Articles | Volume 19, issue 9
https://doi.org/10.5194/gmd-19-3875-2026
© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/gmd-19-3875-2026
© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Development and technical paper
| 
13 May 2026
Development and technical paper |
| 13 May 2026
| 13 May 2026
Representing subgrid-scale cloud effects in a radiation parameterization using machine learning: MLe-radiation v1.0
Katharina Hafner
CORRESPONDING AUTHOR
University of Bremen, Institute of Environmental Physics (IUP), Bremen, Germany
Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Physik der Atmosphäre, Oberpfaffenhofen, Germany
Sara Shamekh
Courant Institute of Mathematical Sciences, New York University (NYU), New York, NY, USA
Guillaume Bertoli
Department of Earth and Environmental Engineering, Columbia University, New York, NY, USA
Axel Lauer
Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Physik der Atmosphäre, Oberpfaffenhofen, Germany
Robert Pincus
Lamont-Doherty Earth Observatory, Palisades, New York, USA
Julien Savre
Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Physik der Atmosphäre, Oberpfaffenhofen, Germany
Veronika Eyring
Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Physik der Atmosphäre, Oberpfaffenhofen, Germany
University of Bremen, Institute of Environmental Physics (IUP), Bremen, Germany
Related authors
No articles found.
Diajeng W. Atmojo, Katja Weigel, Arthur Grundner, Marika M. Holland, Dmitry Sidorenko, and Veronika Eyring
The Cryosphere, 20, 4437–4464, https://doi.org/10.5194/tc-20-4437-2026, https://doi.org/10.5194/tc-20-4437-2026, 2026
Short summary
Short summary
This study presents an observation-driven sea ice albedo parametrisation by discovering an equation using symbolic regression, an interpretable machine learning method. Leveraging satellite and reanalyses data, our discovered equation identifies high sensitivity to thin snow and the weighted temperature difference between sea ice surface and 2 m air as critical to determine sea ice albedo. Our findings contribute to improving Arctic climate projections and understanding.
Forrest M. Hoffman, Birgit Hassler, Ranjini Swaminathan, Jared Lewis, Bouwe Andela, Nathan Collier, Dóra Hegedűs, Jiwoo Lee, Charlotte Pascoe, Mika Pflüger, Martina Stockhause, Paul Ullrich, Min Xu, Lisa Bock, Felicity Chun, Bettina K. Gier, Douglas I. Kelley, Axel Lauer, Julien Lenhardt, Manuel Schlund, Mohanan G. Sreeush, Katja Weigel, Ed Blockley, Rebecca Beadling, Romain Beucher, Demiso D. Dugassa, Valerio Lembo, Jianhua Lu, Swen Brands, Jerry Tjiputra, Elizaveta Malinina, Brian Medeiros, Enrico Scoccimarro, Jeremy Walton, Phil Kershaw, André Lanfer Marquez, Malcolm J. Roberts, Eleanor O'Rourke, Beth Dingley, Briony Turner, Helene Hewitt, and John P. Dunne
Geosci. Model Dev., 19, 7415–7455, https://doi.org/10.5194/gmd-19-7415-2026, https://doi.org/10.5194/gmd-19-7415-2026, 2026
Short summary
Short summary
The Rapid Evaluation Framework (REF) is a community-driven platform for benchmarking and performance assessment of Earth system models. Built upon four disparate community evaluation tools, the REF is designed to provide model-data comparisons for the Assessment Fast Track for the Seventh Phase of the Coupled Model Intercomparison Project. The REF will be run at the Earth System Grid Federation to enable model devleopers and scientists to quickly identify model biases and performance issues.
Lara Leko, Gunnar Behrens, Nils Müller, Adrià Amell, Axel Lauer, and Patrick Eriksson
EGUsphere, https://doi.org/10.5194/egusphere-2026-3869, https://doi.org/10.5194/egusphere-2026-3869, 2026
This preprint is open for discussion and under review for Atmospheric Chemistry and Physics (ACP).
Short summary
Short summary
We look at daily cycles of tropical ice clouds and their representation in models. New satellite data of ice clouds captures daily cycles over land and oceans, which was not possible yet. These are often shifted in time in reanalysis, and biased depending on the variable. High-resolution models show improvements but still have regional biases. We show that combining different satellite observations helps to understand daily cycles and improve the simulation of tropical ice clouds and rainfall.
Aytaç Paçal, Birgit Hassler, Katja Weigel, Miguel-Ángel Fernández-Torres, Gustau Camps-Valls, and Veronika Eyring
Earth Syst. Dynam., 17, 955–986, https://doi.org/10.5194/esd-17-955-2026, https://doi.org/10.5194/esd-17-955-2026, 2026
Short summary
Short summary
Heatwaves are among the deadliest natural hazards, yet their evolving atmospheric drivers remain unclear. We apply an unsupervised machine learning model to reanalysis data to identify multivariate heatwave patterns over Europe. The model captures seasonally distinct regimes and reveals significant shifts between historical and recent events across all seasons, highlighting changes in heatwave dynamics over time.
Axel Lauer, Manuel Schlund, Lisa Bock, Birgit Hassler, Gunnar Behrens, Bettina Gier, Lukas Lindenlaub, Stephan Lorenz, Jan-Hendrik Malles, Wolfgang A. Müller, Trang van Pham, Katja Weigel, Guang Zeng, and Veronika Eyring
EGUsphere, https://doi.org/10.5194/egusphere-2026-2288, https://doi.org/10.5194/egusphere-2026-2288, 2026
Short summary
Short summary
ICONEval is a new framework to facilitate researchers to continuously monitor and check their climate models during the development phase. It builds on the ESMValTool software package to run tests verifying that the results are plausible, follow physical laws, and that the model has sufficient skill in reproducing the observed climate. An important aim is to make it easier to spot model errors early, for example when implementing new model components that use machine learning.
