On the role of GRACE for the joint assimilation of altimetry and in-situ data
Frederic Castruccio(1) , Jacques Verron(1) , Lionel Gourdeau(2) , Jean-Michel Brankart(1) , and Pierre Basseur(1)
LEGI - CNRS,
38041 Grenoble cedex9,
(2) IRD, BPA5, 98848 Noumea, New Caledonia
The measured altimetric signal, i.e. the Sea Surface Height (SSH) can only be used in oceanography in its “residual” component, i.e. the Sea Surface Anomaly (SLA). Because of geoid uncertainties, altimetric applications have indeed concentrated on ocean variability using SLA. The lack of an adequate mean dynamic topography (MDT) is a recurrent issue on physical oceanography and especially on altimetric data assimilation. The recent advances of our knowledge of the geoid brought by the satellite gravity mission such as CHAMP launched in 2000 and the GRACE experiment launched in 2002 provides a more accurate geoid reference and therefore the possibility to exploit the absolute sea surface height measurements in physical oceanography studies.
In this study, we investigate the impact of the absolute altimetric signal in numerical model via data assimilation method. Assimilation experiments are performed with the OPA 8.2 general circulation model in its ORCA configuration (free surface global 2° low resolution with a meridional grid spacing refinement down to 0.5° in tropical region to improve the equatorial dynamics). The assimilation scheme is a reduced-order sequential Kalman filter (SEEK filter). To avoid the shocks that are induced into the model by intermittent analysis, an Incremental Analysis Update (IAU) scheme is added to the regular SEEK filter procedure.
The contribution brought by the use of an observed absolute MDT is illustrated by multi-annual experiments jointly assimilating altimetric SSH (from TOPEX/POSEIDON and ERS1&2) referenced to the GRACE geoid and the in situ TAO temperature profiles. The focus is made on the tropical Pacific region, which is the region where data assimilation is performed within the global model. The objective of the study is to assess how the assimilation of the GRACE geoid into a realistic model can improve the modelling of the Tropical Pacific dynamics and what impact it may have on the multiple data assimilation process. The solution is evaluated with regards to our present knowledge of the Pacific tropical dynamics as given by various data bases. In particular, independent XBT profiles coming from the CORIOLIS project data base are used in order to validate the thermal structure of the assimilated run.
The first results are promising, especially in term of compatibility between in situ and satellite data observations. It is clear that by providing an observed absolute reference for altimeter data, GRACE data make possible the assimilation of multiple data sets such as SSH and in situ temperature profiles. The use of an absolute dynamic topography opens the ways for the development of efficient assimilation systems based on multiple data sets using altimetry and in situ data.