The Climate Modeling Branch (CMB) of the National Centers for Environmental Prediction (NCEP) routinely produces weekly ocean analyses using a data assimilation system based on an oceanic general circulation model. At present the system assimilates only temperature data, primarily satellite estimates of SST and vertical profiles from XBTs. Further details of the system can be found in the December 1994 issue of ACCP Notes and in the literature (Ji et. al., 1994 and Derber and Rosati, 1989). A comparison of the Atlantic system with TOPEX altimetry data has also been published (Behringer, 1994).
Figure 1 shows anomalous SST from the model averaged for March 1997 though May of 1997. The SST climatology is due to Reynolds (personal communication) and is for the period 1982-92. Since the December 1996 through February 1997 period (ACCP Notes, Vol. IV, No.1) the South Atlantic has generally cooled while the eastern margin of the North Atlantic has warmed. In the southern hemisphere an earlier warm patch south of 20° S and west of 10° W, has disappeared and an anomalous cool patch has appeared between 20° S and 30° S. In the North Atlantic, negative anomalies of 1° C, or more, which had extended along the model Gulf Stream in DJF, have decreased in area and appear to be embedded in patchy warm anomalies of about l° C. Much of the eastern margin of the North Atlantic, north of 5° N and east of 20° W, is anomalously warm by about l° C. In the figure, anomalies with magnitudes greater than 1° C are shaded. Given the abundance of satellite SST data, the assimilation of these data will strongly influence the model SST, masking inadequacies in the surface heat flux and model dynamics. The patterns shown here strongly resemble Reynolds' SST analyses (see e.g. Climate Diagnostics Bulletin) which are based on the same data.
Figure 2 shows dynamic height and anomalous dynamic height at the surface relative to 1000m averaged for March 1997 though May of 1997. The anomalous dynamic height is relative to a short climatology based on three years, 1993-5, of data from this analysis system. Large anomalies of about 10 dyn. cm. which appeared in the center and along the southern edge of the North Atlantic subtropical gyre during the previous period (DJF) have persisted into the present period. A comparison of model output with TOPEX sea surface height anomalies averaged over l0° x l0° squares (Behringer, 1994) shows that rms differences between the model and TOPEX are about 10 cm along western boundaries and about 3 cm in the tropics. If these rms differences are regarded as rough estimates of the error in model dynamic height, then only the largest DJF anomalies in the North Atlantic subtropical gyre exceed the uncertainty in the model.
Derber, J. and A. Rosati. 1989. A global oceanic assimilation system. J. Phys. oceanogr., 19: 1,333-1,347, 1989.
Ji, M., A. Leetmaa and J. Derber. 1994. An Ocean Analysis System for Seasonal to Interannual Climate Studies. Mon. Wea. Rev. 123: 46O-481.