skip to main content
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Could Machine Learning Break the Convection Parameterization Deadlock?

Abstract

Representing unresolved moist convection in coarse-scale climate models remains one of the main bottlenecks of current climate simulations. Many of the biases present with parameterized convection are strongly reduced when convection is explicitly resolved (i.e., in cloud resolving models at high spatial resolution approximately a kilometer or so). We here present a novel approach to convective parameterization based on machine learning, using an aquaplanet with prescribed sea surface temperatures as a proof of concept. A deep neural network is trained with a superparameterized version of a climate model in which convection is resolved by thousands of embedded 2-D cloud resolving models. The machine learning representation of convection, which we call the Cloud Brain (CBRAIN), can skillfully predict many of the convective heating, moistening, and radiative features of superparameterization that are most important to climate simulation, although an unintended side effect is to reduce some of the superparameterization’s inherent variance. Since as few as three months’ high-frequency global training data prove sufficient to provide this skill, the approach presented here opens up a new possibility for a future class of convection parameterizations in climate models that are built “top-down,” that is, by learning salient features of convection from unusually explicit simulations.

Authors:
ORCiD logo [1]; ORCiD logo [2]; ORCiD logo [3];  [1]; ORCiD logo [2]
  1. Columbia Univ., New York, NY (United States). Earth and Environmental Engineering
  2. Univ. of California, Irvine, CA (United States). Earth System Science
  3. Ludwig Maximilian Univ., Munich (Germany). Faculty of Physics
Publication Date:
Research Org.:
Univ. of California, Irvine, CA (United States)
Sponsoring Org.:
USDOE; National Science Foundation (NSF)
OSTI Identifier:
1452657
Alternate Identifier(s):
OSTI ID: 1452658; OSTI ID: 1510994
Grant/Contract Number:  
SC0012548; SC0012152
Resource Type:
Published Article
Journal Name:
Geophysical Research Letters
Additional Journal Information:
Journal Volume: 45; Journal Issue: 11; Journal ID: ISSN 0094-8276
Publisher:
American Geophysical Union
Country of Publication:
United States
Language:
English
Subject:
54 ENVIRONMENTAL SCIENCES; 97 MATHEMATICS AND COMPUTING

Citation Formats

Gentine, P., Pritchard, M., Rasp, S., Reinaudi, G., and Yacalis, G. Could Machine Learning Break the Convection Parameterization Deadlock?. United States: N. p., 2018. Web. doi:10.1029/2018gl078202.
Gentine, P., Pritchard, M., Rasp, S., Reinaudi, G., & Yacalis, G. Could Machine Learning Break the Convection Parameterization Deadlock?. United States. doi:10.1029/2018gl078202.
Gentine, P., Pritchard, M., Rasp, S., Reinaudi, G., and Yacalis, G. Wed . "Could Machine Learning Break the Convection Parameterization Deadlock?". United States. doi:10.1029/2018gl078202.
@article{osti_1452657,
title = {Could Machine Learning Break the Convection Parameterization Deadlock?},
author = {Gentine, P. and Pritchard, M. and Rasp, S. and Reinaudi, G. and Yacalis, G.},
abstractNote = {Representing unresolved moist convection in coarse-scale climate models remains one of the main bottlenecks of current climate simulations. Many of the biases present with parameterized convection are strongly reduced when convection is explicitly resolved (i.e., in cloud resolving models at high spatial resolution approximately a kilometer or so). We here present a novel approach to convective parameterization based on machine learning, using an aquaplanet with prescribed sea surface temperatures as a proof of concept. A deep neural network is trained with a superparameterized version of a climate model in which convection is resolved by thousands of embedded 2-D cloud resolving models. The machine learning representation of convection, which we call the Cloud Brain (CBRAIN), can skillfully predict many of the convective heating, moistening, and radiative features of superparameterization that are most important to climate simulation, although an unintended side effect is to reduce some of the superparameterization’s inherent variance. Since as few as three months’ high-frequency global training data prove sufficient to provide this skill, the approach presented here opens up a new possibility for a future class of convection parameterizations in climate models that are built “top-down,” that is, by learning salient features of convection from unusually explicit simulations.},
doi = {10.1029/2018gl078202},
journal = {Geophysical Research Letters},
number = 11,
volume = 45,
place = {United States},
year = {2018},
month = {5}
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record
DOI: 10.1029/2018gl078202

Citation Metrics:
Cited by: 16 works
Citation information provided by
Web of Science

Save / Share: