Search Menu
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information
Search Scholarly Publications

Title: Data Imbalance, Uncertainty Quantification, and Transfer Learning in Data‐Driven Parameterizations: Lessons From the Emulation of Gravity Wave Momentum Transport in WACCM

Journal Article · · Journal of Advances in Modeling Earth Systems
ORCiD logo [1]; ORCiD logo [2];  [3]; ORCiD logo [1]; ORCiD logo [4]; ORCiD logo [5];  [6]; ORCiD logo [7]; ORCiD logo [1]
  1. Rice University Houston TX USA, University of Chicago Chicago IL USA
  2. Rice University Houston TX USA, NorthWest Research Associates Boulder CO USA
  3. Rice University Houston TX USA, University of California, Santa Cruz Santa Cruz CA USA
  4. Rice University Houston TX USA, Pacific Northwest National Laboratory Richland WA USA
  5. NorthWest Research Associates Boulder CO USA
  6. New York University New York NY USA
  7. Stanford University Palo Alto CA USA
+ Show Author Affiliations

Abstract Neural networks (NNs) are increasingly used for data‐driven subgrid‐scale parameterizations in weather and climate models. While NNs are powerful tools for learning complex non‐linear relationships from data, there are several challenges in using them for parameterizations. Three of these challenges are (a) data imbalance related to learning rare, often large‐amplitude, samples; (b) uncertainty quantification (UQ) of the predictions to provide an accuracy indicator; and (c) generalization to other climates, for example, those with different radiative forcings. Here, we examine the performance of methods for addressing these challenges using NN‐based emulators of the Whole Atmosphere Community Climate Model (WACCM) physics‐based gravity wave (GW) parameterizations as a test case. WACCM has complex, state‐of‐the‐art parameterizations for orography‐, convection‐, and front‐driven GWs. Convection‐ and orography‐driven GWs have significant data imbalance due to the absence of convection or orography in most grid points. We address data imbalance using resampling and/or weighted loss functions, enabling the successful emulation of parameterizations for all three sources. We demonstrate that three UQ methods (Bayesian NNs, variational auto‐encoders, and dropouts) provide ensemble spreads that correspond to accuracy during testing, offering criteria for identifying when an NN gives inaccurate predictions. Finally, we show that the accuracy of these NNs decreases for a warmer climate (4 × CO 2 ). However, their performance is significantly improved by applying transfer learning, for example, re‐training only one layer using ∼1% new data from the warmer climate. The findings of this study offer insights for developing reliable and generalizable data‐driven parameterizations for various processes, including (but not limited to) GWs.

Sponsoring Organization:
USDOE
OSTI ID:
2429750
Journal Information:
Journal of Advances in Modeling Earth Systems, Journal Name: Journal of Advances in Modeling Earth Systems Journal Issue: 7 Vol. 16; ISSN 1942-2466
Publisher:
American Geophysical Union (AGU)Copyright Statement
Country of Publication:
United States
Language:
English

References (63)

