Super parameterization of ocean dynamics for tracer transport models

Clement Bricaud1, Julien Le Sommer2, Gurvan Madec3, Christophe Calone4, Jerome Chanut5, Christian Ethe6 and Coralie Perruche1, (1)Mercator-Ocean, Ramonville Saint-Agne, France, (2)CNRS, IGE, Grenoble, France, (3)Sorbonne Universités (UPMC, Univ Paris 06)-CNRS-IRD-MNHN, LOCEAN Laboratory, IPSL, Paris, France, (4)CNRS, LGGE, Grenoble, France, (5)Mercator Ocean, Ramonville St Agne, France, (6)Institut Pierre Simon Laplace, LOCEAN/CNRS, Paris, France
Abstract:
Ocean mesoscale and submesoscale turbulence contribute to ocean tracer transport and to shaping ocean biogeochemical tracers distribution. Representing adequately tracer transport in ocean models therefore requires to increase model resolution so that the impact of ocean turbulence is adequately accounted for. But due to supercomputers power and storage limitations, global biogeochemical models are not yet run routinely at eddying resolution. Still, because the ”effective resolution” of eddying ocean models is much coarser than the physical model grid resolution, tracer transport can be reconstructed to a large extent by computing tracer transport and diffusion with a model grid resolution close to the effective resolution of the physical model. This observation has motivated the implementation of a new capability in NEMO ocean model (http://www.nemo-ocean.eu/) that allows to run the physical model and the tracer transport model at different grid resolutions. Here, we present results obtained with this new capability applied to a synthetic age tracer in a global eddying model configuration. In this model configuration, ocean dynamic is computed at ¼° resolution but tracer transport is computed at 3/4° resolution. The solution obtained is compared to a reference setup, where age tracer is computed at the same grid resolution as ocean dynamics. We discuss possible options for defining the vertical diffusivity coefficient for the tracer transport model based on information from the high resolution grid. We describe the impact of this choice on the distribution and one the penetration of the age tracer. The method described here can found applications in ocean forecasting, such as the Copernicus Marine service operated by Mercator-Ocean, and in Earth System Models for climate applications.