Signal Energy and CO2 Footprint of Climate Model Intercomparison Projects

Summary

Earth System Models rely heavily on high-performance computing, and their computational demands keep growing due to finer spatial resolution, integration of complex biogeochemical processes, and larger ensembles used to manage uncertainty. Historically, growth in peak computing performance has outpaced improvements in energy efficiency, increasing total HPC power consumption, yet the environmental costs of Model Intercomparison Projects (MIPs) have received limited systematic attention. The paper traces the evolution of climate model carbon accounting from voluntary post-hoc estimation under CMIP6 to standardized accounting under the newly established CMIP7 Task Team on Energy Consumption. Using high-resolution Destination Earth simulations run on the MareNostrum 5 supercomputer, the authors empirically evaluate how different accounting boundaries—operational, active-only, and embodied carbon—affect reported energy use, carbon emissions and financial costs. The paper concludes by outlining methodological considerations for standardizing energy and carbon accounting across future Model Intercomparison Projects.