Articles | Volume 17, issue 7
https://doi.org/10.5194/tc-17-2965-2023
© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/tc-17-2965-2023
© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Research article
| 
21 Jul 2023
Research article |
| 21 Jul 2023
| 21 Jul 2023
Deep learning subgrid-scale parametrisations for short-term forecasting of sea-ice dynamics with a Maxwell elasto-brittle rheology
Tobias Sebastian Finn
CORRESPONDING AUTHOR
CEREA, École des Ponts and EDF R&D, Île-de-France, France
Charlotte Durand
CEREA, École des Ponts and EDF R&D, Île-de-France, France
Alban Farchi
CEREA, École des Ponts and EDF R&D, Île-de-France, France
Marc Bocquet
CEREA, École des Ponts and EDF R&D, Île-de-France, France
Yumeng Chen
Dept. of Meteorology and NCEO, University of Reading, Reading, United Kingdom
Alberto Carrassi
Dept. of Meteorology and NCEO, University of Reading, Reading, United Kingdom
Dept. of Physics and Astronomy “Augusto Righi”, University of Bologna, Bologna, Italy
Véronique Dansereau
Laboratoire 3SR, Grenoble INP, CNRS, Université Grenoble Alpes, Grenoble, France
Related authors
Wenbo Yu, Anirbit Ghosh, Tobias Sebastian Finn, Rossella Arcucci, Marc Bocquet, and Sibo Cheng
EGUsphere, https://doi.org/10.5194/egusphere-2025-2836, https://doi.org/10.5194/egusphere-2025-2836, 2025
This preprint is open for discussion and under review for Geoscientific Model Development (GMD).
Short summary
Short summary
We introduce the first denoising diffusion model for wildfire spread prediction, a new kind of generative AI model that learns to simulate fires not just as one fixed outcome, but as a range of possible scenarios. This allows us to capture the inherent uncertainty of wildfire dynamics. Our model produces ensembles of forecasts that reflect physically meaningful distributions of where fire might go next.
Charlotte Durand, Tobias Sebastian Finn, Alban Farchi, Marc Bocquet, Julien Brajard, and Laurent Bertino
EGUsphere, https://doi.org/10.5194/egusphere-2024-4028, https://doi.org/10.5194/egusphere-2024-4028, 2025
Short summary
Short summary
This paper presents a four-dimensional variational data assimilation system based on a neural network emulator for sea-ice thickness, learned from neXtSIM simulation outputs. Testing with simulated and real observation retrievals, the system improves forecasts and bias error, performing comparably to operational methods, demonstrating the promise of sea-ice data-driven data assimilation systems.
Tobias Sebastian Finn, Lucas Disson, Alban Farchi, Marc Bocquet, and Charlotte Durand
Nonlin. Processes Geophys., 31, 409–431, https://doi.org/10.5194/npg-31-409-2024, https://doi.org/10.5194/npg-31-409-2024, 2024
Short summary
Short summary
We train neural networks as denoising diffusion models for state generation in the Lorenz 1963 system and demonstrate that they learn an internal representation of the system. We make use of this learned representation and the pre-trained model in two downstream tasks: surrogate modelling and ensemble generation. For both tasks, the diffusion model can outperform other more common approaches. Thus, we see a potential of representation learning with diffusion models for dynamical systems.
Yumeng Chen, Polly Smith, Alberto Carrassi, Ivo Pasmans, Laurent Bertino, Marc Bocquet, Tobias Sebastian Finn, Pierre Rampal, and Véronique Dansereau
The Cryosphere, 18, 2381–2406, https://doi.org/10.5194/tc-18-2381-2024, https://doi.org/10.5194/tc-18-2381-2024, 2024
Short summary
Short summary
We explore multivariate state and parameter estimation using a data assimilation approach through idealised simulations in a dynamics-only sea-ice model based on novel rheology. We identify various potential issues that can arise in complex operational sea-ice models when model parameters are estimated. Even though further investigation will be needed for such complex sea-ice models, we show possibilities of improving the observed and the unobserved model state forecast and parameter accuracy.
Charlotte Durand, Tobias Sebastian Finn, Alban Farchi, Marc Bocquet, Guillaume Boutin, and Einar Ólason
The Cryosphere, 18, 1791–1815, https://doi.org/10.5194/tc-18-1791-2024, https://doi.org/10.5194/tc-18-1791-2024, 2024
Short summary
Short summary
This paper focuses on predicting Arctic-wide sea-ice thickness using surrogate modeling with deep learning. The model has a predictive power of 12 h up to 6 months. For this forecast horizon, persistence and daily climatology are systematically outperformed, a result of learned thermodynamics and advection. Consequently, surrogate modeling with deep learning proves to be effective at capturing the complex behavior of sea ice.
Tobias Sebastian Finn, Gernot Geppert, and Felix Ament
Hydrol. Earth Syst. Sci. Discuss., https://doi.org/10.5194/hess-2020-672, https://doi.org/10.5194/hess-2020-672, 2021
Revised manuscript not accepted
Short summary
Short summary
Through the lens of recent developments in hydrological modelling and data assimilation, we hourly update the soil moisture with ensemble data assimilation and sparse 2-metre-temperature observations in a coupled limited area model system. In idealized experiments, we improve the soil moisture analysis by coupled data assimilation across the atmosphere-land interface. We conclude that we can merge the separated assimilation cycles for the atmosphere and land surface into one single cycle.
Jozef Skákala, David Ford, Keith Haines, Amos Lawless, Matthew J. Martin, Philip Browne, Marcin Chrust, Stefano Ciavatta, Alison Fowler, Daniel Lea, Matthew Palmer, Andrea Rochner, Jennifer Waters, Hao Zuo, Deep S. Banerjee, Mike Bell, Davi M. Carneiro, Yumeng Chen, Susan Kay, Dale Partridge, Martin Price, Richard Renshaw, Georgy Shapiro, and James While
Ocean Sci., 21, 1709–1734, https://doi.org/10.5194/os-21-1709-2025, https://doi.org/10.5194/os-21-1709-2025, 2025
Short summary
Short summary
UK marine data assimilation (MDA) involves a closely collaborating research community. In this paper, we offer both an overview of the state of the art and a vision for the future across all of the main areas of UK MDA, ranging from physics to biogeochemistry to coupled DA. We discuss the current UK MDA stakeholder applications, highlight theoretical developments needed to advance our systems, and reflect upon upcoming opportunities with respect to hardware and observational missions.