Detlef P. Van Vuuren, Brian C. O'Neill, Claudia Tebaldi, Benjamin M. Sanderson, Louise P. Chini, Pierre Friedlingstein, Tomoko Hasegawa, Keywan Riahi, Bala Govindasamy, Nico Bauer, Veronika Eyring, Cheikh M. N. Fall, Katja Frieler, Matthew J. Gidden, Laila K. Gohar, Annika Högner, Andrew D. Jones, Jarmo Kikstra, Andrew King, Reto Knutti, Elmar Kriegler, Peter Lawrence, Chris Lennard, Jason Lowe, Camilla Mathison, Shahbaz Mehmood, Zebedee Nicholls, Luciana F. Prado, Qiang Zhang, Steven K. Rose, Alex C. Ruane, Marit Sandstad, Carl-Friedrich Schleussner, Roland Seferian, Jana Sillmann, Chris Smith, Anna A. Sörensson, Swapna Panickal, Kaoru Tachiiri, Naomi Vaughan, Saritha S. Vishwanathan, Tokuta Yokohata, Marco Zecchetto, and Tilo Ziehn
Geosci. Model Dev., 19, 2627–2656, https://doi.org/10.5194/gmd-19-2627-2026, https://doi.org/10.5194/gmd-19-2627-2026, 2026
Short summary
Short summary
We propose a set of seven plausible 21st century emission scenarios, and their multi-century extensions, that will be used by the international community of climate modeling centers to produce the next generation of climate projections. These projections will support climate, impact and mitigation researchers, provide information to practitioners to address future risks from climate change, and contribute to policymakers’ considerations of the trade-offs among various levels of mitigation.
Lukas Lindenlaub, Katja Weigel, Birgit Hassler, Colin Jones, and Veronika Eyring
Earth Syst. Dynam., 17, 81–105, https://doi.org/10.5194/esd-17-81-2026, https://doi.org/10.5194/esd-17-81-2026, 2026
Short summary
Short summary
This study explores changes in drought characteristic based on projections by 18 different Earth system models. Their performance is evaluated by comparing historical simulations to observation based reanalysis. The analysis of a standardized drought index under different future scenarios revealed that the harvest area that is projected to experience extreme drought conditions towards the end of this century ranges from 10 % to 40 % depending on the emission scenario.
John P. Dunne, Helene T. Hewitt, Julie M. Arblaster, Frédéric Bonou, Olivier Boucher, Tereza Cavazos, Beth Dingley, Paul J. Durack, Birgit Hassler, Martin Juckes, Tomoki Miyakawa, Matt Mizielinski, Vaishali Naik, Zebedee Nicholls, Eleanor O'Rourke, Robert Pincus, Benjamin M. Sanderson, Isla R. Simpson, and Karl E. Taylor
Geosci. Model Dev., 18, 6671–6700, https://doi.org/10.5194/gmd-18-6671-2025, https://doi.org/10.5194/gmd-18-6671-2025, 2025
Short summary
Short summary
The seventh phase of the Coupled Model Intercomparison Project (CMIP7) coordinates efforts to answer key and timely climate science questions and facilitate delivery of relevant multi-model simulations for prediction and projection; characterization, attribution, and process understanding; and vulnerability, impact, and adaptation analysis. Key to the CMIP7 design are the mandatory Diagnostic, Evaluation and Characterization of Klima and optional Assessment Fast Track experiments.
Manuel Schlund, Bouwe Andela, Jörg Benke, Ruth Comer, Birgit Hassler, Emma Hogan, Peter Kalverla, Axel Lauer, Bill Little, Saskia Loosveldt Tomas, Francesco Nattino, Patrick Peglar, Valeriu Predoi, Stef Smeets, Stephen Worsley, Martin Yeo, and Klaus Zimmermann
Geosci. Model Dev., 18, 4009–4021, https://doi.org/10.5194/gmd-18-4009-2025, https://doi.org/10.5194/gmd-18-4009-2025, 2025
Short summary
Short summary
The Earth System Model Evaluation Tool (ESMValTool) is a community diagnostics and performance metrics tool for the evaluation of Earth system models. Here, we describe recent significant improvements of ESMValTool’s computational efficiency including parallel, out-of-core, and distributed computing. Evaluations with the enhanced version of ESMValTool are faster, use less computational resources, and can handle input data larger than the available memory.
Pauline Bonnet, Lorenzo Pastori, Mierk Schwabe, Marco Giorgetta, Fernando Iglesias-Suarez, and Veronika Eyring
Geosci. Model Dev., 18, 3681–3706, https://doi.org/10.5194/gmd-18-3681-2025, https://doi.org/10.5194/gmd-18-3681-2025, 2025
Short summary
Short summary
Tuning a climate model means adjusting uncertain parameters in the model to best match observations like the global radiation balance and cloud cover. This is usually done by running many simulations of the model with different settings, which can be time-consuming and relies heavily on expert knowledge. To make this process faster and more objective, we developed a machine learning emulator to create a large ensemble and apply a method called history matching to find the best settings.
