This project has ended | -
As a Horizon2020 Centre of Excellence, ESiWACE has contributed significantly to strengthen the links between the different initiatives for weather and climate model development in Europe and to meet and discuss progress. ESiWACE has brought together the European community of weather and climate modellers to make our models fit for the next, heavy-parallelised generation of supercomputers. In particular, ESiWACE has improved scalability, usability and exploitability of high-resolution simulations of global atmosphere and ocean models.
As highlights, ESiWACE has developed configurations of the IFS and ICON models that can run global simulations at O(1 km) resolution, made significant progress on the development of a data middleware to take care of the processing of huge amounts of model output, and has supported the DYAMOND model inter-comparison project for global, storm-resolving simulations (https://www.esiwace.eu/services/dyamond).
ESiWACE has allowed ECMWF to explore scalability of IFS towards simulations with 1.45 km grid-spacing as well as the concurrent integration of the radiation scheme, the wave model and the rest of IFS within forecast simulations. Furthermore, the single precision version of IFS could be improved significantly as part of the work on the high-resolution demonstrator, and a method for the compression of ensemble model output has been investigated.
More details on the results and outcomes can be found on the project website https://www.esiwace.eu/ .
Numerical weather prediction and climate modelling are highly dependent on the available
computing power in terms of the achievable spatial resolution, the number of members run in
ensemble simulations as well as the completeness of physical processes that can be represented.
Both domains are also highly dependent on the ability to produce, store and analyse large amounts
of simulated data, often with time constraints from operational schedules or international
coordinated experiments. The ever increasing complexity of both numerical models and high
performance computing (HPC) systems has led to the situation that today, one major limiting factor
is no longer the theoretical peak performance of available HPC systems, but the relatively low
sustained efficiency that can be obtained with complex numerical models of the Earth system.
The differences in model complexity as well as the temporal and spatial scales that were
historically characteristic for climate and weather modelling are vanishing since both applications
ultimately require complex Earth system modelling capabilities which resolve the same physical
process detail across atmosphere, ocean, cryosphere and biosphere. With increasing compute
power and data handling needs, both communities must exploit synergies to tackle common
scientific and technical challenges. A joint climate and weather community engagement in research
and service provision is urgently required and timely given the substantial societal investment in
European wide infrastructures through the Copernicus services that are coordinated and managed
by the European Commission.
ESiWACE will deliver the required research and services through leveraging two established
European networks, namely (1) the European Network for Earth System modelling (ENES),
representing the European climate modelling community contributing to the internationally
coordinated experiments of the World Climate Research Program (WCRP) and the
Intergovernmental Panel on Climate Change (IPCC) assessments, and (2) the world leading
European Centre for Medium-Range Weather Forecasts (ECMWF), which is an independent
European organisation supported by 34 member and cooperating states, and also the operator of
the Copernicus services for atmospheric monitoring (CAMS) and climate change (CCCS).