3. Decadal-scale variability in an OGCM experiment

To illustrate the effects of ocean dynamics on SSTs, we start with an analysis of output from an OGCM that is forced by the NCEP/NCAR reanalysis wind-stress for 40 years from 1958 to 1997. This OGCM simulation is unique in that the model SST and surface salinity are restored toward monthly climatology. Thus, we know for sure that anomalous atmospheric thermal forcing vanishes and all the SST variability in this model is caused by ocean dynamics. An analysis of this simulation can therefore shed light on the effects of ocean dynamics on SST variability. We ask the following questions here. Where does the ocean dynamics have strong effects on SST and by what mechanisms? Is this SST simulation realistic at all?

Fig. 1 Comparison of the model simulation (lower two panels) with observations [top four panels from Fig. 13 of Deser et al. (1996) ]. All are the area average in 140ｰE-180, 34ｰ-42ｰN.

The model results compare quite well observations. In the KOE region, the submixed layer (400-m depth) temperature anomalies were positive during the 1970s and negative during the 1980s in both the model output and observations. The vertical structure of temperature anomalies is largely in phase from surface to at least 400-m depth in the KOE region, suggsting that geostrophic advection acts not only to submixed layer temperature variabilty but also to SST variability.

Fig. 2 Rms of wintertime temperature variability with periods longer than five years at the model surface (solid lines in ｰC) and 200-m depth (color shade>0.3ｰC). The right panel displays the zonal mean rms for winter (white solid) and the summer (green dashed).

The regions with large rms in both the surface and submixed layers see the in-phase vartical thermal structure, and appear in the midlatitude western Pacific centered at 42ｰN and in the so-called subduction region to the southeast (Deser et al. 1996; Schneider et al. 1999; Nonaka et al. 2000). The corresponding large rms is due to only the wintertime variation and is concentrated in the two latitude bands 37ｰ-47ｰN and 30ｰ-35ｰN. The former rms is much larger than the latter one.

Fig. 3 First EOFs of (a) model decadal SST (66.7%) and (b) 26.5 kgm-3 isopycnal depth (26.5s) anomalies (71.1%), and (c) their time coefficients (green for SST and blue for isopycnal depth).

The spatial pattern in model DSV EOF is similar to that in the variance map (Fig. 2) exhibitting a westward-intensified structure (cf. Miller et al. 1998). There is another negative pole in the DSV EOF to the south that tilts northeastward slightly. In contrast to this SST dipole, the decadal 26.5s depth variability is characterized by a monopole pattern in the midlatitude western Pacific. The high correlation between the two time coefficients (0.96) supports that the anomalous large-scale gyre circualtion regulates the model DSV through geostrophic advection. In fact, the effect of Ekman advection is relatively small on the DSV compared with that of geostrophic advection in the KOE region (Qiu 2000; Xie et al. 2000). The anomalous eastward geostrophic currents in the midlatitude western Pacific maintain the positive SST anomalies in the KOE region, and the negative SST anomalies to the south are attributed to the weaker-than-normal Kuroshio Extension.

Fig. 4ﾅ@Wintertime mean mixed layer depth (color shade) and the rms of its decadal variability (contour in m).

The mixed layer is deep in the KOE region especially near the date line between 33ｰN and 42ｰN. Mode waters with small values of potential vorticity form on the edge of the KOE region with the deep mixed layer (Xie et al. 2000). The large rms of decadal variability is found in the northern and southeastern parts of the deep mixed layer region. Since the effect of geostrophic heat transport is propotional to the depth of mixed layer, the deep mixed layer allows geostrophic thermal advection to cause substantial SST variability in the KOE region. This explains why large SST rms appears where the mixed layer is deep in this model without anomalous atmospheric heat flux forcing.