid Dynamics Laboratory global ice‐ocean model. For this, we use a convolutional neural network (CNN) which was developed in a previous study (Gregory et al., 2023, https://doi.org/10.1029/2023ms003757) for the purpose of predicting sea ice concentration (SIC) data assimilation (DA) increments. An initial implementation of the CNN shows systematic improvements in SIC biases relative to the free‐running model, however large summertime errors remain. We show that these residual errors can be significantly improved with a novel sea ice data augmentation approach. This approach applies sequential CNN and DA corrections to a new simulation over the training period, which then provides a new training data set to refine the weights of the initial network. We propose that this machine‐learned correction scheme could be utilized for generating improved initial conditions, and also for real‐time sea ice bias correction within seasonal‐to‐subseasonal sea ice forecasts.

show abstract

“…Even though hybrid models can be more difficult to implement than surrogate models, they are often more accurate while reducing data demands (Watson, 2019;Farchi et al, 2021b). There are many examples of hybrid modelling in the geosciences, ranging from data-driven subgrid scale parametrisations (Rasp et al, 2018;Bolton and Zanna, 2019;Gagne et al, 2020;Finn et al, 2023;Ross et al, 2023) to generic model error correction (Bonavita and Laloyaux, 2020;Wikner et al, 2020;Farchi et al, 2021b,a;Brajard et al, 2021;Chen et al, 2022) and super resolution (Barthélémy et al, 2022).…”

Section: Introductionmentioning

confidence: 99%

Online model error correction with neural networks in the incremental 4D-Var framework

Farchi

Chrust

Bocquet

et al. 2022

Preprint

View full text Add to dashboard Cite

Recent studies have demonstrated that it is possible to combine machine learning with data assimilation to reconstruct the dynamics of a physical model partially and imperfectly observed. Data assimilation is used to estimate the system state from the observations, while machine learning computes a surrogate model of the dynamical system based on those estimated states. The surrogate model can be defined as an hybrid combination where a physical model based on prior knowledge is enhanced with a statistical model estimated by a neural network. The training of the neural network is typically done offline, once a large enough dataset of model state estimates is available. By contrast, with online approaches the surrogate model is improved each time a new system state estimate is computed. Online approaches naturally fit the sequential framework encountered in geosciences where new observations become available with time. In a recent methodology paper, we have developed a new weak-constraint 4D-Var formulation which can be used to train a neural network for online model error correction. In the present article, we develop a simplified version of that method, in the incremental 4D-Var framework adopted by most operational weather centres. The simplified method is implemented in the ECMWF Object-Oriented Prediction System, with the help of a newly developed Fortran neural network library, and tested with a two-layer two-dimensional quasi geostrophic model. The results confirm that online learning is effective and yields a more accurate model error correction than offline learning. Finally, the simplified method is compatible with future applications to state-of-the-art models such as the ECMWF Integrated Forecasting System.

show abstract

Deep learning subgrid-scale parametrisations for short-term forecasting of sea-ice dynamics with a Maxwell elasto-brittle rheology

Cited by 6 publications

References 61 publications

Deep Learning of Systematic Sea Ice Model Errors From Data Assimilation Increments

Deep Learning of Systematic Sea Ice Model Errors From Data Assimilation Increments

Machine Learning for Online Sea Ice Bias Correction Within Global Ice‐Ocean Simulations

Online model error correction with neural networks in the incremental 4D-Var framework

Contact Info

Product

Resources

About