Sebastien Kuchly, Baptiste Auvity, Nicolas Mokus, Matilde Bureau, Paul Nicot, Amaury Fourgeaud, Véronique Dansereau, Antonin Eddi, Stéphane Perrard, Dany Dumont, and Ludovic Moreau
EGUsphere, https://doi.org/10.5194/egusphere-2025-3304, https://doi.org/10.5194/egusphere-2025-3304, 2025
This preprint is open for discussion and under review for The Cryosphere (TC).
Short summary
Short summary
During February and March 2024, we realized a multi-instrument field campaign in the St. Lawrence Estuary, to capture swell-driven sea ice fragmentation. The dataset combines geophones, wave buoys, smartphones, and video recordings with drones, to study wave-ice interactions under natural conditions. It enables analysis of ice thickness, wave properties, and ice motion. Preliminary results show strong consistency across instruments, offering a valuable resource to improve sea ice models.
Wenbo Yu, Anirbit Ghosh, Tobias Sebastian Finn, Rossella Arcucci, Marc Bocquet, and Sibo Cheng
EGUsphere, https://doi.org/10.5194/egusphere-2025-2836, https://doi.org/10.5194/egusphere-2025-2836, 2025
This preprint is open for discussion and under review for Geoscientific Model Development (GMD).
Short summary
Short summary
We introduce the first denoising diffusion model for wildfire spread prediction, a new kind of generative AI model that learns to simulate fires not just as one fixed outcome, but as a range of possible scenarios. This allows us to capture the inherent uncertainty of wildfire dynamics. Our model produces ensembles of forecasts that reflect physically meaningful distributions of where fire might go next.
Joffrey Dumont Le Brazidec, Pierre Vanderbecken, Alban Farchi, Grégoire Broquet, Gerrit Kuhlmann, and Marc Bocquet
Geosci. Model Dev., 18, 3607–3622, https://doi.org/10.5194/gmd-18-3607-2025, https://doi.org/10.5194/gmd-18-3607-2025, 2025
Short summary
Short summary
We developed a deep learning method to estimate CO2 emissions from power plants using satellite images. Trained and validated on simulated data, our model accurately predicts emissions despite challenges like cloud cover. When applied to real OCO3 satellite images, the results closely match reported emissions. This study shows that neural networks trained on simulations can effectively analyse real satellite data, offering a new way to monitor CO2 emissions from space.
Nicolas Guillaume Alexandre Mokus, Véronique Dansereau, Guillaume Boutin, Jean-Pierre Auclair, and Alexandre Tlili
EGUsphere, https://doi.org/10.5194/egusphere-2025-1831, https://doi.org/10.5194/egusphere-2025-1831, 2025
This preprint is open for discussion and under review for Geoscientific Model Development (GMD).
Short summary
Short summary
Arctic sea ice recedes and it is more exposed to waves. Waves can then fracture continuous pack ice into floes, which are more mobile and easier to melt. The fracture process itself is not well understood, because of harsh field conditions. We propose a novel sea ice fracture criterion incorporated into a numerical model that simulates wave propagation. This criterion can be compared to existing ones. We relate our results to laboratory experiments, and find qualitative agreement.
Ieuan Higgs, Ross Bannister, Jozef Skákala, Alberto Carrassi, and Stefano Ciavatta
EGUsphere, https://doi.org/10.48550/arXiv.2504.05218, https://doi.org/10.48550/arXiv.2504.05218, 2025
Short summary
Short summary
We explored how machine learning can improve computer models that simulate ocean ecosystems. These models help us understand how the ocean works, but they often struggle due to limited observations and complex processes. Our approach uses machine learning to better connect the parts of the system we can observe with those we cannot. This leads to more accurate and efficient predictions, offering a promising way to improve future ocean monitoring and forecasting tools.
Charlotte Durand, Tobias Sebastian Finn, Alban Farchi, Marc Bocquet, Julien Brajard, and Laurent Bertino
EGUsphere, https://doi.org/10.5194/egusphere-2024-4028, https://doi.org/10.5194/egusphere-2024-4028, 2025
Short summary
Short summary
This paper presents a four-dimensional variational data assimilation system based on a neural network emulator for sea-ice thickness, learned from neXtSIM simulation outputs. Testing with simulated and real observation retrievals, the system improves forecasts and bias error, performing comparably to operational methods, demonstrating the promise of sea-ice data-driven data assimilation systems.
Simon Driscoll, Alberto Carrassi, Julien Brajard, Laurent Bertino, Einar Ólason, Marc Bocquet, and Amos Lawless
EGUsphere, https://doi.org/10.5194/egusphere-2024-2476, https://doi.org/10.5194/egusphere-2024-2476, 2024
Short summary
Short summary
The formation and evolution of sea ice melt ponds (ponds of melted water) are complex, insufficiently understood and represented in models with considerable uncertainty. These uncertain representations are not traditionally included in climate models potentially causing the known underestimation of sea ice loss in climate models. Our work creates the first observationally based machine learning model of melt ponds that is also a ready and viable candidate to be included in climate models.
Tobias Sebastian Finn, Lucas Disson, Alban Farchi, Marc Bocquet, and Charlotte Durand
Nonlin. Processes Geophys., 31, 409–431, https://doi.org/10.5194/npg-31-409-2024, https://doi.org/10.5194/npg-31-409-2024, 2024
Short summary
Short summary
We train neural networks as denoising diffusion models for state generation in the Lorenz 1963 system and demonstrate that they learn an internal representation of the system. We make use of this learned representation and the pre-trained model in two downstream tasks: surrogate modelling and ensemble generation. For both tasks, the diffusion model can outperform other more common approaches. Thus, we see a potential of representation learning with diffusion models for dynamical systems.
