DECADAL AND ENSO-SCALE THERMOCLINE VARIATIONS IN THE NORTH PACIFIC OCEAN
Arthur J. Miller
Scripps Institution of Oceanography
La Jolla CA 92093-0224
Abstract. Temperature variations which occurred in the North Pacific thermocline (nominally 400m depth) during the 1970s and 1980s are described in both a numerical simulation and XBT observations, We identify, in both model and observations, two distinct midlatitude variations, (i) a decadal change in the thermocline which is western intensified and (ii) westward propagating thermocline anomalies with ENSO time scales. The decadal change is associated with a raised thermocline off the coast of Japan and a deepened thermocline off the coast of California after 1976-77. The ENSO scale variations are associated with shorter-lived thermocline changes which originate off the coast of California and propagate at least to 170W between 30-40N, and much further westward at lower latitudes. The decadal change is driven by a basin-scale change in wind stress curl and the ENSO scale thermocline anomalies appear to be at least partly forced by large-scale midlatitude wind stress curl variations as well. The observed and modeled ENSO-scale anomalies are distinct from the baroclinic Rossby waves radiated from coastally trapped Kelvin-like waves previously identified in high-resolution numerical models because they have larger wavelengths and appear to he wind forced.
1. Introduction
Decadal scale changes have been observed in numerous physical, chemical and biological variables in the North Pacific (e.g., Douglas et al., 1982; Ebbesmeyer, et al., 1991; Roemmich and McGowan, 1995). One especially interesting change occurred in the mid-1970s (e.g., Trenberth, 1990; Graham, 1994; Trenberth and Hurrell 1994) for which, in previous work (Miller et al., 1994a,b), we attempted to explain the oceanic physics involved in switching the ocean from a warm central (cool eastern) North Pacific state to an oppositely signed regime after the winter of 1976-77.
In this study, we take a look underneath the mixed layer to identify (a) what changes were observed to occur in the oceanic thermocline and (b) what controls these changes. Towards that end, we analyze a dataset of upper-ocean (to 400m depth) XBT observations from the period 1970-88 (see White and Cayan, 1996, for details) and compare it with the response of an ocean general circulation model forced by observed fluxes and winds over the same period. By identifying what signals are common to both the model and observatious, we are able to have more faith in the data (which is sparsely sampled in space and time) and to interpret the variations with the model output which is dynamically consistent with the input forcing.
Figure 1 shows differences from the early 1970s to the early 1980s of the upper ocean temperature field of the North Pacific in both the observations and in the model. (The time differencing interval is motivated by results to be presented in subsequent sections.) In the surface mixed layer, one sees the well known structure of the cooling in the central Pacific and the warming of the eastern Pacific. This structure is commensurate with the large-scale structure of the atmospheric variables (heat fluxes and wind stresses) which drive the variability (Cayan, 1992; Miller et al., 1994a.b). The time series of SST variability in a rectangular region in the central North Pacific (180W-150W, 30N-40N) reveals a good correspondence between model (dotted) and observations (dashed) and shows the step-like character of the SST change during the winter of 1976-77.
Looking beneath the mixed layer at 200m and 400m (Figure 1), where little or no direct contact with the atmosphere occurs, one sees a progression into region where ocean dynamics are expected to dominate the response. Indeed the thermal structures at 400m bear a western intensified structure reminiscent of gyre-scale circulation theory. The variability at 400m depth is much weaker and, though the model does not capture the higher-frequency variability, both model and data reveal more gradual change in temperature from the 1970s to the 1980s in contrast to the more step-like change in the surface layer.
