Energy-Aware Earth System Science: A Community Roadmap
Abstract
Earth system models underpin the evidence that informs net-zero policy, yet their energy demands are rarely considered when judging model quality. Using archived output from 159 models spanning five generations of the Coupled Model Intercomparison Project (CMIP), we show that model error has fallen at a constant rate of 13.5 per cent per doubling of compute since 1995. Halving CMIP6 error would require 28 times more compute. Improvements are uneven, with sea-level pressure and surface air temperature errors declining 3.5-fold and 2.7-fold since CMIP1, but precipitation error by only 7 per cent. Machine-learning weather models match or exceed operational forecast skill at a fraction of the runtime, but none publishes measured energy, and including training changes how they rank. We propose a skill-energy framework that combines an empirical Pareto front with a ranking score. These findings argue for standardised reporting of energy use and performance ahead of CMIP8.
Related
- How to Use Score-Based Diffusion in Earth System Science: A Satellite Nowcasting Example
- Engagement in food system-related controversies can strengthen science–policy–society interfaces
- From mercury to plastics: Earth system science and governance strategies
- A Rapid Projection of Water Carrying Capacity in Areas with Community Based Clean Water Supply System in Semarang City
- Uncertainty-aware bilevel energy storage planning for active distribution networks: A fixed-topology DistFlow-based MISOCP framework with voltage-security assessment
BibTeX
@misc{colfescu2026energyaware,
title = {{Energy-Aware Earth System Science: A Community Roadmap}},
author = {Ioana Colfescu and Annalisa Bracco and Kara Moraw and Chen Lu and Michael Bareford and Tom Beucler and Massimo Bollasina and Juan Carlos Bilbao-Ludena and Hannah Christensen and Serafina DiGioia and Gregory J. Hakim and Thomas Hauser and Adrian Jackson and Phong Le and Ilenia Manco and Stephen Mobbs and Amirpasha Mozaffari and Antonio Navarra and Mark Parsons and James W. B. Rae and Thomas Reichler and Carlos Rodriguez-Pardo and Graham Stacey and Michèle Weiland},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.23216},
eprint = {2610.23216},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.23216},
}
RIS
TY - UNPB
TI - Energy-Aware Earth System Science: A Community Roadmap
AU - Ioana Colfescu
AU - Annalisa Bracco
AU - Kara Moraw
AU - Chen Lu
AU - Michael Bareford
AU - Tom Beucler
AU - Massimo Bollasina
AU - Juan Carlos Bilbao-Ludena
AU - Hannah Christensen
AU - Serafina DiGioia
AU - Gregory J. Hakim
AU - Thomas Hauser
AU - Adrian Jackson
AU - Phong Le
AU - Ilenia Manco
AU - Stephen Mobbs
AU - Amirpasha Mozaffari
AU - Antonio Navarra
AU - Mark Parsons
AU - James W. B. Rae
AU - Thomas Reichler
AU - Carlos Rodriguez-Pardo
AU - Graham Stacey
AU - Michèle Weiland
PY - 2026
DA - 2026/10/16
PB - arXiv
DO - 10.48550/arxiv.2610.23216
UR - https://arxiv.org/abs/2610.23216
AB - Earth system models underpin the evidence that informs net-zero policy, yet their energy demands are rarely considered when judging model quality. Using archived output from 159 models spanning five generations of the Coupled Model Intercomparison Project (CMIP), we show that model error has fallen at a constant rate of 13.5 per cent per doubling of compute since 1995. Halving CMIP6 error would require 28 times more compute. Improvements are uneven, with sea-level pressure and surface air temperature errors declining 3.5-fold and 2.7-fold since CMIP1, but precipitation error by only 7 per cent. Machine-learning weather models match or exceed operational forecast skill at a fraction of the runtime, but none publishes measured energy, and including training changes how they rank. We propose a skill-energy framework that combines an empirical Pareto front with a ranking score. These findings argue for standardised reporting of energy use and performance ahead of CMIP8.
ER -