Laurent Brodeau, Pierre Rampal, Einar Ólason, and Véronique Dansereau
Geosci. Model Dev., 17, 6051–6082, https://doi.org/10.5194/gmd-17-6051-2024, https://doi.org/10.5194/gmd-17-6051-2024, 2024
Short summary
Short summary
A new brittle sea ice rheology, BBM, has been implemented into the sea ice component of NEMO. We describe how a new spatial discretization framework was introduced to achieve this. A set of idealized and realistic ocean and sea ice simulations of the Arctic have been performed using BBM and the standard viscous–plastic rheology of NEMO. When compared to satellite data, our simulations show that our implementation of BBM leads to a fairly good representation of sea ice deformations.
Marc Bocquet, Pierre J. Vanderbecken, Alban Farchi, Joffrey Dumont Le Brazidec, and Yelva Roustan
Nonlin. Processes Geophys., 31, 335–357, https://doi.org/10.5194/npg-31-335-2024, https://doi.org/10.5194/npg-31-335-2024, 2024
Short summary
Short summary
A novel approach, optimal transport data assimilation (OTDA), is introduced to merge DA and OT concepts. By leveraging OT's displacement interpolation in space, it minimises mislocation errors within DA applied to physical fields, such as water vapour, hydrometeors, and chemical species. Its richness and flexibility are showcased through one- and two-dimensional illustrations.
Lilian Garcia-Oliva, Alberto Carrassi, and François Counillon
EGUsphere, https://doi.org/10.5194/egusphere-2024-1843, https://doi.org/10.5194/egusphere-2024-1843, 2024
Short summary
Short summary
We used a simple coupled model and a data assimilation method to find the correct initialisation for climate predictions. We aim to clarify when weakly or strongly coupled data assimilation (WCDA or SCDA) is best, depending on the system's dynamical characteristics (spatio-temporal) and data coverage.
We found that WCDA is better in full data coverage. When we have a partially observed system, SCDA is better. This result depends on the temporal and spatial scale of the observed quantity.
We found that WCDA is better in full data coverage. When we have a partially observed system, SCDA is better. This result depends on the temporal and spatial scale of the observed quantity.
Yumeng Chen, Lars Nerger, and Amos S. Lawless
EGUsphere, https://doi.org/10.5194/egusphere-2024-1078, https://doi.org/10.5194/egusphere-2024-1078, 2024
Short summary
Short summary
In this paper, we present pyPDAF, a Python interface to the parallel data assimilation framework (PDAF) allowing for coupling with Python-based models. We demonstrate the capability and efficiency of pyPDAF under a coupled data assimilation setup.
Yumeng Chen, Polly Smith, Alberto Carrassi, Ivo Pasmans, Laurent Bertino, Marc Bocquet, Tobias Sebastian Finn, Pierre Rampal, and Véronique Dansereau
The Cryosphere, 18, 2381–2406, https://doi.org/10.5194/tc-18-2381-2024, https://doi.org/10.5194/tc-18-2381-2024, 2024
Short summary
Short summary
We explore multivariate state and parameter estimation using a data assimilation approach through idealised simulations in a dynamics-only sea-ice model based on novel rheology. We identify various potential issues that can arise in complex operational sea-ice models when model parameters are estimated. Even though further investigation will be needed for such complex sea-ice models, we show possibilities of improving the observed and the unobserved model state forecast and parameter accuracy.
Charlotte Durand, Tobias Sebastian Finn, Alban Farchi, Marc Bocquet, Guillaume Boutin, and Einar Ólason
The Cryosphere, 18, 1791–1815, https://doi.org/10.5194/tc-18-1791-2024, https://doi.org/10.5194/tc-18-1791-2024, 2024
Short summary
Short summary
This paper focuses on predicting Arctic-wide sea-ice thickness using surrogate modeling with deep learning. The model has a predictive power of 12 h up to 6 months. For this forecast horizon, persistence and daily climatology are systematically outperformed, a result of learned thermodynamics and advection. Consequently, surrogate modeling with deep learning proves to be effective at capturing the complex behavior of sea ice.
Joffrey Dumont Le Brazidec, Pierre Vanderbecken, Alban Farchi, Grégoire Broquet, Gerrit Kuhlmann, and Marc Bocquet
Geosci. Model Dev., 17, 1995–2014, https://doi.org/10.5194/gmd-17-1995-2024, https://doi.org/10.5194/gmd-17-1995-2024, 2024
Short summary
Short summary
Our research presents an innovative approach to estimating power plant CO2 emissions from satellite images of the corresponding plumes such as those from the forthcoming CO2M satellite constellation. The exploitation of these images is challenging due to noise and meteorological uncertainties. To overcome these obstacles, we use a deep learning neural network trained on simulated CO2 images. Our method outperforms alternatives, providing a positive perspective for the analysis of CO2M images.
Ieuan Higgs, Jozef Skákala, Ross Bannister, Alberto Carrassi, and Stefano Ciavatta
Biogeosciences, 21, 731–746, https://doi.org/10.5194/bg-21-731-2024, https://doi.org/10.5194/bg-21-731-2024, 2024
Short summary
Short summary
A complex network is a way of representing which parts of a system are connected to other parts. We have constructed a complex network based on an ecosystem–ocean model. From this, we can identify patterns in the structure and areas of similar behaviour. This can help to understand how natural, or human-made, changes will affect the shelf sea ecosystem, and it can be used in multiple future applications such as improving modelling, data assimilation, or machine learning.
Bo Dong, Ross Bannister, Yumeng Chen, Alison Fowler, and Keith Haines
Geosci. Model Dev., 16, 4233–4247, https://doi.org/10.5194/gmd-16-4233-2023, https://doi.org/10.5194/gmd-16-4233-2023, 2023
Short summary
Short summary
Traditional Kalman smoothers are expensive to apply in large global ocean operational forecast and reanalysis systems. We develop a cost-efficient method to overcome the technical constraints and to improve the performance of existing reanalysis products.