Kevin Debeire, Lisa Bock, Peer Nowack, Jakob Runge, and Veronika Eyring
Earth Syst. Dynam., 16, 607–630, https://doi.org/10.5194/esd-16-607-2025, https://doi.org/10.5194/esd-16-607-2025, 2025
Short summary
Short summary
Projecting future precipitation is essential for preparing for climate change, but current climate models still have large uncertainties, especially over land. This study presents a new method to improve precipitation projections by identifying which models best capture key climate patterns. By giving more weight to models that better represent these patterns, our approach leads to more reliable future precipitation projections over land.
Axel Lauer, Lisa Bock, Birgit Hassler, Patrick Jöckel, Lukas Ruhe, and Manuel Schlund
Geosci. Model Dev., 18, 1169–1188, https://doi.org/10.5194/gmd-18-1169-2025, https://doi.org/10.5194/gmd-18-1169-2025, 2025
Short summary
Short summary
Earth system models are important tools to improve our understanding of current climate and to project climate change. Thus, it is crucial to understand possible shortcomings in the models. New features of the ESMValTool software package allow one to compare and visualize a model's performance with respect to reproducing observations in the context of other climate models in an easy and user-friendly way. We aim to help model developers assess and monitor climate simulations more efficiently.
Paul J. Durack, Karl E. Taylor, Peter J. Gleckler, Gerald A. Meehl, Bryan N. Lawrence, Curt Covey, Ronald J. Stouffer, Guillaume Levavasseur, Atef Ben-Nasser, Sebastien Denvil, Martina Stockhause, Jonathan M. Gregory, Martin Juckes, Sasha K. Ames, Fabrizio Antonio, David C. Bader, John P. Dunne, Daniel Ellis, Veronika Eyring, Sandro L. Fiore, Sylvie Joussaume, Philip Kershaw, Jean-Francois Lamarque, Michael Lautenschlager, Jiwoo Lee, Chris F. Mauzey, Matthew Mizielinski, Paola Nassisi, Alessandra Nuzzo, Eleanor O’Rourke, Jeffrey Painter, Gerald L. Potter, Sven Rodriguez, and Dean N. Williams
EGUsphere, https://doi.org/10.5194/egusphere-2024-3729, https://doi.org/10.5194/egusphere-2024-3729, 2025
Short summary
Short summary
CMIP6 was the most expansive and ambitious Model Intercomparison Project (MIP), the latest in a history, extending four decades. CMIP engaged a growing community focused on improving climate understanding, and quantifying and attributing observed climate change being experienced today. The project's profound impact is due to the combining the latest climate science and technology, enabling the latest-generation climate simulations and increasing community attention in every successive phase.
Bettina K. Gier, Manuel Schlund, Pierre Friedlingstein, Chris D. Jones, Colin Jones, Sönke Zaehle, and Veronika Eyring
Biogeosciences, 21, 5321–5360, https://doi.org/10.5194/bg-21-5321-2024, https://doi.org/10.5194/bg-21-5321-2024, 2024
Short summary
Short summary
This study investigates present-day carbon cycle variables in CMIP5 and CMIP6 simulations. Overall, CMIP6 models perform better but also show many remaining biases. A significant improvement in the simulation of photosynthesis in models with a nitrogen cycle is found, with only small differences between emission- and concentration-based simulations. Thus, we recommend using emission-driven simulations in CMIP7 by default and including the nitrogen cycle in all future carbon cycle models.
Benjamin M. Sanderson, Ben B. B. Booth, John Dunne, Veronika Eyring, Rosie A. Fisher, Pierre Friedlingstein, Matthew J. Gidden, Tomohiro Hajima, Chris D. Jones, Colin G. Jones, Andrew King, Charles D. Koven, David M. Lawrence, Jason Lowe, Nadine Mengis, Glen P. Peters, Joeri Rogelj, Chris Smith, Abigail C. Snyder, Isla R. Simpson, Abigail L. S. Swann, Claudia Tebaldi, Tatiana Ilyina, Carl-Friedrich Schleussner, Roland Séférian, Bjørn H. Samset, Detlef van Vuuren, and Sönke Zaehle
Geosci. Model Dev., 17, 8141–8172, https://doi.org/10.5194/gmd-17-8141-2024, https://doi.org/10.5194/gmd-17-8141-2024, 2024
Short summary
Short summary
We discuss how, in order to provide more relevant guidance for climate policy, coordinated climate experiments should adopt a greater focus on simulations where Earth system models are provided with carbon emissions from fossil fuels together with land use change instructions, rather than past approaches that have largely focused on experiments with prescribed atmospheric carbon dioxide concentrations. We discuss how these goals might be achieved in coordinated climate modeling experiments.
Colin G. Jones, Fanny Adloff, Ben B. B. Booth, Peter M. Cox, Veronika Eyring, Pierre Friedlingstein, Katja Frieler, Helene T. Hewitt, Hazel A. Jeffery, Sylvie Joussaume, Torben Koenigk, Bryan N. Lawrence, Eleanor O'Rourke, Malcolm J. Roberts, Benjamin M. Sanderson, Roland Séférian, Samuel Somot, Pier Luigi Vidale, Detlef van Vuuren, Mario Acosta, Mats Bentsen, Raffaele Bernardello, Richard Betts, Ed Blockley, Julien Boé, Tom Bracegirdle, Pascale Braconnot, Victor Brovkin, Carlo Buontempo, Francisco Doblas-Reyes, Markus Donat, Italo Epicoco, Pete Falloon, Sandro Fiore, Thomas Frölicher, Neven S. Fučkar, Matthew J. Gidden, Helge F. Goessling, Rune Grand Graversen, Silvio Gualdi, José M. Gutiérrez, Tatiana Ilyina, Daniela Jacob, Chris D. Jones, Martin Juckes, Elizabeth Kendon, Erik Kjellström, Reto Knutti, Jason Lowe, Matthew Mizielinski, Paola Nassisi, Michael Obersteiner, Pierre Regnier, Romain Roehrig, David Salas y Mélia, Carl-Friedrich Schleussner, Michael Schulz, Enrico Scoccimarro, Laurent Terray, Hannes Thiemann, Richard A. Wood, Shuting Yang, and Sönke Zaehle
Earth Syst. Dynam., 15, 1319–1351, https://doi.org/10.5194/esd-15-1319-2024, https://doi.org/10.5194/esd-15-1319-2024, 2024
Short summary
Short summary
We propose a number of priority areas for the international climate research community to address over the coming decade. Advances in these areas will both increase our understanding of past and future Earth system change, including the societal and environmental impacts of this change, and deliver significantly improved scientific support to international climate policy, such as future IPCC assessments and the UNFCCC Global Stocktake.
