DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Spatially Extended Tests of a Neural Network Parametrization Trained by Coarse-Graining

Abstract

General circulation models (GCMs) typically have a grid size of 25–200 km. Parametrizations are used to represent diabatic processes such as radiative transfer and cloud microphysics and account for subgrid-scale motions and variability. Unlike traditional approaches, neural networks (NNs) can readily exploit recent observational data sets and global cloud-system resolving model (CRM) simulations to learn subgrid variability. This article describes an NN parametrization trained by coarse-graining a near-global CRM simulation with a 4-km horizontal grid spacing. The NN predicts the residual heating and moistening averaged over (160 km)2 grid boxes as a function of the coarse-resolution fields within the same atmospheric column. This NN is coupled to the dynamical core of a GCM with the same 16-km resolution. A recent study described how to train such an NN to be stable when coupled to specified time-evolving advective forcings in a single-column model, but feedbacks between NN and GCM components cause spatially extended simulations to crash within a few days. Analyzing the linearized response of such an NN reveals that it learns to exploit a strong synchrony between precipitation and the atmospheric state above 10 km. Removing these variables from the NN's inputs stabilizes the coupled simulations, which predict the future state more accuratelymore » than a coarse-resolution simulation without any parametrizations of subgrid-scale variability, although the mean state slowly drifts.« less

Authors:
ORCiD logo [1]; ORCiD logo [1]
  1. Univ. of Washington, Seattle, WA (United States)
Publication Date:
Research Org.:
Univ. of Washington, Seattle, WA (United States)
Sponsoring Org.:
USDOE Office of Science (SC); Univ. of Washington, Seattle, WA (United States)
OSTI Identifier:
1567951
Alternate Identifier(s):
OSTI ID: 1567953; OSTI ID: 1611814
Grant/Contract Number:  
SC0012451; SC0016433
Resource Type:
Published Article
Journal Name:
Journal of Advances in Modeling Earth Systems
Additional Journal Information:
Journal Volume: 11; Journal Issue: 8; Journal ID: ISSN 1942-2466
Publisher:
American Geophysical Union (AGU)
Country of Publication:
United States
Language:
English
Subject:
54 ENVIRONMENTAL SCIENCES; Meteorology & Atmospheric Sciences

Citation Formats

Brenowitz, Noah D., and Bretherton, Christopher S. Spatially Extended Tests of a Neural Network Parametrization Trained by Coarse-Graining. United States: N. p., 2019. Web. doi:10.1029/2019ms001711.
Brenowitz, Noah D., & Bretherton, Christopher S. Spatially Extended Tests of a Neural Network Parametrization Trained by Coarse-Graining. United States. https://doi.org/10.1029/2019ms001711
Brenowitz, Noah D., and Bretherton, Christopher S. Thu . "Spatially Extended Tests of a Neural Network Parametrization Trained by Coarse-Graining". United States. https://doi.org/10.1029/2019ms001711.
@article{osti_1567951,
title = {Spatially Extended Tests of a Neural Network Parametrization Trained by Coarse-Graining},
author = {Brenowitz, Noah D. and Bretherton, Christopher S.},
abstractNote = {General circulation models (GCMs) typically have a grid size of 25–200 km. Parametrizations are used to represent diabatic processes such as radiative transfer and cloud microphysics and account for subgrid-scale motions and variability. Unlike traditional approaches, neural networks (NNs) can readily exploit recent observational data sets and global cloud-system resolving model (CRM) simulations to learn subgrid variability. This article describes an NN parametrization trained by coarse-graining a near-global CRM simulation with a 4-km horizontal grid spacing. The NN predicts the residual heating and moistening averaged over (160 km)2 grid boxes as a function of the coarse-resolution fields within the same atmospheric column. This NN is coupled to the dynamical core of a GCM with the same 16-km resolution. A recent study described how to train such an NN to be stable when coupled to specified time-evolving advective forcings in a single-column model, but feedbacks between NN and GCM components cause spatially extended simulations to crash within a few days. Analyzing the linearized response of such an NN reveals that it learns to exploit a strong synchrony between precipitation and the atmospheric state above 10 km. Removing these variables from the NN's inputs stabilizes the coupled simulations, which predict the future state more accurately than a coarse-resolution simulation without any parametrizations of subgrid-scale variability, although the mean state slowly drifts.},
doi = {10.1029/2019ms001711},
journal = {Journal of Advances in Modeling Earth Systems},
number = 8,
volume = 11,
place = {United States},
year = {2019},
month = {8}
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record
https://doi.org/10.1029/2019ms001711

Citation Metrics:
Cited by: 63 works
Citation information provided by
Web of Science

Save / Share:

Works referenced in this record:

Sensitivity of Coupled Tropical Pacific Model Biases to Convective Parameterization in CESM1
journal, January 2018

  • Woelfle, M. D.; Yu, S.; Bretherton, C. S.
  • Journal of Advances in Modeling Earth Systems, Vol. 10, Issue 1
  • DOI: 10.1002/2017MS001176

