2. MODEL
We use the Widely used OGCM (GFDL MOM 1.1; Pacanowski et, al. 1991). The moclel covers the portion of the Paciflc from 50°S to 6O°N, and has realistic coastline and bottom topography. The horizontaland vertical eddy viscosities are constant at 5.0x107 and 10.0cm2s-1, respectively. Tracers are mixed both along isopycnal surfaces and diapycnally with diffusivities of 2.0×107 and 0.3cm2s-1, respectively. At the sea surfac.e, the climatological wind stress (Hellerman and Rosenstein 1983) is applied and temperature and salinity are restored to the climat, ological value (Levitus and Boyer 1994; Levitus et al. 1994) with seasonal variation. The horizontal resolution is 1 degree, and the vertical resolutions are 10 to 15m in the top 300m. The model is integrated for 31 years and the resultaveraged for the last year is used in the following passive tracer calculation. See Nonaka (1998) for more details.
To examine how the SST anomalies (Fig. 1) spread along the thermocline, the passive tracer corresponding to the temperature anomaly is injected to the model ocean, by restoring its concentration at the most upper level to the observed SST anomalies north of 20｡?. The restoring time is 14 days, the same as for temperature. As We are concerned with a tropical signature of SST anomalies injected into the thermocline in the mid-latitudes, the surface tracer field is restored to zero south of 20｡?. To decide the isopycnal and diapycnal mixing direction. the stratification is fixed at its March 1 values, when the mixed layer is deepest, and SST anomalies are subducted into the thermocline in theNorth Pacific. This tracer is advected by the annual mean velocity field. Initially, tracer has uniform zero fleld. The tracer field is in a nearly steady state after 30 years integration. The results at the 30th year are analyzed.
3. RESULTS
Fig. 2 shows that negative tracers subducted at the center of the oval-shaped SST anomalies extend southrvestward like a tongue to the western boundary, following the water parcel trajectories. The majority of negative tracers do not go to the Tropics but recirculate within the subtropical gyre. Only tracers on the southern flank of the negative tracer tongue are advected into the equatorial region through both the interior and the western boundary exchange routes as revealed by the trajectories.