Arndt Kaps, Axel Lauer, Rémi Kazeroni, Martin Stengel, and Veronika Eyring
Earth Syst. Sci. Data, 16, 3001–3016, https://doi.org/10.5194/essd-16-3001-2024, https://doi.org/10.5194/essd-16-3001-2024, 2024
Short summary
Short summary
CCClim displays observations of clouds in terms of cloud classes that have been in use for a long time. CCClim is a machine-learning-powered product based on multiple existing observational products from different satellites. We show that the cloud classes in CCClim are physically meaningful and can be used to study cloud characteristics in more detail. The goal of this is to make real-world clouds more easily understandable to eventually improve the simulation of clouds in climate models.
Soufiane Karmouche, Evgenia Galytska, Gerald A. Meehl, Jakob Runge, Katja Weigel, and Veronika Eyring
Earth Syst. Dynam., 15, 689–715, https://doi.org/10.5194/esd-15-689-2024, https://doi.org/10.5194/esd-15-689-2024, 2024
Short summary
Short summary
This study explores Atlantic–Pacific interactions and their response to external factors. Causal analysis of 1950–2014 data reveals a shift from a Pacific- to an Atlantic-driven regime. Contrasting impacts between El Niño and tropical Atlantic temperatures are highlighted, along with different pathways connecting the two oceans. The findings also suggest increasing remote contributions of forced Atlantic responses in modulating local Pacific responses during the most recent analyzed decades.
Bjorn Stevens, Stefan Adami, Tariq Ali, Hartwig Anzt, Zafer Aslan, Sabine Attinger, Jaana Bäck, Johanna Baehr, Peter Bauer, Natacha Bernier, Bob Bishop, Hendryk Bockelmann, Sandrine Bony, Guy Brasseur, David N. Bresch, Sean Breyer, Gilbert Brunet, Pier Luigi Buttigieg, Junji Cao, Christelle Castet, Yafang Cheng, Ayantika Dey Choudhury, Deborah Coen, Susanne Crewell, Atish Dabholkar, Qing Dai, Francisco Doblas-Reyes, Dale Durran, Ayoub El Gaidi, Charlie Ewen, Eleftheria Exarchou, Veronika Eyring, Florencia Falkinhoff, David Farrell, Piers M. Forster, Ariane Frassoni, Claudia Frauen, Oliver Fuhrer, Shahzad Gani, Edwin Gerber, Debra Goldfarb, Jens Grieger, Nicolas Gruber, Wilco Hazeleger, Rolf Herken, Chris Hewitt, Torsten Hoefler, Huang-Hsiung Hsu, Daniela Jacob, Alexandra Jahn, Christian Jakob, Thomas Jung, Christopher Kadow, In-Sik Kang, Sarah Kang, Karthik Kashinath, Katharina Kleinen-von Königslöw, Daniel Klocke, Uta Kloenne, Milan Klöwer, Chihiro Kodama, Stefan Kollet, Tobias Kölling, Jenni Kontkanen, Steve Kopp, Michal Koran, Markku Kulmala, Hanna Lappalainen, Fakhria Latifi, Bryan Lawrence, June Yi Lee, Quentin Lejeun, Christian Lessig, Chao Li, Thomas Lippert, Jürg Luterbacher, Pekka Manninen, Jochem Marotzke, Satoshi Matsouoka, Charlotte Merchant, Peter Messmer, Gero Michel, Kristel Michielsen, Tomoki Miyakawa, Jens Müller, Ramsha Munir, Sandeep Narayanasetti, Ousmane Ndiaye, Carlos Nobre, Achim Oberg, Riko Oki, Tuba Özkan-Haller, Tim Palmer, Stan Posey, Andreas Prein, Odessa Primus, Mike Pritchard, Julie Pullen, Dian Putrasahan, Johannes Quaas, Krishnan Raghavan, Venkatachalam Ramaswamy, Markus Rapp, Florian Rauser, Markus Reichstein, Aromar Revi, Sonakshi Saluja, Masaki Satoh, Vera Schemann, Sebastian Schemm, Christina Schnadt Poberaj, Thomas Schulthess, Cath Senior, Jagadish Shukla, Manmeet Singh, Julia Slingo, Adam Sobel, Silvina Solman, Jenna Spitzer, Philip Stier, Thomas Stocker, Sarah Strock, Hang Su, Petteri Taalas, John Taylor, Susann Tegtmeier, Georg Teutsch, Adrian Tompkins, Uwe Ulbrich, Pier-Luigi Vidale, Chien-Ming Wu, Hao Xu, Najibullah Zaki, Laure Zanna, Tianjun Zhou, and Florian Ziemen
Earth Syst. Sci. Data, 16, 2113–2122, https://doi.org/10.5194/essd-16-2113-2024, https://doi.org/10.5194/essd-16-2113-2024, 2024
Short summary
Short summary
To manage Earth in the Anthropocene, new tools, new institutions, and new forms of international cooperation will be required. Earth Virtualization Engines is proposed as an international federation of centers of excellence to empower all people to respond to the immense and urgent challenges posed by climate change.
