DOE OSTI · 3395280
Crowdsourcing the Frontier: Advancing Hybrid Physics‐ML Climate Simulation via a $\$$50,000 Kaggle Competition
Abstract
Subgrid machine-learning (machine learning [ML]) parameterizations have the potential to introduce a new generation of climate models that incorporate the effects of higher-resolution physics without incurring the prohibitive computational cost associated with more explicit physics-based simulations. However, important issues, ranging from online instability to inconsistent online performance, have limited their operational use for long-term climate projections. To more rapidly drive progress in solving these issues, domain scientists and ML researchers opened up the offline aspect of this problem to the broader ML and data science community with the release of ClimSim, a NeurIPS Data sets and Benchmarks publication, and an associated Kaggle competition. This paper reports on the downstream results of the Kaggle competition by coupling emulators inspired by the winning teams' architectures to an interactive climate model (including full cloud microphysics, a regime historically prone to online instability) and systematically evaluating their online performance. Our results demonstrate that online stability in the low-resolution real-geography setting is reproducible across multiple diverse architectures, which we consider a key milestone. All tested architectures exhibit strikingly similar offline and online biases, though their responses to architecture-agnostic design choices (e.g., expanding the list of input variables) can differ significantly. Multiple Kaggle-inspired architectures achieve state-of-the-art results on certain metrics such as zonal mean bias patterns and global Root Mean Squared Error, indicating that crowdsourcing the essence of the offline problem is one path to improving online performance in hybrid physics-AI climate simulation.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Lin, Jerry [University of California, Irvine, CA (United States); Boston University, MA (United States)] (ORCID:0000000317789426), Hu, Zeyuan [NVIDIA Corporation, Santa Clara, CA (United States)] (ORCID:000000032041879X), Beucler, Tom [University of Lausanne (Switzerland)] (ORCID:0000000257311040), Frields, Katherine [University of California, Irvine, CA (United States)], Christensen, Hannah [University of Oxford (United Kingdom)] (ORCID:0000000182440218), Hannah, Walter [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)] (ORCID:0000000248111928), Heuer, Helge [German Aerospace Center (DLR), Wessling (Germany). Institute of Atmospheric Physics (IPA)] (ORCID:0000000324117150), Ukkonen, Peter [University of Oxford (United Kingdom)] (ORCID:0000000185658079), Mansfield, Laura A. [University of Oxford (United Kingdom)] (ORCID:0000000262856045), Zheng, Tian [Columbia University, New York, NY (United States)] (ORCID:0000000348890391), Peng, Liran [University of California, Irvine, CA (United States)] (ORCID:0000000329935208), Gupta, Ritwik [University of California, Berkeley, CA (United States); University of Maryland, College Park, MD (United States)], Gentine, Pierre [Columbia University, New York, NY (United States)] (ORCID:0000000208458345), Al‐Naher, Yusef [Unaffiliated, Amsterdam (Netherlands)], Duan, Mingjiang [Z Lab, Hangzhou (China)], Hattori, Kyo [ABEJA Inc., Nagoya (Japan)], Ji, Weiliang [Z Lab, Hangzhou (China)], Li, Chunhan [Z Lab, Hangzhou (China)], Matsuda, Kippei [Kawasaki Heavy Industries, Ltd., Kobe (Japan)], Murakami, Naoki [DeNA Co., Ltd., Tokyo (Japan)], Ron, Shlomo [Unaffiliated, Ramat Gan (Israel)], Serlin, Marec [Uber Technologies, Inc., Seattle, WA (United States)], Song, Hongjian [Z Lab, Hangzhou (China)], Tanabe, Yuma [Unaffiliated, Kyoto (Japan)], Yamamoto, Daisuke [Unaffiliated, Nara (Japan)], Zhou, Jianyao [Unaffiliated, Melbourne, VIC (Australia)], Pritchard, Michael [University of California, Irvine, CA (United States); NVIDIA Corporation, Santa Clara, CA (United States)] (ORCID:0000000203406327). 2026-05-13. Crowdsourcing the Frontier: Advancing Hybrid Physics‐ML Climate Simulation via a $\$$50,000 Kaggle Competition. https://doi.org/10.1029/2025ms005643
Cite the original work for its findings. Save a collection to share your selection of sources.