Skip Navigation
menu close modal

High resolution monitoring and forecasting of the Global Ocean

Author:

Elizabeth Rémy - (MERCATOR OCEAN)

Co-Investigator(s):

Mounir Benkiran (MERCATOR OCEAN)
Marie Drevillon (MERCATOR OCEAN)
Yann Drillet (MERCATOR OCEAN)
Mathieu Hamon (MERCATOR OCEAN)
Jean-Michel Lellouche (MERCATOR OCEAN)
Angélique Melet (MERCATOR OCEAN)
Giovanni Ruggiero (MERCATOR OCEAN)

Abstract:

Mercator Ocean is in charge of the Global Monitoring and Forecasting Centre (GLO MFC) for Copernicus Marine Environment Monitoring Service (CMEMS).

In this context, a daily operational production is performed to deliver near real time products including hindcasts and 10-day forecasts of the global ocean, making use of the most recent atmospheric forcing. In addition, multi-year time series covering at least the altimetry era (from 1992) are regularly produced. These ocean “reanalyses” are homogeneous gridded long time series of the ocean 3D state, which benefit from improved atmospheric forcing and observation data sets compared to the real time production.

Mercator Ocean‘s real time monitoring and forecasting system, hereafter referred to as GLO12, and the global ocean reanalysis system, called GLORYS12, are based on the same configuration at 1/12° horizontal resolution and 50 vertical levels. Both real time and reanalysis systems are constrained by data assimilation of along track altimetry observations (AVISO/CMEMS Sea Level TAC), satellite sea surface temperature maps, in situ temperature and salinity profiles (CORIOLIS/CORA/CMEMS IN Situ TAC) and satellite sea ice concentration (CERSAT or OSI-SAF). The multi year time series assimilating all available altimetry missions are extended every year to be as close as possible to the real time. This operational production is consolidated in CMEMS until 2021 and disseminated to users thanks to the CMEMS Central Information System (http://marine.copernicus.eu/).

In addition to the operational production, Research and Development activities are performed to improve the operational systems. The development plan for the next 4 years concerns:

  • The ocean model component towards an increased resolution, improved numerical schemes and parametrizations as well as the inclusion of new processes or coupling (wave, tide, …), addition of physical processes (as for example the tide) to gain in realism,
  • The data assimilation scheme that will allow to control higher frequency and smaller spatial scales, with high resolution observation data sets. A scheme based on ensemble method is currently being developed.
  • The assimilated observations taking into account new kind or new processing of observations when available. For the altimetry observations, the spatial and temporal filtering of along track observations and the different corrections applied to the observations (DAC, tide, inverse barometer …) have a large impact on the analysis. Different strategies are tested to assimilate them. The Mean Dynamic Topography (MDT) is also updated regularly based on the version released by CNES/CLS.

Those scientific evolutions of the different system components should lead to an increased accuracy of the analysis and forecasts with a better coherency with the assimilated in situ and satellite observations.

The discussion on impact of those evolutions on a better reconstruction of the ocean surface topography and improved benefit of along track altimetry data assimilation is the core of this proposal.

A monitoring of the quality of the production is performed routinely and metrics and diagnostics are available online. New metrics are regularly developed to quantify improvement of the main oceanic fields in comparison with observations for the analysis and/or for the forecast. Dedicated metrics to quantify the realism of some oceanic processes or structures as mesoscale phenomena are developed. They complement the classical data assimilation statistics based on bias, RMS misfit to observations. Today, correlation diagnostics but also spectral analysis and coherence are performed and will be complemented with front or eddy detection, particle displacement, etc. Those diagnostics help to quantify the sea surface height estimation accuracy but are also applied to currents and other quantities in the ocean analysis and forecasts produced in real time and in reanalysis. They are more oriented toward user need and physical content evaluation of the real time and delayed time production. A specific attention will be given to the estimation of the mesoscale features that are represented in the model forecasts.

We will also discuss the sensitivity of the operational system to altimetry constellation changes, with a focus on Jason and Sentinel series. This will be based on dedicated experiments and diagnostics.

Specific ocean circulation analysis, at regional scale, based on the operational production, will also be presented. Today, it concerns specifically the Malvinas current system and the projection at regional scale of the sea level rise, taking into account additional important processes in coastal zone such as waves, tides and atmospheric surges.

Supported by CNES