Lisa Bock and Axel Lauer
Atmos. Chem. Phys., 24, 1587–1605, https://doi.org/10.5194/acp-24-1587-2024, https://doi.org/10.5194/acp-24-1587-2024, 2024
Short summary
Short summary
Climate model simulations still show a large range of effective climate sensitivity (ECS) with high uncertainties. An important contribution to ECS is cloud climate feedback. We investigate the representation of cloud physical and radiative properties from Coupled Model Intercomparison Project models grouped by ECS. We compare the simulated cloud properties of today’s climate from three ECS groups and quantify how the projected changes in cloud properties and cloud radiative effects differ.
Soufiane Karmouche, Evgenia Galytska, Jakob Runge, Gerald A. Meehl, Adam S. Phillips, Katja Weigel, and Veronika Eyring
Earth Syst. Dynam., 14, 309–344, https://doi.org/10.5194/esd-14-309-2023, https://doi.org/10.5194/esd-14-309-2023, 2023
Short summary
Short summary
This study uses a causal discovery method to evaluate the ability of climate models to represent the interactions between the Atlantic multidecadal variability (AMV) and the Pacific decadal variability (PDV). The approach and findings in this study present a powerful methodology that can be applied to a number of environment-related topics, offering tremendous insights to improve the understanding of the complex Earth system and the state of the art of climate modeling.
Manuel Schlund, Birgit Hassler, Axel Lauer, Bouwe Andela, Patrick Jöckel, Rémi Kazeroni, Saskia Loosveldt Tomas, Brian Medeiros, Valeriu Predoi, Stéphane Sénési, Jérôme Servonnat, Tobias Stacke, Javier Vegas-Regidor, Klaus Zimmermann, and Veronika Eyring
Geosci. Model Dev., 16, 315–333, https://doi.org/10.5194/gmd-16-315-2023, https://doi.org/10.5194/gmd-16-315-2023, 2023
Short summary
Short summary
The Earth System Model Evaluation Tool (ESMValTool) is a community diagnostics and performance metrics tool for routine evaluation of Earth system models. Originally, ESMValTool was designed to process reformatted output provided by large model intercomparison projects like the Coupled Model Intercomparison Project (CMIP). Here, we describe a new extension of ESMValTool that allows for reading and processing native climate model output, i.e., data that have not been reformatted before.
Marco A. Giorgetta, William Sawyer, Xavier Lapillonne, Panagiotis Adamidis, Dmitry Alexeev, Valentin Clément, Remo Dietlicher, Jan Frederik Engels, Monika Esch, Henning Franke, Claudia Frauen, Walter M. Hannah, Benjamin R. Hillman, Luis Kornblueh, Philippe Marti, Matthew R. Norman, Robert Pincus, Sebastian Rast, Daniel Reinert, Reiner Schnur, Uwe Schulzweida, and Bjorn Stevens
Geosci. Model Dev., 15, 6985–7016, https://doi.org/10.5194/gmd-15-6985-2022, https://doi.org/10.5194/gmd-15-6985-2022, 2022
Short summary
Short summary
This work presents a first version of the ICON atmosphere model that works not only on CPUs, but also on GPUs. This GPU-enabled ICON version is benchmarked on two GPU machines and a CPU machine. While the weak scaling is very good on CPUs and GPUs, the strong scaling is poor on GPUs. But the high performance of GPU machines allowed for first simulations of a short period of the quasi-biennial oscillation at very high resolution with explicit convection and gravity wave forcing.
