Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated 8 min ago
PreprintCC BY 4.0arXiv

EC-EarthFlow: Probabilistic emulation of daily transient global climate model simulations with flow matching

  • Kirien Whan
  • Nikolaj T. Mücke
  • Karin van der Wiel
Open at sourcePDFDOICite

Abstract

We introduce EC-EarthFlow, a generative flow matching model that emulates simulations from the physical climate model EC-Earth3. The model is trained on transient simulations from EC-Earth3 (1950-2166, SSP2-4.5) to predict the day ahead temperature field from the previous days temperature as well as annual mean temperature. Predictions are made auto-regressively with rollout periods of between a month and an extended season. Using only this variable of interest, we are able to reproduce the daily variability, spatial patterns, annual cycle and long-term trend from EC-Earth3 at a substantially lower computational cost than the physical model. We demonstrate that EC-EarthFlow is stable for long inference periods, and that it can learn the physical relationships as simulated in EC-Earth3.

Related

  1. GLAMODAT (v1.0): a model-data comparison tool for assessing transient palaeoglacier simulations
    Andreas Henz, Sarah Kamleitner, Samuel U. Nussbaumer +3
    EGUspherePreprintCC BY 4.0General earth scienceEGUsphere
  2. Scenario-conditioned flow matching for probabilistic generation of three-component ground-motion waveforms
    Yi Ding, Jinjun Hu, Su Chen +5
    arXivPreprintCC BY 4.0Full textGeophysics & seismologyarXiv
  3. Dynamical-Model-Informed Stochastic Flow Matching for Global Probabilistic Subseasonal Forecasting of Means and Extremes
    Abdullah A. Fahad, Manmeet Singh
    arXivPreprintCC BY 4.0Full textAtmospheric science & meteorologyStatisticsarXiv
  4. The calibration deficit: a challenge for Earth system simulation and emulation
    Peter M. Cox, Dale Durran, Veronika Eyring +4
    EGUspherePreprintCC BY 4.0General earth scienceEGUsphere
  5. Variable Masking for Climate Emulation in ACE
    Alexey Yermakov, Troy Arcomano, Jeremy McGibbon +2
    arXivPreprintCC BY 4.0Full textAtmospheric science & meteorologyarXiv

Cite

BibTeXRIS
BibTeX
@misc{whan2026ecearthflow,
  title = {{EC-EarthFlow: Probabilistic emulation of daily transient global climate model simulations with flow matching}},
  author = {Kirien Whan and Nikolaj T. Mücke and Karin van der Wiel},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.09715},
  eprint = {2610.09715},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.09715},
  note = {CC BY 4.0},
}
RIS
TY  - UNPB
TI  - EC-EarthFlow: Probabilistic emulation of daily transient global climate model simulations with flow matching
AU  - Kirien Whan
AU  - Nikolaj T. Mücke
AU  - Karin van der Wiel
PY  - 2026
DA  - 2026/10/07
PB  - arXiv
DO  - 10.48550/arxiv.2610.09715
UR  - https://arxiv.org/abs/2610.09715
AB  - We introduce EC-EarthFlow, a generative flow matching model that emulates simulations from the physical climate model EC-Earth3. The model is trained on transient simulations from EC-Earth3 (1950-2166, SSP2-4.5) to predict the day ahead temperature field from the previous days temperature as well as annual mean temperature. Predictions are made auto-regressively with rollout periods of between a month and an extended season. Using only this variable of interest, we are able to reproduce the daily variability, spatial patterns, annual cycle and long-term trend from EC-Earth3 at a substantially lower computational cost than the physical model. We demonstrate that EC-EarthFlow is stable for long inference periods, and that it can learn the physical relationships as simulated in EC-Earth3.
ER  -