mlsanityData to 14 Oct 2026

A posteriori learning for quasi-geostrophic turbulence parametrization

arXiv 2204.03911 ↗First posted 8 Apr 2022physics.flu-dyn cs.LG physics.ao-phcited by 36

Not on the map. Why?

Abstract (arXiv)

The use of machine learning to build subgrid parametrizations for climate models is receiving growing attention. State-of-the-art strategies address the problem as a supervised learning task and optimize algorithms that predict subgrid fluxes based on information from coarse resolution models. In practice, training data are generated from higher resolution numerical simulations transformed in order to mimic coarse resolution simulations. By essence, these strategies optimize subgrid parametrizations to meet so-called textita priori criteria. But the actual purpose of a subgrid parametrization is to obtain good performance in terms of textita posteriori metrics which imply computing entire model trajectories. In this paper, we focus on the representation of energy backscatter in two dimensional quasi-geostrophic turbulence and compare parametrizations obtained with different learning strategies at fixed computational complexity. We show that strategies based on textita priori criteria yield parametrizations that tend to be unstable in direct simulations and describe how subgrid parametrizations can alternatively be trained end-to-end in order to meet textita posteriori criteria. We illustrate that end-to-end learning strategies yield parametrizations that outperform known empirical and data-driven schemes in terms of performance, stability and ability to apply to different flow configurations. These results support the relevance of differentiable programming paradigms for climate models in the future.

Key facts. Cited by 36 papers in the map (data through 14 Oct 2026). Builds on 26 of its 118 references that are in the map. First posted on arXiv 8 Apr 2022.

Builds on

26 of 118 references.

References · 26

  1. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  2. 2018Neural Ordinary Differential Equationscited by 3,779
  3. 2016Group Equivariant Convolutional Networkscited by 1,165
  4. 2017DGM: A deep learning algorithm for solving partial differential equationscited by 678
  5. 2021Machine learning accelerated computational fluid dynamicscited by 471
  6. 2020Lagrangian Neural Networkscited by 398
  7. 2019Machine Learning for Fluid Mechanicscited by 385
  8. 2017PDE-Net: Learning PDEs from Datacited by 311
  9. 2020Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solverscited by 205
  10. 2018Deep learning to represent sub-grid processes in climate modelscited by 156
  11. 2020Learning to Control PDEs with Differentiable Physicscited by 132
  12. 2021Learned Coarse Models for Efficient Turbulence Simulationcited by 91
14 more
  1. 2021Sub-seasonal forecasting with a large ensemble of deep-learning weather prediction modelscited by 79
  2. 2021Physics-based Deep Learningcited by 68
  3. 2017Earth System Modeling 2.0: A Blueprint for Models That Learn From Observations and Targeted High-Resolution Simulationscited by 51
  4. 2019DPM: A deep learning PDE augmentation method (with application to large-eddy simulation)cited by 41
  5. 2020Using machine learning to correct model error in data assimilation and forecast applicationscited by 41
  6. 2020Building high accuracy emulators for scientific simulations with deep neural architecture searchcited by 36
  7. 2020Learning Variational Data Assimilation Models and Solverscited by 36
  8. 2017Limits of End-to-End Learningcited by 35
  9. 2021Noether Networks: Meta-Learning Useful Conserved Quantitiescited by 26
  10. 2020Data-Driven Discovery of Coarse-Grained Equationscited by 14
  11. 2020Physical invariance in neural networks for subgrid-scale scalar flux modelingcited by 13
  12. 2021Embedded training of neural-network sub-grid-scale turbulence modelscited by 13
  13. 2021Machine Learning Emulation of 3D Cloud Radiative Effectscited by 5
  14. 2021Learning Runge-Kutta Integration Schemes for ODE Simulation and Identificationcited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 3; 2023: 14; 2024: 7; 2025: 6; 2026: 6.

36 papers.