Cited articles
Barker, H. W., Stephens, G. L., and Fu, Q.: The sensitivity of domain-averaged solar fluxes to assumptions about cloud geometry, Q. J. Roy. Meteor. Soc., 125, 2127–2152, https://doi.org/10.1002/qj.49712555810, 1999. a
Bertoli, G., Mohebi, S., Ozdemir, F., Jucker, J., Rüdisühli, S., Perez-Cruz, F., Salzmann, M., and Schemm, S.: Revisiting Machine Learning Approaches for Short- and Longwave Radiation Inference in Weather and Climate Models, J. Adv. Model. Earth Sy., 17, https://doi.org/10.1029/2025ms004956, 2025. a, b, c
Beucler, T., Gentine, P., Yuval, J., Gupta, A., Peng, L., Lin, J., Yu, S., Rasp, S., Ahmed, F., O’Gorman, P. A., Neelin, J. D., Lutsko, N. J., and Pritchard, M.: Climate-invariant machine learning, Science Advances, 10, eadj7250, https://doi.org/10.1126/sciadv.adj7250, 2024. a
Bock, L. and Lauer, A.: Cloud properties and their projected changes in CMIP models with low to high climate sensitivity, Atmos. Chem. Phys., 24, 1587–1605, https://doi.org/10.5194/acp-24-1587-2024, 2024. a, b
Bony, S., Stevens, B., Frierson, D. M. W., Jakob, C., Kageyama, M., Pincus, R., Shepherd, T. G., Sherwood, S. C., Siebesma, A. P., Sobel, A. H., Watanabe, M., and Webb, M. J.: Clouds, circulation and climate sensitivity, Nat. Geosci., 8, 261–268, https://doi.org/10.1038/ngeo2398, 2015. a, b
Bretherton, C. S., Henn, B., Kwa, A., Brenowitz, N. D., Watt-Meyer, O., McGibbon, J., Perkins, W. A., Clark, S. K., and Harris, L.: Correcting Coarse-Grid Weather and Climate Models by Machine Learning From Global Storm-Resolving Simulations, J. Adv. Model. Earth Sy., 14, https://doi.org/10.1029/2021ms002794, 2022. a
Busecke, J. J. M., Balwada, D., Martin, P. E., Nicholas, T. E. G., Johnson, Z. C. P., Nalluri, P., Stern, C. I., and Abernathey, R. P.: The Impact of Sub-Grid Heterogeneity on Air-Sea Turbulent Heat Flux in Coupled Climate Models, Geophys. Res. Lett., 52, https://doi.org/10.1029/2025gl114951, 2025. a
Chen, D., Rojas, M., Samset, B., Cobb, K., Diongue Niang, A., Edwards, P., Emori, S., Faria, S., Hawkins, E., Hope, P., Huybrechts, P., Meinshausen, M., Mustafa, S., Plattner, G.-K., and Tréguier, A.-M.: Framing, Context, and Methods, in: Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, edited by: Masson-Delmotte, V., Zhai, P., Pirani, A., Connors, S. L., Péan, C., Berger, S., Caud, N., Chen, Y., Goldfarb, L., Gomis, M. I., Huang, M., Leitzell, K., Lonnoy, E., Matthews, J. B. R., Maycock, T. K., Waterfield, T., Yelekçi, O., Yu, R., and Zhou, B., book section 1, pp. 147–286, Cambridge University Press, Cambridge, UK and New York, NY, USA, https://doi.org/10.1017/9781009157896.003, 2021. a
Chevallier, F., Chéruy, F., Scott, N. A., and Chédin, A.: A Neural Network Approach for a Fast and Accurate Computation of a Longwave Radiative Budget, J. Appl. Meteorol. Clim., 37, 1385–1397, https://doi.org/10.1175/1520-0450(1998)037<1385:ANNAFA>2.0.CO;2, 1998. a
Cotronei, A. and Slawig, T.: Single-precision arithmetic in ECHAM radiation reduces runtime and energy consumption, Geosci. Model Dev., 13, 2783–2804, https://doi.org/10.5194/gmd-13-2783-2020, 2020. a
Doktorowski, S., Kretzschmar, J., Quaas, J., Salzmann, M., and Sourdeval, O.: Subgrid-scale variability of cloud ice in the ICON-AES 1.3.00, Geosci. Model Dev., 17, 3099–3110, https://doi.org/10.5194/gmd-17-3099-2024, 2024. a
Doms, G., Förstner, G., Heise, E., Herzog, H.-J., Mironov, D., Raschendorfer, M., Reinhardt, T., Ritter, B., Schrodin, R., Schulz, J.-P., and Vogel, G.: A Description of the Nonhydrostatic Regional COSMO Model. Part II: Physical Parameterization, Consortium for Small-Scale Modelling, http://www.cosmo-model.org/ (last access: 3 May 2026), 2011. a, b
Giorgetta, M. A., Brokopf, R., Crueger, T., Esch, M., Fiedler, S., Helmert, J., Hohenegger, C., Kornblueh, L., Köhler, M., Manzini, E., Mauritsen, T., Nam, C., Raddatz, T., Rast, S., Reinert, D., Sakradzija, M., Schmidt, H., Schneck, R., Schnur, R., Silvers, L., Wan, H., Zängl, G., and Stevens, B.: ICON-A, the Atmosphere Component of the ICON Earth System Model: I. Model Description, J. Adv. Model. Earth Sy., 10, 1613–1637, https://doi.org/10.1029/2017MS001242, 2018. a, b, c, d
Giorgetta, M. A., Sawyer, W., Lapillonne, X., Adamidis, P., Alexeev, D., Clément, V., Dietlicher, R., Engels, J. F., Esch, M., Franke, H., Frauen, C., Hannah, W. M., Hillman, B. R., Kornblueh, L., Marti, P., Norman, M. R., Pincus, R., Rast, S., Reinert, D., Schnur, R., Schulzweida, U., and Stevens, B.: The ICON-A model for direct QBO simulations on GPUs (version icon-cscs:baf28a514), Geosci. Model Dev., 15, 6985–7016, https://doi.org/10.5194/gmd-15-6985-2022, 2022. a, b, c, d, e