Joffrey Dumont Le Brazidec, Pierre Vanderbecken, Alban Farchi, Marc Bocquet, Jinghui Lian, Grégoire Broquet, Gerrit Kuhlmann, Alexandre Danjou, and Thomas Lauvaux
Geosci. Model Dev., 16, 3997–4016, https://doi.org/10.5194/gmd-16-3997-2023, https://doi.org/10.5194/gmd-16-3997-2023, 2023
Short summary
Short summary
Monitoring of CO2 emissions is key to the development of reduction policies. Local emissions, from cities or power plants, may be estimated from CO2 plumes detected in satellite images. CO2 plumes generally have a weak signal and are partially concealed by highly variable background concentrations and instrument errors, which hampers their detection. To address this problem, we propose and apply deep learning methods to detect the contour of a plume in simulated CO2 satellite images.
Thomas Richter, Véronique Dansereau, Christian Lessig, and Piotr Minakowski
Geosci. Model Dev., 16, 3907–3926, https://doi.org/10.5194/gmd-16-3907-2023, https://doi.org/10.5194/gmd-16-3907-2023, 2023
Short summary
Short summary
Sea ice covers not only the pole regions but affects the weather and climate globally. For example, its white surface reflects more sunlight than land. The oceans around the poles are therefore kept cool, which affects the circulation in the oceans worldwide. Simulating the behavior and changes in sea ice on a computer is, however, very difficult. We propose a new computer simulation that better models how cracks in the ice change over time and show this by comparing to other simulations.
Sukun Cheng, Yumeng Chen, Ali Aydoğdu, Laurent Bertino, Alberto Carrassi, Pierre Rampal, and Christopher K. R. T. Jones
The Cryosphere, 17, 1735–1754, https://doi.org/10.5194/tc-17-1735-2023, https://doi.org/10.5194/tc-17-1735-2023, 2023
Short summary
Short summary
This work studies a novel application of combining a Lagrangian sea ice model, neXtSIM, and data assimilation. It uses a deterministic ensemble Kalman filter to incorporate satellite-observed ice concentration and thickness in simulations. The neXtSIM Lagrangian nature is handled using a remapping strategy on a common homogeneous mesh. The ensemble is formed by perturbing air–ocean boundary conditions and ice cohesion. Thanks to data assimilation, winter Arctic sea ice forecasting is enhanced.
Pierre J. Vanderbecken, Joffrey Dumont Le Brazidec, Alban Farchi, Marc Bocquet, Yelva Roustan, Élise Potier, and Grégoire Broquet
Atmos. Meas. Tech., 16, 1745–1766, https://doi.org/10.5194/amt-16-1745-2023, https://doi.org/10.5194/amt-16-1745-2023, 2023
Short summary
Short summary
Instruments dedicated to monitoring atmospheric gaseous compounds from space will provide images of urban-scale plumes. We discuss here the use of new metrics to compare observed plumes with model predictions that will be less sensitive to meteorology uncertainties. We have evaluated our metrics on diverse plumes and shown that by eliminating some aspects of the discrepancies, they are indeed less sensitive to meteorological variations.
Joffrey Dumont Le Brazidec, Marc Bocquet, Olivier Saunier, and Yelva Roustan
Geosci. Model Dev., 16, 1039–1052, https://doi.org/10.5194/gmd-16-1039-2023, https://doi.org/10.5194/gmd-16-1039-2023, 2023
Short summary
Short summary
When radionuclides are released into the atmosphere, the assessment of the consequences depends on the evaluation of the magnitude and temporal evolution of the release, which can be highly variable as in the case of Fukushima Daiichi.
Here, we propose Bayesian inverse modelling methods and the reversible-jump Markov chain Monte Carlo technique, which allows one to evaluate the temporal variability of the release and to integrate different types of information in the source reconstruction.
Colin Grudzien and Marc Bocquet
Geosci. Model Dev., 15, 7641–7681, https://doi.org/10.5194/gmd-15-7641-2022, https://doi.org/10.5194/gmd-15-7641-2022, 2022
Short summary
Short summary
Iterative optimization techniques, the state of the art in data assimilation, have largely focused on extending forecast accuracy to moderate- to long-range forecast systems. However, current methodology may not be cost-effective in reducing forecast errors in online, short-range forecast systems. We propose a novel optimization of these techniques for online, short-range forecast cycles, simultaneously providing an improvement in forecast accuracy and a reduction in the computational cost.
Francine Schevenhoven and Alberto Carrassi
Geosci. Model Dev., 15, 3831–3844, https://doi.org/10.5194/gmd-15-3831-2022, https://doi.org/10.5194/gmd-15-3831-2022, 2022
Short summary
Short summary
In this study, we present a novel formulation to build a dynamical combination of models, the so-called supermodel, which needs to be trained based on data. Previously, we assumed complete and noise-free observations. Here, we move towards a realistic scenario and develop adaptations to the training methods in order to cope with sparse and noisy observations. The results are very promising and shed light on how to apply the method with state of the art general circulation models.
Yumeng Chen, Alberto Carrassi, and Valerio Lucarini
Nonlin. Processes Geophys., 28, 633–649, https://doi.org/10.5194/npg-28-633-2021, https://doi.org/10.5194/npg-28-633-2021, 2021
Short summary
Short summary
Chaotic dynamical systems are sensitive to the initial conditions, which are crucial for climate forecast. These properties are often used to inform the design of data assimilation (DA), a method used to estimate the exact initial conditions. However, obtaining the instability properties is burdensome for complex problems, both numerically and analytically. Here, we suggest a different viewpoint. We show that the skill of DA can be used to infer the instability properties of a dynamical system.
Joffrey Dumont Le Brazidec, Marc Bocquet, Olivier Saunier, and Yelva Roustan
Atmos. Chem. Phys., 21, 13247–13267, https://doi.org/10.5194/acp-21-13247-2021, https://doi.org/10.5194/acp-21-13247-2021, 2021
Short summary
Short summary
The assessment of the environmental consequences of a radionuclide release depends on the estimation of its source. This paper aims to develop inverse Bayesian methods which combine transport models with measurements, in order to reconstruct the ensemble of possible sources.
