Project description
Improved Earth system modelling and climate prediction considering ocean mesoscale eddies
Climate is shaped by numerous processes, not only the greenhouse gases emitted into the atmosphere as a result of human activity. Earth system models simulate physical, chemical and biological processes and how these work together, integrating the interactions of atmosphere, ocean, land, ice and biosphere to estimate climate under various conditions. Until now, these models have omitted ocean mesoscale eddies, i.e. time-dependent swirls with distance scales of under 100 kilometres and time scales of around a month. The ambitious EU-funded EERIE project will develop improved Earth system models, harnessing Europe’s pre-exascale computers and considering European ocean mesoscale eddies in preparation for the next Intergovernmental Panel on Climate Change report.
Objective
Eddy Rich Earth System Models (EERIE) will develop a new generation of Earth System Models (ESMs) that are capable of explicitly representing a crucially important, yet unexplored regime of the Earth system, the ocean mesoscale. Leveraging the latest advances in science and technology, EERIE will substantially improve the ability of such ESMs to faithfully represent the centennial-scale evolution of the global climate, especially its variability, extremes and how tipping points may unfold under the influence of the ocean mesoscale. Model improvements include new dynamical cores, new components (particularly sea ice), scale-aware parametrization and the complementary use of Machine Learning (ML)
The technological challenge associated with this ambition is very high. EERIE’s goal is to achieve a simulation speed of up to 5 simulated years per day (5 SYPDs) and to make efficient use (reduction in power consumption by 50%) of the pre-exascale supercomputers now available in Europe. The technological solutions that are to be leveraged in EERIE are the use of reduced precision, GPUs, ML and reduced I/O. Alongside model improvements, EERIE will develop innovative experimental simulation protocols that are suitable for the mesoscale, to be pioneered on behalf of the global climate modelling community, in preparation for the next IPCC.
EERIE will produce useful and usable climate information that will contribute to national and international climate change assessments such as IPCC; it will incorporate model variability and extremes within an Integrated Assessment Model (IAM) via a MLemulato;, and it will deliver storyline approaches to the consequences of climate tipping points and disease outbreaks.
Fields of science
- natural sciencesmathematicsapplied mathematicsdynamical systems
- natural sciencesmathematicsapplied mathematicsstatistics and probability
- natural sciencesearth and related environmental sciencesoceanography
- natural sciencescomputer and information sciencesartificial intelligencemachine learning
- natural sciencescomputer and information sciencessoftwaresoftware applicationssimulation software
Keywords
Programme(s)
Funding Scheme
HORIZON-RIA - HORIZON Research and Innovation ActionsCoordinator
27570 Bremerhaven
Germany