Grundner, A., Beucler, T., Gentine, P., Iglesias-Suarez, F., Giorgetta, M. A., and Eyring, V.: Deep Learning Based Cloud Cover Parameterization for ICON, J. Adv. Model. Earth Sy., 14, e2021MS002959, https://doi.org/10.1029/2021MS002959, 2022. a, b
Hafner, K.: Representing Subgrid-Scale Cloud Effects in a Radiation Parameterization using Machine Learning: MLe-radiation v1.0, Zenodo [code], https://doi.org/10.5281/ZENODO.17280639, 2025. a
Hafner, K.: Code and Data for “Representing Subgrid-Scale Cloud Effects in a Radiation Parameterization using Machine Learning: MLe-radiation v1.0”, Zenodo [code and data set], https://doi.org/10.5281/zenodo.18853569, 2026. a, b
Hafner, K., Iglesiaz-Suarez, F., Shamekh, S., Gentine, P., Pincus, R., Giorgetta, M., and Eyring, V.: Stable Machine Learning based Radiation Emulation for ICON, J. Adv. Model. Earth Sy., https://doi.org/10.22541/essoar.174708082.27787580/v1, in review, 2025a. a, b, c
Hafner, K., Iglesiaz-Suarez, F., Shamekh, S., Gentine, P., Pincus, R., Girgetta, M., and Eyring, V.: Interpretable machine learning radiation parameterization for ICON, J. Geophys. Res.-Mach. Learn. Comput., 2, e2024JH000501, https://doi.org/10.1029/2024JH000501, 2025b. a, b, c
Henn, B., Jauregui, Y. R., Clark, S. K., Brenowitz, N. D., McGibbon, J., Watt-Meyer, O., Pauling, A. G., and Bretherton, C. S.: A Machine Learning Parameterization of Clouds in a Coarse-Resolution Climate Model for Unbiased Radiation, J. Adv. Model. Earth Sy., 16, https://doi.org/10.1029/2023ms003949, 2024. a
Heuer, H., Schwabe, M., Gentine, P., Giorgetta, M. A., and Eyring, V.: Interpretable Multiscale Machine Learning-Based Parameterizations of Convection for ICON, J. Adv. Model. Earth Sy., 16, https://doi.org/10.1029/2024ms004398, 2024. a
Hogan, R. J. and Matricardi, M.: A Tool for Generating Fast k-Distribution Gas-Optics Models for Weather and Climate Applications, J. Adv. Model. Earth Sy., 14, e2022MS003033, https://doi.org/10.1029/2022MS003033, 2022. a
Hohenegger, C., Korn, P., Linardakis, L., Redler, R., Schnur, R., Adamidis, P., Bao, J., Bastin, S., Behravesh, M., Bergemann, M., Biercamp, J., Bockelmann, H., Brokopf, R., Brüggemann, N., Casaroli, L., Chegini, F., Datseris, G., Esch, M., George, G., Giorgetta, M., Gutjahr, O., Haak, H., Hanke, M., Ilyina, T., Jahns, T., Jungclaus, J., Kern, M., Klocke, D., Kluft, L., Kölling, T., Kornblueh, L., Kosukhin, S., Kroll, C., Lee, J., Mauritsen, T., Mehlmann, C., Mieslinger, T., Naumann, A. K., Paccini, L., Peinado, A., Praturi, D. S., Putrasahan, D., Rast, S., Riddick, T., Roeber, N., Schmidt, H., Schulzweida, U., Schütte, F., Segura, H., Shevchenko, R., Singh, V., Specht, M., Stephan, C. C., von Storch, J.-S., Vogel, R., Wengel, C., Winkler, M., Ziemen, F., Marotzke, J., and Stevens, B.: ICON-Sapphire: simulating the components of the Earth system and their interactions at kilometer and subkilometer scales, Geosci. Model Dev., 16, 779–811, https://doi.org/10.5194/gmd-16-779-2023, 2023. a, b
ICON Partnership (DWD, MPI-M, DKRZ, KIT, and C2SM): ICON release 2024.01, ICON Partnership [code], https://doi.org/10.35089/WDCC/IconRelease01, 2024. a, b, c
Kingma, D. P. and Ba, J.: Adam: A Method for Stochastic Optimization, 3rd International Conference on Learning Representations, ICLR2015, San Diego, CA, USA, 7–9 May 2015, Conference Track Proceedings, https://doi.org/10.48550/arXiv.1412.6980, 2017. a
Klöwer, M., Gelbrecht, M., Hotta, D., Willmert, J., Silvestri, S., Wagner, G. L., White, A., Hatfield, S., Kimpson, T., Constantinou, N. C., and Hill, C.: SpeedyWeather.jl: Reinventing atmospheric general circulation models towards interactivity and extensibility, Journal of Open Source Software, 9, 6323, https://doi.org/10.21105/joss.06323, 2024. a
Kochkov, D., Yuval, J., Langmore, I., Norgaard, P., Smith, J., Mooers, G., Klöwer, M., Lottes, J., Rasp, S., Düben, P., Hatfield, S., Battaglia, P., Sanchez-Gonzalez, A., Willson, M., Brenner, M. P., and Hoyer, S.: Neural general circulation models for weather and climate, Nature, 632, 1060–1066, https://doi.org/10.1038/s41586-024-07744-y, 2024. a
Krasnopolsky, V. M., Fox-Rabinovitz, M. S., and Chalikov, D. V.: New Approach to Calculation of Atmospheric Model Physics: Accurate and Fast Neural Network Emulation of Longwave Radiation in a Climate Model, Mon. Weather Rev., 133, 1370–1383, https://doi.org/10.1175/MWR2923.1, 2005. a
Lagerquist, R., Turner, D. D., Ebert-Uphoff, I., and Stewart, J. Q.: Estimating Full Longwave and Shortwave Radiative Transfer with Neural Networks of Varying Complexity, J. Atmos. Ocean. Tech., 40, 1407–1432, https://doi.org/10.1175/JTECH-D-23-0012.1, 2023. a
LeCun, Y. A., Bottou, L., Orr, G. B., and Müller, K.-R.: Efficient BackProp, Springer Berlin Heidelberg, p. 9–48, https://doi.org/10.1007/978-3-642-35289-8_3, 2012. a