Sort:

By area:

  1. 2023Machine learning with data assimilation and uncertainty quantification for dynamical systems: a reviewcited by 75
  2. 2022Learning to correct spectral methods for simulating turbulent flowscited by 49
  3. 2024DySLIM: Dynamics Stable Learning by Invariant Measure for Chaotic Systemscited by 30
  4. 2022Differentiable Programming for Earth System Modelingcited by 22
  5. 2023A Neural PDE Solver with Temporal Stencil Modelingcited by 22
  6. 2022Explaining the physics of transfer learning a data-driven subgrid-scale closure to a different turbulent flowcited by 20
  7. 2023Evolve Smoothly, Fit Consistently: Learning Smooth Latent Dynamics For Advection-Dominated Systemscited by 14
  8. 2023Learning Closed-form Equations for Subgrid-scale Closures from High-fidelity Data: Promises and Challengescited by 9
  9. 2023Gradient-free online learning of subgrid-scale dynamics with neural emulatorscited by 5
  10. 2024Joint Parameter and Parameterization Inference with Uncertainty Quantification through Differentiable Programmingcited by 5
  11. 2024What You See is Not What You Get: Neural Partial Differential Equations and The Illusion of Learningcited by 5
  12. 2023AttNS: Attention-Inspired Numerical Solving For Limited Data Scenarioscited by 4
24 more
  1. 2023Multi-fidelity climate model parameterization for better generalization and extrapolationcited by 4
  2. 2023OceanBench: The Sea Surface Height Editioncited by 4
  3. 2023Online Calibration of Deep Learning Sub-Models for Hybrid Numerical Modeling Systemscited by 4
  4. 2024Decoding complexity: how machine learning is redefining scientific discoverycited by 3
  5. 2025INC: An Indirect Neural Corrector for Auto-Regressive Hybrid PDE Solverscited by 3
  6. 2023Data-driven multiscale modeling for correcting dynamical systemscited by 2
  7. 2023On the choice of training data for machine learning of geostrophic mesoscale turbulencecited by 2
  8. 2023Extreme Event Prediction with Multi-agent Reinforcement Learning-based Parametrization of Atmospheric and Oceanic Turbulencecited by 2
  9. 2024Development of an offline and online hybrid model for the Integrated Forecasting Systemcited by 2
  10. 2025Guided Unconditional and Conditional Generative Models for Super-Resolution and Inference of Quasi-Geostrophic Turbulencecited by 2
  11. 2025Electron neural closure for turbulent magnetosheath simulations: energy channelscited by 2
  12. 2025Towards fully differentiable neural ocean model with Veroscited by 2
  13. 2025Calibrating Geophysical Predictions under Constrained Probabilistic Distributionscited by 2
  14. 2023Algorithmically Designed Artificial Neural Networks (ADANNs): Higher order deep operator learning for parametric partial differential equationscited by 1
  15. 2023Differentiable Turbulence: Closure as a partial differential equation constrained optimizationcited by 1
  16. 2024Adjoint-based online learning of two-layer quasi-geostrophic baroclinic turbulencecited by 1
  17. 2025NORi: An ML-Augmented Ocean Boundary Layer Parameterizationcited by 1
  18. 2026Towards Efficient and Stable Ocean State Forecasting: A Continuous-Time Koopman Approachcited by 1
  19. 2026Samudra 2: Scaling Ocean Emulators across Resolutionscited by 1
  20. 2024Rational-WENO: A lightweight, physically-consistent three-point weighted essentially non-oscillatory schemecited by 0
  21. 2026Calibration of a neural network ocean closure for improved mean state and variabilitycited by 0
  22. 2026Deep Learning-based Algebraic Reynolds Stress Closures for RANS Simulations of Turbulent Flowscited by 0
  23. 2026Spectrally Regularized Latent Flow Matching for Turbulence Generationcited by 0
  24. 2026Improving Reduced-Order Rotating Detonation Engine Models with Data Assimilation and Machine Learningcited by 0