Three methods to quantify uncertainties based on the definition of probability distributions and the physical models are proposed and evaluated for the case of 106Ru releases over Europe in 2017.
Yumeng Chen, Konrad Simon, and Jörn Behrens
Geosci. Model Dev., 14, 2289–2316, https://doi.org/10.5194/gmd-14-2289-2021, https://doi.org/10.5194/gmd-14-2289-2021, 2021
Short summary
Short summary
Mesh adaptivity can reduce overall model error by only refining meshes in specific areas where it us necessary in the runtime. Here we suggest a way to integrate mesh adaptivity into an existing Earth system model, ECHAM6, without having to redesign the implementation from scratch. We show that while the additional computational effort is manageable, the error can be reduced compared to a low-resolution standard model using an idealized test and relatively realistic dust transport tests.
Einar Ólason, Pierre Rampal, and Véronique Dansereau
The Cryosphere, 15, 1053–1064, https://doi.org/10.5194/tc-15-1053-2021, https://doi.org/10.5194/tc-15-1053-2021, 2021
Short summary
Short summary
We analyse the fractal properties observed in the pattern of the long, narrow openings that form in Arctic sea ice known as leads. We use statistical tools to explore the fractal properties of the lead fraction observed in satellite data and show that our sea-ice model neXtSIM displays the same behaviour. Building on this result we then show that the pattern of heat loss from ocean to atmosphere in the model displays similar fractal properties, stemming from the fractal properties of the leads.
Tobias Sebastian Finn, Gernot Geppert, and Felix Ament
Hydrol. Earth Syst. Sci. Discuss., https://doi.org/10.5194/hess-2020-672, https://doi.org/10.5194/hess-2020-672, 2021
Revised manuscript not accepted
Short summary
Short summary
Through the lens of recent developments in hydrological modelling and data assimilation, we hourly update the soil moisture with ensemble data assimilation and sparse 2-metre-temperature observations in a coupled limited area model system. In idealized experiments, we improve the soil moisture analysis by coupled data assimilation across the atmosphere-land interface. We conclude that we can merge the separated assimilation cycles for the atmosphere and land surface into one single cycle.
Cited articles
Amitrano, D., Grasso, J.-R., and Hantz, D.: From Diffuse to Localised Damage
through Elastic Interaction, Geophys. Res. Lett., 26, 2109–2112,
https://doi.org/10.1029/1999GL900388, 1999. a
Andersson, T. R., Hosking, J. S., Pérez-Ortiz, M., Paige, B., Elliott,
A., Russell, C., Law, S., Jones, D. C., Wilkinson, J., Phillips, T., Byrne,
J., Tietsche, S., Sarojini, B. B., Blanchard-Wrigglesworth, E., Aksenov,
Y., Downie, R., and Shuckburgh, E.: Seasonal Arctic Sea Ice Forecasting
with Probabilistic Deep Learning, Nat. Commun., 12, 5124,
https://doi.org/10.1038/s41467-021-25257-4, 2021. a
Ayzel, G., Scheffer, T., and Heistermann, M.: RainNet v1.0: a convolutional neural network for radar-based precipitation nowcasting, Geosci. Model Dev., 13, 2631–2644, https://doi.org/10.5194/gmd-13-2631-2020, 2020. a
Ba, J. L., Kiros, J. R., and Hinton, G. E.: Layer Normalization,
arXiv [preprint], https://doi.org/10.48550/arXiv.1607.06450, 2016. a
Bachlechner, T., Majumder, B. P., Mao, H. H., Cottrell, G. W., and McAuley, J.:
ReZero Is All You Need: Fast Convergence at Large Depth,
arXiv [preprint], https://doi.org/10.48550/arXiv.2003.04887, 2020. a
Beck, A. and Kurz, M.: A Perspective on Machine Learning Methods in Turbulence
Modeling, GAMM-Mitteilungen, 44, e202100002, https://doi.org/10.1002/gamm.202100002,
2021. a
Beucler, T., Ebert-Uphoff, I., Rasp, S., Pritchard, M., and Gentine, P.:
Machine Learning for Clouds and Climate (Invited Chapter for
the AGU Geophysical Monograph Series “Clouds and Climate”) [preprint],
https://doi.org/10.1002/essoar.10506925.1, 2021. a
Bloom, S. C., Takacs, L. L., da Silva, A. M., and Ledvina, D.: Data
Assimilation Using Incremental Analysis Updates, Mon. Weather Rev.,
124, 1256–1271, https://doi.org/10.1175/1520-0493(1996)124<1256:DAUIAU>2.0.CO;2, 1996. a
Bocquet, M., Brajard, J., Carrassi, A., and Bertino, L.: Data assimilation as a learning tool to infer ordinary differential equation representations of dynamical models, Nonlin. Processes Geophys., 26, 143–162, https://doi.org/10.5194/npg-26-143-2019, 2019. a, b
Bocquet, M., Brajard, J., Carrassi, A., and Bertino, L.: Bayesian Inference of
Chaotic Dynamics by Merging Data Assimilation, Machine Learning and
Expectation-Maximization, Foundations of Data Science, 2, 55,
https://doi.org/10.3934/fods.2020004, 2020. a
Bouchat, A., Hutter, N., Chanut, J., Dupont, F., Dukhovskoy, D., Garric, G.,
Lee, Y. J., Lemieux, J.-F., Lique, C., Losch, M., Maslowski, W., Myers,
P. G., Ólason, E., Rampal, P., Rasmussen, T., Talandier, C., Tremblay,
B., and Wang, Q.: Sea Ice Rheology Experiment (SIREx): 1. Scaling
and Statistical Properties of Sea-Ice Deformation Fields, J.
