A global ocean data assimilation system has been developed for initializing coupled ocean-atmosphere general circulation models and analyzing climatic variability. It uses a variationally formulated optimal interpolation method to analyze oceanic fields from temperature observations and a first guess field provided by integrating a global ocean general circulation model (OGCM). The obsevations consist of the most comprehensive data dets [resently available, including weekly sea surface temperature fields from the Climte Prediction Center (Reynolds and Smith 1994) and temperature profiles from the World Ocean Atlas 1994 CD-ROM series (Levitus and Boyer 1994) and from the ANtional Centers for Environmental Prediction (NCEP). The OGCM is forced by the European Centre for Medium Range Weather Forecasts (ECMWF) surface wind stress and parameterized surface heat fluxes based on operational analyses.
Two retrospective analyses have been made using this assimilation system, one for 1986-1992 and the other for 1986-1989, which are referrec to as ASM1 and ASM2 respectively. The difference between these two analyses is that in ASM1 different time steps are used to integrate the tracer and momentum equations whereas ASM2 uses the same time step for both temperature and momentum equations. The time integration scheme for ASM1 is more computationally efficient. the two analyzed temperature have similar means and variability in their common period. Comparisons between ASM1 and the NCEP analysis also show good overall agreement in the analyzed temperature fields. These comparisons suggest that with the insertion of the observational data, these different analyses seem to converge to each other. Further comparison with a corresponding OGCM simulation shows that assimilating temperature observations into the OGCM seems to significantly reduce the systematic errors of the model mean state, such as the diffusing of the thermocline gradient and the excessive cooling of the equatorial ocean at the sea surface. IT also improves the annual cycle and interannual variability in the upper ocean, both on and off the equator.
The assimilation of hte ocean temperature also has profound effects on the model produced currents. Outside the equatorial wave guide, the improvements of the temperature field seem to improve the velocity fields. However, a comparison with independent in situ current meter measurements from TOGA TAO moorings shows that the assimilated currents near the equator have larger errors than those derived from driving the OGCM with observed surface forcings only. Moreover, these are substantial differences among ASM1, ASM2 and NCEP analyses. Some of the errors in ASM1 are caused by the split time step which distort the equatorial waves. Other problems, hawever, are from deficiencies in teh surface wind forcing that causes a large imbalance between the wind stress and the pressure gradient.
Complete copies of this report are available from: Center for Ocean-Land-Atmosphere Studies
last update: 27 March 1997
comments to: www@grads.iges.org