Lohmann, U. and Roeckner, E.: Design and performance of a new cloud microphysics scheme developed for the ECHAM general circulation model, Clim. Dynam., 12, 557–572, https://doi.org/10.1007/bf00207939, 1996. a, b
Meyer, D., Hogan, R. J., Dueben, P. D., and Mason, S. L.: Machine Learning Emulation of 3D Cloud Radiative Effects, J. Adv. Model. Earth Sy., 14, e2021MS002550, https://doi.org/10.1029/2021MS002550, 2022. a
Pal, A., Mahajan, S., and Norman, M. R.: Using Deep Neural Networks as Cost-Effective Surrogate Models for Super-Parameterized E3SM Radiative Transfer, Geophys. Res. Lett., 46, 6069–6079, https://doi.org/10.1029/2018GL081646, 2019. a
Pincus, R., makepath LLC, and Sehnem, J. M.: pyRTE-RRTMGP, Zenodo [code], https://doi.org/10.5281/zenodo.16644555, 2025. a, b
Räisänen, P., Barker, H. W., Khairoutdinov, M. F., Li, J., and Randall, D. A.: Stochastic generation of subgrid-scale cloudy columns for large-scale models, Q. J. Roy. Meteor. Soc., 130, 2047–2067, https://doi.org/10.1256/qj.03.99, 2004. a, b, c
Rasp, S.: Coupled online learning as a way to tackle instabilities and biases in neural network parameterizations: general algorithms and Lorenz 96 case study (v1.0), Geosci. Model Dev., 13, 2185–2196, https://doi.org/10.5194/gmd-13-2185-2020, 2020. a
Roh, S. and Song, H.-J.: Evaluation of Neural Network Emulations for Radiation Parameterization in Cloud Resolving Model, Geophys. Res. Lett., 47, e2020GL089444, https://doi.org/10.1029/2020GL089444, 2020. a
Shamekh, S., Lamb, K. D., Huang, Y., and Gentine, P.: Implicit learning of convective organization explains precipitation stochasticity, P. Natl. Acad. Sci. USA, 120, https://doi.org/10.1073/pnas.2216158120, 2023. a
Stevens, B., Satoh, M., Auger, L., Biercamp, J., Bretherton, C. S., Chen, X., Düben, P., Judt, F., Khairoutdinov, M., Klocke, D., Kodama, C., Kornblueh, L., Lin, S.-J., Neumann, P., Putman, W. M., Röber, N., Shibuya, R., Vanniere, B., Vidale, P. L., Wedi, N., and Zhou, L.: DYAMOND: the DYnamics of the Atmospheric general circulation Modeled On Non-hydrostatic Domains, Progress in Earth and Planetary Science, 6, https://doi.org/10.1186/s40645-019-0304-z, 2019. a, b
Ukkonen, P.: Exploring Pathways to More Accurate Machine Learning Emulation of Atmospheric Radiative Transfer, J. Adv. Model. Earth Syst., 14, e2021MS002875, https://doi.org/10.1029/2021MS002875, 2022. a, b, c
Ukkonen, P. and Chantry, M.: Vertically Recurrent Neural Networks for Sub-Grid Parameterization, J. Adv. Model. Earth Sy., 17, e2024MS004833, https://doi.org/10.1029/2024MS004833, 2025. a, b, c, d
Ukkonen, P. and Hogan, R. J.: Twelve Times Faster yet Accurate: A New State-Of-The-Art in Radiation Schemes via Performance and Spectral Optimization, J. Adv. Model. Earth Sy., 16, https://doi.org/10.1029/2023ms003932, 2024. a
Watt-Meyer, O., Brenowitz, N. D., Clark, S. K., Henn, B., Kwa, A., McGibbon, J., Perkins, W. A., Harris, L., and Bretherton, C. S.: Neural Network Parameterization of Subgrid-Scale Physics From a Realistic Geography Global Storm-Resolving Simulation, J. Adv. Model. Earth Sy., 16, https://doi.org/10.1029/2023ms003668, 2024. a
Wood, R.: Stratocumulus Clouds, Mon. Weather Rev., 140, 2373–2423, https://doi.org/10.1175/mwr-d-11-00121.1, 2012. a
Yao, Y., Zhong, X., Zheng, Y., and Wang, Z.: A Physics-Incorporated Deep Learning Framework for Parameterization of Atmospheric Radiative Transfer, J. Adv. Model. Earth Sy., 15, e2022MS003445, https://doi.org/10.1029/2022MS003445, 2023. a, b, c
Zängl, G., Reinert, D., Rípodas, P., and Baldauf, M.: The ICON (ICOsahedral Non-hydrostatic) modelling framework of DWD and MPI-M: Description of the non-hydrostatic dynamical core, Q. J. Roy. Meteor. Soc., 141, 563–579, https://doi.org/10.1002/qj.2378, 2014. a
Zhao, Y., Li, J., Wen, D., Li, Y., Wang, Y., and Huang, J.: Distinct structure, radiative effects, and precipitation characteristics of deep convection systems in the Tibetan Plateau compared to the tropical Indian Ocean, Atmos. Chem. Phys., 24, 9435–9457, https://doi.org/10.5194/acp-24-9435-2024, 2024. a
Short summary
Most climate models cannot resolve clouds and cloud-radiation interactions at coarse horizontal resolutions of about 100 km, which introduces uncertainties. High-resolution models resolve clouds better but are expensive to run. We use short high-resolution simulations and artificial intelligence to learn the cloud-radiation interactions without making any assumptions about the small scales. We propose a new method that significantly reduces cloud related errors.
Most climate models cannot resolve clouds and cloud-radiation interactions at coarse horizontal...