Cloud Resolving Modeling of the ARM Summer 1997 IOP: Model Formulation, Results, Uncertainties, and Sensitivities
journal, February 2003


A Proposal for the Intercomparison of the Dynamical Cores of Atmospheric General Circulation Models
journal, October 1994


Improving the diurnal cycle of convection in GCMs
journal, December 2011

  • Stratton, R. A.; Stirling, A. J.
  • Quarterly Journal of the Royal Meteorological Society, Vol. 138, Issue 666
  • DOI: 10.1002/qj.991

Vertical structure and physical processes of the Madden-Julian oscillation: Exploring key model physics in climate simulations: KEY PHYSICS IN MODELING THE MJO
journal, May 2015

  • Jiang, Xianan; Waliser, Duane E.; Xavier, Prince K.
  • Journal of Geophysical Research: Atmospheres, Vol. 120, Issue 10
  • DOI: 10.1002/2014JD022375

Adjoint-Based Climate Model Tuning: Application to the Planet Simulator
journal, January 2018

  • Lyu, Guokun; Köhl, Armin; Matei, Ion
  • Journal of Advances in Modeling Earth Systems, Vol. 10, Issue 1
  • DOI: 10.1002/2017MS001194

Using Ensemble of Neural Networks to Learn Stochastic Convection Parameterizations for Climate and Numerical Weather Prediction Models from Data Simulated by a Cloud Resolving Model
journal, January 2013

  • Krasnopolsky, Vladimir M.; Fox-Rabinovitz, Michael S.; Belochitski, Alexei A.
  • Advances in Artificial Neural Systems, Vol. 2013
  • DOI: 10.1155/2013/485913

Using Machine Learning to Parameterize Moist Convection: Potential for Modeling of Climate, Climate Change, and Extreme Events
journal, October 2018

  • O'Gorman, Paul A.; Dwyer, John G.
  • Journal of Advances in Modeling Earth Systems, Vol. 10, Issue 10
  • DOI: 10.1029/2018MS001351

Relationships between Water Vapor Path and Precipitation over the Tropical Oceans
journal, April 2004


Tropical Super Clusters within Intraseasonal Variations over the Western Pacific [西太平洋の季節内変動中の熱帯域超雲団]
journal, January 1988


Reexamining the Nonlinear Moisture-Precipitation Relationship Over the Tropical Oceans
journal, January 2018

  • Rushley, S. S.; Kim, D.; Bretherton, C. S.
  • Geophysical Research Letters, Vol. 45, Issue 2
  • DOI: 10.1002/2017GL076296

A Neural Network Approach for a Fast and Accurate Computation of a Longwave Radiative Budget
journal, November 1998


Nonhydrostatic icosahedral atmospheric model (NICAM) for global cloud resolving simulations
journal, March 2008


Approximation by superpositions of a sigmoidal function
journal, December 1989

  • Cybenko, G.
  • Mathematics of Control, Signals, and Systems, Vol. 2, Issue 4
  • DOI: 10.1007/BF02551274

Development of neural network convection parameterizations for numerical climate and weather prediction models using cloud resolving model simulations
conference, July 2010

  • Krasnopolsky, Vladimir M.; Fox-Rabinovitz, Michael S.; Belochitski, Alexei A.
  • The 2010 International Joint Conference on Neural Networks (IJCNN)
  • DOI: 10.1109/IJCNN.2010.5596766

On large-scale circulations in convecting atmospheres
journal, July 1994

  • Emanuel, Kerry A.; David Neelin, J.; Bretherton, Christopher S.
  • Quarterly Journal of the Royal Meteorological Society, Vol. 120, Issue 519
  • DOI: 10.1002/qj.49712051902

A nonlinear dynamical perspective on model error: A proposal for non-local stochastic-dynamic parametrization in weather and climate prediction models
journal, January 2001

  • Palmer, T. N.
  • Quarterly Journal of the Royal Meteorological Society, Vol. 127, Issue 572
  • DOI: 10.1002/qj.49712757202

Prognostic Validation of a Neural Network Unified Physics Parameterization
journal, June 2018

  • Brenowitz, N. D.; Bretherton, C. S.
  • Geophysical Research Letters, Vol. 45, Issue 12
  • DOI: 10.1029/2018GL078510

Toward mitigating the double ITCZ problem in NCAR CCSM3
journal, January 2006

  • Zhang, Guang J.; Wang, Houjun
  • Geophysical Research Letters, Vol. 33, Issue 6
  • DOI: 10.1029/2005GL025229

Deep learning to represent subgrid processes in climate models
journal, September 2018

  • Rasp, Stephan; Pritchard, Michael S.; Gentine, Pierre
  • Proceedings of the National Academy of Sciences, Vol. 115, Issue 39
  • DOI: 10.1073/pnas.1810286115

Parameterizing Convective Organization to Escape the Entrainment Dilemma: PARAMETERIZING CONVECTIVE ORGANIZATION
journal, February 2011