Geophys. Res.-Oceans, 127, e2021JC017667,
https://doi.org/10.1029/2021JC017667, 2022. a, b
Boutin, G., Williams, T., Horvat, C., and Brodeau, L.: Modelling the Arctic
Wave-Affected Marginal Ice Zone: A Comparison with ICESat-2 Observations,
Philos. T. R. Soc. A, 380, 20210262, https://doi.org/10.1098/rsta.2021.0262, 2022. a
Brajard, J., Carrassi, A., Bocquet, M., and Bertino, L.: Combining Data
Assimilation and Machine Learning to Emulate a Dynamical Model from Sparse
and Noisy Observations: A Case Study with the Lorenz 96 Model,
J. Comput. Sci., 44, 101171,
https://doi.org/10.1016/j.jocs.2020.101171, 2020. a
Brajard, J., Carrassi, A., Bocquet, M., and Bertino, L.: Combining Data
Assimilation and Machine Learning to Infer Unresolved Scale Parametrization,
Philos. T. R. Soc. A, 379, 20200 086, https://doi.org/10.1098/rsta.2020.0086, 2021. a
Breiman, L.: Statistical Modeling: The Two Cultures (with Comments and
a Rejoinder by the Author), Stat. Sci., 16, 199–231,
https://doi.org/10.1214/ss/1009213726, 2001. a
Brenowitz, N. D. and Bretherton, C. S.: Prognostic Validation of a Neural
Network Unified Physics Parameterization, Geophys. Res. Lett., 45,
6289–6298, https://doi.org/10.1029/2018GL078510, 2018. a
Carrassi, A., Weber, R. J. T., Guemas, V., Doblas-Reyes, F. J., Asif, M., and Volpi, D.: Full-field and anomaly initialization using a low-order climate model: a comparison and proposals for advanced formulations, Nonlin. Processes Geophys., 21, 521–537, https://doi.org/10.5194/npg-21-521-2014, 2014. a
Cheng, Y., Giometto, M. G., Kauffmann, P., Lin, L., Cao, C., Zupnick, C., Li,
H., Li, Q., Huang, Y., Abernathey, R., and Gentine, P.: Deep Learning for
Subgrid-Scale Turbulence Modeling in Large-Eddy Simulations of the
Convective Atmospheric Boundary Layer, J. Adv. Model.
Earth Sy., 14, e2021MS002 847, https://doi.org/10.1029/2021MS002847, 2022. a
Cipolla, R., Gal, Y., and Kendall, A.: Multi-Task Learning Using
Uncertainty to Weigh Losses for Scene Geometry and Semantics,
in: 2018 IEEE/CVF Conference on Computer Vision and Pattern
Recognition, 7482–7491, IEEE, Salt Lake City, UT, USA,
https://doi.org/10.1109/CVPR.2018.00781, 2018. a, b
Dansereau, V., Weiss, J., and Saramito, P.: A Continuum
Viscous-Elastic-Brittle, Finite Element DG Model for the Fracture
and Drift of Sea Ice, in: Challenges and Innovations in
Geomechanics, edited by: Barla, M., Di Donna, A., and Sterpi, D., Lecture
Notes in Civil Engineering, 125–139, Springer International
Publishing, Cham, https://doi.org/10.1007/978-3-030-64514-4_8, 2021. a
De, S. and Smith, S. L.: Batch Normalization Biases Residual Blocks Towards
the Identity Function in Deep Networks, arXiv [preprint],
https://doi.org/10.48550/arXiv.2002.10444, 2020. a
Falcon, W. and The PyTorch Lightning team: PyTorch Lightning, Zenodo [code], https://doi.org/10.5281/zenodo.3530844, 2019. a
Farchi, A., Bocquet, M., Laloyaux, P., Bonavita, M., and Malartic, Q.: A
Comparison of Combined Data Assimilation and Machine Learning Methods for
Offline and Online Model Error Correction, J. Comput. Sci.,
55, 101468, https://doi.org/10.1016/j.jocs.2021.101468, 2021a. a, b, c, d
Farchi, A., Laloyaux, P., Bonavita, M., and Bocquet, M.: Using Machine Learning
to Correct Model Error in Data Assimilation and Forecast Applications,
Q. J. Roy. Meteorol. Soc., 147, 3067–3084,
https://doi.org/10.1002/qj.4116, 2021b. a
Finn, T.: cerea-daml/hybrid_nn_meb_model: Preprint submission (Preprint), Zenodo [code], https://doi.org/10.5281/zenodo.7997435, 2023. a
Fisher, R. A.: Frequency Distribution of the Values of the
Correlation Coefficient in Samples from an Indefinitely Large
Population, Biometrika, 10, 507–521, https://doi.org/10.2307/2331838, 1915. a
Geer, A. J.: Learning Earth System Models from Observations: Machine Learning
or Data Assimilation?, Philos. T. R. Soc. A, 379, 20200089,
https://doi.org/10.1098/rsta.2020.0089, 2021. a
Gentine, P., Pritchard, M., Rasp, S., Reinaudi, G., and Yacalis, G.: Could
Machine Learning Break the Convection Parameterization Deadlock?,
Geophys. Res. Lett., 45, 5742–5751, https://doi.org/10.1029/2018GL078202,
2018. a
Girard, L., Bouillon, S., Weiss, J., Amitrano, D., Fichefet, T., and Legat, V.:
A New Modeling Framework for Sea-Ice Mechanics Based on Elasto-Brittle