A review on regional convection‐permitting climate modeling: Demonstrations, prospects, and challenges journal May 2015
Earth System Modeling 2.0: A Blueprint for Models That Learn From Observations and Targeted High-Resolution Simulations: EARTH SYSTEM MODELING 2.0 journal December 2017
Response of the Quasi‐Biennial Oscillation to a warming climate in global climate models journal February 2020
The parametrization of drag induced by stratified flow over anisotropic orography journal July 2000
Recent developments in gravity-wave effects in climate models and the global distribution of gravity-wave momentum flux from observations and models journal January 2010
A Survey on Deep Transfer Learning book January 2018
Robust weighted kernel logistic regression in imbalanced and rare events data journal January 2011
A review of uncertainty quantification in deep learning: Techniques, applications and challenges journal December 2021
Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data journal October 2019
A new efficient parameter estimation algorithm for high-dimensional complex nonlinear turbulent dynamical systems with partial observations journal November 2019
Deep neural networks for data-driven LES closure models journal December 2019
Stable a posteriori LES of 2D turbulence using convolutional neural networks: Backscattering analysis and generalization to higher Re via transfer learning journal June 2022
Uncertainty quantification in scientific machine learning: Methods, metrics, and comparisons journal March 2023
Weighted logistic regression for large-scale imbalanced and rare events data journal March 2014
A systematic study of the class imbalance problem in convolutional neural networks journal October 2018
Learning physics-constrained subgrid-scale closures in the small-data regime for stable and accurate LES journal January 2023
Parameterization Schemes: Keys to Understanding Numerical Weather Prediction Models book January 2007
Sinh-arcsinh-normal distributions to add uncertainty to neural network regression tasks: Applications to tropical cyclone intensity forecasts journal January 2023
Subgrid modelling for two-dimensional turbulence using neural networks journal November 2018
On the origins of mesospheric gravity waves journal October 2009
Using Machine Learning to Parameterize Moist Convection: Potential for Modeling of Climate, Climate Change, and Extreme Events journal October 2018
Applications of Deep Learning to Ocean Data Inference and Subgrid Parameterization journal January 2019
The Whole Atmosphere Community Climate Model Version 6 (WACCM6) journal December 2019
Spatially Extended Tests of a Neural Network Parametrization Trained by Coarse‐Graining journal August 2019
Machine Learning for Stochastic Parameterization: Generative Adversarial Networks in the Lorenz '96 Model journal March 2020
Analog Forecasting of Extreme‐Causing Weather Patterns Using Deep Learning journal February 2020
Machine Learning the Warm Rain Process journal February 2021
Potential and Limitations of Machine Learning for Modeling Warm‐Rain Cloud Microphysical Processes journal November 2020
Application of Deep Learning to Estimate Atmospheric Gravity Wave Parameters in Reanalysis Data Sets journal September 2020
Data‐Driven Super‐Parameterization Using Deep Learning: Experimentation With Multiscale Lorenz 96 Systems and Transfer Learning journal November 2020
A Baseline for Global Weather and Climate Simulations at 1 km Resolution journal October 2020
Probabilistic Machine Learning Estimation of Ocean Mixed Layer Depth From Dense Satellite and Sparse In Situ Observations journal November 2021
Machine Learning Emulation of Gravity Wave Drag in Numerical Weather Forecasting journal July 2021
Improved Weather Forecasting Using Neural Network Emulation for Radiation Parameterization journal October 2021
Stochastic‐Deep Learning Parameterization of Ocean Momentum Forcing journal September 2021
Incorporating Uncertainty Into a Regression Neural Network Enables Identification of Decadal State‐Dependent Predictability in CESM2 journal August 2022
Machine Learning Gravity Wave Parameterization Generalizes to Capture the QBO and Response to Increased CO2 journal April 2022
Quantifying 3D Gravity Wave Drag in a Library of Tropical Convection‐Permitting Simulations for Data‐Driven Parameterizations journal May 2023
Revealing the Statistics of Extreme Events Hidden in Short Weather Forecast Data journal March 2023
Accelerating Atmospheric Gravity Wave Simulations Using Machine Learning: Kelvin‐Helmholtz Instability and Mountain Wave Sources Driving Gravity Wave Breaking and Secondary Gravity Wave Generation journal August 2023
Causally‐Informed Deep Learning to Improve Climate Models and Projections journal February 2024
Implementation and Evaluation of a Machine Learned Mesoscale Eddy Parameterization Into a Numerical Ocean Circulation Model journal October 2023
Explainable Offline‐Online Training of Neural Networks for Parameterizations: A 1D Gravity Wave‐QBO Testbed in the Small‐Data Regime journal January 2024
Comparing Loon Superpressure Balloon Observations of Gravity Waves in the Tropics With Global Storm‐Resolving Models journal August 2023
Searching for exotic particles in high-energy physics with deep learning journal July 2014
Evaluation of machine learning algorithms for prediction of regions of high Reynolds averaged Navier Stokes uncertainty journal August 2015
Data-driven subgrid-scale modeling of forced Burgers turbulence using deep learning with generalization to higher Reynolds numbers via transfer learning journal March 2021
Accelerating progress in climate science journal June 2021
Deep learning to represent subgrid processes in climate models journal September 2018
Using machine learning to predict extreme events in complex systems journal December 2019
Implicit learning of convective organization explains precipitation stochasticity journal May 2023
Explaining the physics of transfer learning in data-driven turbulence modeling journal January 2023
Climbing down Charney’s ladder: machine learning and the post-Dennard era of computational climate science journal February 2021
Probabilistic forecasts of extreme heatwaves using convolutional neural networks in a regime of lack of data journal April 2023
Learning Deep Representation for Imbalanced Classification conference June 2016
Climate-invariant machine learning journal February 2024
Editorial: special issue on learning from imbalanced data sets journal June 2004
Quantifying Uncertainty in Deep Spatiotemporal Forecasting conference August 2021
Toward a Physically Based Gravity Wave Source Parameterization in a General Circulation Model journal January 2010
New Approach to Calculation of Atmospheric Model Physics: Accurate and Fast Neural Network Emulation of Longwave Radiation in a Climate Model journal May 2005
A new parametrization of turbulent orographic form drag
  • Beljaars, Anton C. M.; Brown, Andrew R.; Wood, Nigel
  • Quarterly Journal of the Royal Meteorological Society, Vol. 130, Issue 599 https://doi.org/10.1256/qj.03.73
journal April 2004
An overview of the past, present and future of gravity‐wave drag parametrization for numerical climate and weather prediction models journal March 2003
Data for "Data Imbalance, Uncertainty Quantification, and Generalization via Transfer Learning in Data-driven Parameterizations: Lessons from the Emulation of Gravity Wave Momentum Transport in WACCM" dataset January 2023

Similar Records

Simulating Atmospheric Processes in Earth System Models and Quantifying Uncertainties With Deep Learning Multi‐Member and Stochastic Parameterizations
Journal Article · 2025 · Journal of Advances in Modeling Earth Systems · OSTI ID:2555915
+5 more
Development of neural network convection parameterizations for climate and NWP models using Cloud Resolving Model simulations
Technical Report · 2011 · OSTI ID:1039340
+3 more
An Ensemble of Neural Networks for Moist Physics Processes, Its Generalizability and Stable Integration
Journal Article · 2023 · Journal of Advances in Modeling Earth Systems · OSTI ID:2067628

Related Subjects