  • Mapes, Brian; Neale, Richard
  • Journal of Advances in Modeling Earth Systems, Vol. 3, Issue 2
  • DOI: 10.1029/2011MS000042

Cloud and circulation feedbacks in a near‐global aquaplanet cloud‐resolving model
journal, May 2017

  • Narenpitak, Pornampai; Bretherton, Christopher S.; Khairoutdinov, Marat F.
  • Journal of Advances in Modeling Earth Systems, Vol. 9, Issue 2
  • DOI: 10.1002/2016MS000872

Convective self-aggregation feedbacks in near-global cloud-resolving simulations of an aquaplanet: AGGREGATION FEEDBACKS IN NEAR-GLOBAL CRM
journal, November 2015

  • Bretherton, Christopher S.; Khairoutdinov, Marat F.
  • Journal of Advances in Modeling Earth Systems, Vol. 7, Issue 4
  • DOI: 10.1002/2015MS000499

Pareto-Optimal Estimates of California Precipitation Change: PARETO-OPTIMAL PRECIPITATION CONSTRAINTS
journal, December 2017

  • Langenbrunner, Baird; Neelin, J. David
  • Geophysical Research Letters, Vol. 44, Issue 24
  • DOI: 10.1002/2017GL075226

Determination of Bulk Properties of Tropical Cloud Clusters from Large-Scale Heat and Moisture Budgets
journal, May 1973


New Approach to Calculation of Atmospheric Model Physics: Accurate and Fast Neural Network Emulation of Longwave Radiation in a Climate Model
journal, May 2005

  • Krasnopolsky, Vladimir M.; Fox-Rabinovitz, Michael S.; Chalikov, Dmitry V.
  • Monthly Weather Review, Vol. 133, Issue 5
  • DOI: 10.1175/MWR2923.1

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

  • Schneider, Tapio; Lan, Shiwei; Stuart, Andrew
  • Geophysical Research Letters, Vol. 44, Issue 24
  • DOI: 10.1002/2017GL076101

Fundamentals of synchronization in chaotic systems, concepts, and applications
journal, December 1997

  • Pecora, Louis M.; Carroll, Thomas L.; Johnson, Gregg A.
  • Chaos: An Interdisciplinary Journal of Nonlinear Science, Vol. 7, Issue 4
  • DOI: 10.1063/1.166278

The Cumulus Parameterization Problem: Past, Present, and Future
journal, July 2004


Multiscale Models with Moisture and Systematic Strategies for Superparameterization
journal, July 2007

  • Majda, Andrew J.
  • Journal of the Atmospheric Sciences, Vol. 64, Issue 7
  • DOI: 10.1175/JAS3976.1

El Niño-Southern Oscillation sensitivity to cumulus entrainment in a coupled general circulation model: ENSO SENSITIVE TO CONVECTION SCHEME
journal, November 2011

  • Kim, Daehyun; Jang, Yeon-Soo; Kim, Dong-Hoon
  • Journal of Geophysical Research: Atmospheres, Vol. 116, Issue D22
  • DOI: 10.1029/2011JD016526

Approximation by superpositions of a sigmoidal function
journal, December 1992

  • Cybenko, G.
  • Mathematics of Control, Signals, and Systems, Vol. 5, Issue 4
  • DOI: 10.1007/bf02134016

Deep learning to represent subgrid processes in climate models
text, January 2018

  • Rasp, Stephan; Pritchard, Michael S.; Gentine, Pierre
  • Columbia University
  • DOI: 10.7916/d8bz7pxt

El Niño–Southern Oscillation sensitivity to cumulus entrainment in a coupled general circulation model
text, January 2011

  • Kim, Daehyun; Jang, Yeon-Soo; Kim, Dong-Hoon
  • Columbia University
  • DOI: 10.7916/d8zp4gx1

A nonlinear dynamical perspective on model error: A proposal for non-local stochastic-dynamic parametrization in weather and climate pres-diction models
journal, January 2001

  • Palmer, Tn
  • Quarterly Journal of the Royal Meteorological Society, Vol. 127, Issue 572
  • DOI: 10.1256/smsqj.57201

Coarse-grained outputs from near-global aqua-planet control run with QOBS SST
dataset, April 2018


Prognostic validation of a neural network unified physics parameterization
text, January 2018


Prognostic validation of a neural network unified physics parameterization
text, January 2018


Coarse-grained outputs from near-global aqua-planet control run with QOBS SST
dataset, April 2018


Works referencing / citing this record:

Subgrid-scale parametrization of unresolved scales in forced Burgers equation using generative adversarial networks (GAN)
journal, November 2021


Stable machine-learning parameterization of subgrid processes for climate modeling at a range of resolutions
journal, July 2020


Machine Learning for Stochastic Parameterization: Generative Adversarial Networks in the Lorenz '96 Model
text, January 2019


WeatherBench: A benchmark dataset for data-driven weather forecasting
text, January 2020


Interpreting and Stabilizing Machine-learning Parametrizations of Convection
text, January 2020


Combining distribution-based neural networks to predict weather forecast probabilities
text, January 2021