Rheology, Ann. Glaciol., 52, 123–132,
https://doi.org/10.3189/172756411795931499, 2011. a
Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D.,
Ozair, S., Courville, A., and Bengio, Y.: Generative Adversarial Nets,
in: Advances in Neural Information Processing Systems 27, edited by:
Ghahramani, Z., Welling, M., Cortes, C., Lawrence, N. D., and Weinberger,
K. Q., 2672–2680, Curran Associates, Inc., ISBN 9781510800410, 2014. a
Guillaumin, A. P. and Zanna, L.: Stochastic-Deep Learning Parameterization
of Ocean Momentum Forcing, J. Adv. Model. Earth Sy.,
13, e2021MS002534, https://doi.org/10.1029/2021MS002534, 2021. a
Hendrycks, D. and Gimpel, K.: Gaussian Error Linear Units (GELUs),
arXiv [preprint], https://doi.org/10.48550/arXiv.1606.08415, 2020. a, b
Hodson, T. O.: Root-mean-square error (RMSE) or mean absolute error (MAE): when to use them or not, Geosci. Model Dev., 15, 5481–5487, https://doi.org/10.5194/gmd-15-5481-2022, 2022. a
Horvat, C. and Roach, L. A.: WIFF1.0: a hybrid machine-learning-based parameterization of wave-induced sea ice floe fracture, Geosci. Model Dev., 15, 803–814, https://doi.org/10.5194/gmd-15-803-2022, 2022. a
Houghton, J. T., Ding, Y., Griggs, D. J., Noguer, M., van der
Linden, P. J., Dai, X., Maskell, K., and Johnson, C. A.: Climate
Change 2001: The Scientific Basis: Contribution of Working Group I to the
Third Assessment Report of the Intergovernmental Panel on Climate
Change, Cambridge University Press, ISBN 0-521-01495-6, 2001 a
Irrgang, C., Boers, N., Sonnewald, M., Barnes, E. A., Kadow, C., Staneva, J.,
and Saynisch-Wagner, J.: Towards Neural Earth System Modelling by
Integrating Artificial Intelligence in Earth System Science, Nature
Machine Intelligence, 3, 667–674, https://doi.org/10.1038/s42256-021-00374-3, 2021. a
Kaplan, J., McCandlish, S., Henighan, T., Brown, T. B., Chess, B., Child, R.,
Gray, S., Radford, A., Wu, J., and Amodei, D.: Scaling Laws for Neural
Language Models, arXiv [preprint], https://doi.org/10.48550/arXiv.2001.08361, 2020. a
Kingma, D. P. and Ba, J.: Adam: A Method for Stochastic Optimization,
arXiv [preprint], https://doi.org/10.48550/arXiv.1412.6980, 2017. a
Korosov, A., Rampal, P., Ying, Y., Ólason, E., and Williams, T.: Towards improving short-term sea ice predictability using deformation observations, The Cryosphere Discuss. [preprint], https://doi.org/10.5194/tc-2022-46, in review, 2022. a
Lee, S., Seo, Y., Lee, K., Abbeel, P., and Shin, J.: Addressing Distribution
Shift in Online Reinforcement Learning with Offline Datasets, Offline Reinforcement Learning Workshop at Neural Information Processing Systems, https://offline-rl-neurips.github.io/pdf/13.pdf (last access: 17 July 2023), 2021. a
Levine, S., Kumar, A., Tucker, G., and Fu, J.: Offline Reinforcement
Learning: Tutorial, Review, and Perspectives on Open
Problems, arXiv [preprint], https://doi.org/10.48550/arXiv.2005.01643, 2020. a
Liu, Q., Zhang, R., Wang, Y., Yan, H., and Hong, M.: Daily Prediction of
the Arctic Sea Ice Concentration Using Reanalysis Data Based on a
Convolutional LSTM Network, J. Mar. Sci. Eng., 9,
330, https://doi.org/10.3390/jmse9030330, 2021. a
Liu, Z., Mao, H., Wu, C.-Y., Feichtenhofer, C., Darrell, T., and Xie, S.: A
ConvNet for the 2020s, 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, LA, USA, 18–24 June 2022, 11966–11976,
https://doi.org/10.1109/CVPR52688.2022.01167, 2022. a, b, c, d, e, f, g, h, i, j
Murphy, K. P.: Machine Learning: A Probabilistic Perspective, Adaptive
Computation and Machine Learning Series, MIT Press, Cambridge, MA, ISBN 978-0-262-01802-9,
2012. a
Norton, R. M.: The Double Exponential Distribution: Using Calculus to
Find a Maximum Likelihood Estimator, Am. Stat., 38,
135–136, https://doi.org/10.2307/2683252, 1984. a
Odena, A., Dumoulin, V., and Olah, C.: Deconvolution and Checkerboard
Artifacts, Distill, 1, e3, https://doi.org/10.23915/distill.00003, 2016. a, b
Ólason, E., Rampal, P., and Dansereau, V.: On the statistical properties of sea-ice lead fraction and heat fluxes in the Arctic, The Cryosphere, 15, 1053–1064, https://doi.org/10.5194/tc-15-1053-2021, 2021. a
Ólason, E., Boutin, G., Korosov, A., Rampal, P., Williams, T., Kimmritz,
M., Dansereau, V., and Samaké, A.: A New Brittle Rheology and
Numerical Framework for Large-Scale Sea-Ice Models, J.
