High-resolution ocean and climate modelling on unstructured grids: from finite elements to finite volumes

This talk will give an overview about the ocean model developments at the Alfred Wegener Institute (AWI), Helmholtz Centre for Polar and Marine Research, and summarise example applications made possible by using a novel numerical core. The Finite Element Sea Ice-Ocean Model (FESOM1.4) supports multi...

Full description

Bibliographic Details
Main Author: Rackow, Thomas
Format: Conference Object
Language:unknown
Published: 2018
Subjects:
Online Access:https://epic.awi.de/id/eprint/53596/
https://epic.awi.de/id/eprint/53596/1/Overview_FESOM2_ECMWF.pdf
https://hdl.handle.net/10013/epic.0c3524a9-6df7-418e-bf8b-a53170d1b4b3
https://hdl.handle.net/
Description
Summary:This talk will give an overview about the ocean model developments at the Alfred Wegener Institute (AWI), Helmholtz Centre for Polar and Marine Research, and summarise example applications made possible by using a novel numerical core. The Finite Element Sea Ice-Ocean Model (FESOM1.4) supports multi-resolution triangular grids and has already been used in many studies in the last decade, with a focus on Arctic and Antarctic sea ice, global ocean circulation, and the impact of locally eddy-resolving resolution. In a coupled climate configuration with ECHAM6, the model now also participates in the 6th phase of the Coupled Model Intercomparison Project (CMIP) and in a number of endorsed MIPs like HighResMIP. The change of the dynamical core to a finite volume formulation speeds up the successor model FESOM2 by a factor of 3. In combination with its nearly linear scalability characteristics that retain a short time-to-solution even for extremely large ocean grids, this extends the range of possible applications dramatically. A recent accomplishment is a prototype single-executable coupled configuration of IFS-spectral and FESOM2. This configuration will allow to test the sensitivity of ECMWF’s medium-to-seasonal-range forecasts to the choice of the ocean model formulation, both in terms of computational scalability and possible added value due to locally eddy-resolving surface fields.