Articles | Volume 13, issue 5
Development and technical paper
08 May 2020
Development and technical paper |  | 08 May 2020

Coupled online learning as a way to tackle instabilities and biases in neural network parameterizations: general algorithms and Lorenz 96 case study (v1.0)

Stephan Rasp

Related subject area

Atmospheric sciences
On the use of Infrared Atmospheric Sounding Interferometer (IASI) spectrally resolved radiances to test the EC-Earth climate model (v3.3.3) in clear-sky conditions
Stefano Della Fera, Federico Fabiano, Piera Raspollini, Marco Ridolfi, Ugo Cortesi, Flavio Barbara, and Jost von Hardenberg
Geosci. Model Dev., 16, 1379–1394,,, 2023
Short summary
Incorporation of aerosol into the COSPv2 satellite lidar simulator for climate model evaluation
Marine Bonazzola, Hélène Chepfer, Po-Lun Ma, Johannes Quaas, David M. Winker, Artem Feofilov, and Nick Schutgens
Geosci. Model Dev., 16, 1359–1377,,, 2023
Short summary
The impact of altering emission data precision on compression efficiency and accuracy of simulations of the community multiscale air quality model
Michael S. Walters and David C. Wong
Geosci. Model Dev., 16, 1179–1190,,, 2023
Short summary
AerSett v1.0: a simple and straightforward model for the settling speed of big spherical atmospheric aerosols
Sylvain Mailler, Laurent Menut, Arineh Cholakian, and Romain Pennel
Geosci. Model Dev., 16, 1119–1127,,, 2023
Short summary
Optimization of weather forecasting for cloud cover over the European domain using the meteorological component of the Ensemble for Stochastic Integration of Atmospheric Simulations version 1.0
Yen-Sen Lu, Garrett H. Good, and Hendrik Elbern
Geosci. Model Dev., 16, 1083–1104,,, 2023
Short summary

Cited articles

Berner, J., Fossell, K. R., Ha, S.-Y., Hacker, J. P., and Snyder, C.: Increasing the Skill of Probabilistic Forecasts: Understanding Performance Improvements from Model-Error Representations, Mon. Weather Rev., 143, 1295–1320,, 2015. a
Beucler, T., Rasp, S., Pritchard, M., and Gentine, P.: Achieving Conservation of Energy in Neural Network Emulators for Climate Modeling, available at: (last access: 6 May 2020), 2019. a, b
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,, 2019. a
Bolton, T. and Zanna, L.: Applications of Deep Learning to Ocean Data Inference and Subgrid Parameterization, J. Adv. Model. Earth Sy., 11, 376–399,, 2019. a
Breiman, L.: Random Forests, Mach. Learn., 45, 5–32,, 2001. a
Short summary
Subgrid parameterizations are largely responsible for uncertainties in climate models. Recently, several studies tried to improve the representation of subgrid processes by learning parameterization directly from high-resolution modeling data. In this paper, the current state of the art of this research direction is summarized, and an algorithm is proposed to combat major problems with existing approaches, namely instabilities and biases.