Adv. Model. Earth Sy., 14, e2021MS002685,
https://doi.org/10.1029/2021MS002685, 2022. a, b, c, d
Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen,
T., Lin, Z., Gimelshein, N., Antiga, L., Desmaison, A., Kopf, A., Yang, E.,
DeVito, Z., Raison, M., Tejani, A., Chilamkurthy, S., Steiner, B., Fang, L.,
Bai, J., and Chintala, S.: PyTorch: An Imperative Style,
High-Performance Deep Learning Library, in: Advances in Neural
Information Processing Systems 32, edited by: Wallach, H., Larochelle, H.,
Beygelzimer, A., Alché-Buc, F., Fox, E., and Garnett, R.,
Curran Associates, Inc., 8024–8035, ISBN 9781713807933, 2019. a
Prudencio, R. F., Maximo, M. R. O. A., and Colombini, E. L.: A Survey on
Offline Reinforcement Learning: Taxonomy, Review, and Open
Problems, arXiv [preprint], https://doi.org/10.48550/arXiv.2203.01387, 2022. a
Rabatel, M., Rampal, P., Carrassi, A., Bertino, L., and Jones, C. K. R. T.: Impact of rheology on probabilistic forecasts of sea ice trajectories: application for search and rescue operations in the Arctic, The Cryosphere, 12, 935–953, https://doi.org/10.5194/tc-12-935-2018, 2018. a
Rampal, P., Bouillon, S., Ólason, E., and Morlighem, M.: neXtSIM: a new Lagrangian sea ice model, The Cryosphere, 10, 1055–1073, https://doi.org/10.5194/tc-10-1055-2016, 2016. a, b, c
Rampal, P., Dansereau, V., Olason, E., Bouillon, S., Williams, T., Korosov, A., and Samaké, A.: On the multi-fractal scaling properties of sea ice deformation, The Cryosphere, 13, 2457–2474, https://doi.org/10.5194/tc-13-2457-2019, 2019. a
Rasp, S.: Coupled online learning as a way to tackle instabilities and biases in neural network parameterizations: general algorithms and Lorenz 96 case study (v1.0), Geosci. Model Dev., 13, 2185–2196, https://doi.org/10.5194/gmd-13-2185-2020, 2020. a
Rasp, S., Pritchard, M. S., and Gentine, P.: Deep Learning to Represent Subgrid
Processes in Climate Models, P. Natl. Acad. Sci. USA,
115, 9684–9689, https://doi.org/10.1073/pnas.1810286115, 2018. a
Ravuri, S., Lenc, K., Willson, M., Kangin, D., Lam, R., Mirowski, P.,
Fitzsimons, M., Athanassiadou, M., Kashem, S., Madge, S., Prudden, R.,
Mandhane, A., Clark, A., Brock, A., Simonyan, K., Hadsell, R., Robinson, N.,
Clancy, E., Arribas, A., and Mohamed, S.: Skilful Precipitation Nowcasting
Using Deep Generative Models of Radar, Nature, 597, 672–677,
https://doi.org/10.1038/s41586-021-03854-z, 2021. a
Ronneberger, O., Fischer, P., and Brox, T.: U-Net: Convolutional
Networks for Biomedical Image Segmentation, arXiv [preprint], https://doi.org/10.48550/arXiv.1505.04597,
2015. a, b, c, d
Rybkin, O., Daniilidis, K., and Levine, S.: Simple and Effective VAE
Training with Calibrated Decoders, arXiv [preprint], https://doi.org/10.48550/arXiv.2006.13202,
2020. a
Saramito, P.: Efficient C++ Finite Element Computing with Rheolef, CNRS-CCSD ed., https://cel.hal.science/cel-00573970 (last access: 17 July 2023), 2020. a
Schweiger, A. J. and Zhang, J.: Accuracy of Short-Term Sea Ice Drift Forecasts
Using a Coupled Ice-Ocean Model, J. Geophys. Res.-Oceans,
120, 7827–7841, https://doi.org/10.1002/2015JC011273, 2015. a
Seifert, A. and Rasp, S.: Potential and Limitations of Machine Learning
for Modeling Warm-Rain Cloud Microphysical Processes, J. Adv.
Model. Earth Sy., 12, e2020MS002301, https://doi.org/10.1029/2020MS002301,
2020. a
Simonyan, K., Vedaldi, A., and Zisserman, A.: Deep Inside Convolutional
Networks: Visualising Image Classification Models and Saliency
Maps, arXiv [preprint], https://doi.org/10.48550/arXiv.1312.6034, 2013. a
Smilkov, D., Thorat, N., Kim, B., Viégas, F., and Wattenberg, M.:
SmoothGrad: Removing Noise by Adding Noise, arXiv [preprint],
https://doi.org/10.48550/arXiv.1706.03825, 2017.
a
Sohl-Dickstein, J., Weiss, E. A., Maheswaranathan, N., and Ganguli, S.: Deep
Unsupervised Learning Using Nonequilibrium Thermodynamics, arXiv [preprint],
https://doi.org/10.48550/arXiv.1503.03585, 2015. a
Springenberg, J. T., Dosovitskiy, A., Brox, T., and Riedmiller, M.: Striving
for Simplicity: The All Convolutional Net, arXiv [preprint],
https://doi.org/10.48550/arXiv.1412.6806, 2015. a, b
Stockdale, T. N.: Coupled Ocean–Atmosphere Forecasts in the
Presence of Climate Drift, Mon. Weather Rev., 125, 809–818,
https://doi.org/10.1175/1520-0493(1997)125<0809:COAFIT>2.0.CO;2, 1997. a
Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D.,
Vanhoucke, V., and Rabinovich, A.: Going Deeper with Convolutions,
arXiv [preprint],
https://doi.org/10.48550/arXiv.1409.4842, 2014. a, b
Tang, C.: Numerical Simulation of Progressive Rock Failure and Associated
Seismicity, Int. J. Rock Mech. Min. Sci., 34,
249–261, https://doi.org/10.1016/S0148-9062(96)00039-3, 1997. a
Tomczak, J. M.: Deep Generative Modeling, Springer International
Publishing, Cham, https://doi.org/10.1007/978-3-030-93158-2, 2022. a
Wilchinsky, A. V. and Feltham, D. L.: Modelling the Rheology of Sea Ice as a
Collection of Diamond-Shaped Floes, J. Non-Newtonian Fluid,
138, 22–32, https://doi.org/10.1016/j.jnnfm.2006.05.001, 2006. a
Wilchinsky, A. V. and Feltham, D. L.: Modeling Coulombic Failure of Sea Ice
with Leads, J. Geophys. Res.-Oceans, 116, C08040,
https://doi.org/10.1029/2011JC007071, 2011. a
Yadan, O.: Hydra – A Framework for Elegantly Configuring Complex
Applications, GitHub [code], https://github.com/facebookresearch/hydra (last access: 19 July 2023), 2019. a
Zanna, L. and Bolton, T.: Data-Driven Equation Discovery of Ocean
Mesoscale Closures, Geophys. Res. Lett., 47, e2020GL088376,
https://doi.org/10.1029/2020GL088376, 2020. a
Short summary
We combine deep learning with a regional sea-ice model to correct model errors in the sea-ice dynamics of low-resolution forecasts towards high-resolution simulations. The combined model improves the forecast by up to 75 % and thereby surpasses the performance of persistence. As the error connection can additionally be used to analyse the shortcomings of the forecasts, this study highlights the potential of combined modelling for short-term sea-ice forecasting.
We combine deep learning with a regional sea-ice model to correct model errors in the sea-ice...
