hindcast of which we are aware that uses both observed total heat-flux and wind-stress anomalies as forcing for such a long time interval. There is close agreement between the model SSTs and those observed in many regions of the Pacific, including the tropics and the northern extratropics. Besides performing credibly on the monthly time scale, the model captures the essence of low-frequency variability over the North Pacific, including aspects of a marked basin-wide change that occurred in 1976-1977. In the model's detailed heat budget, the anomalous air-sea heat fluxes, entrainment, and to a lesser extent horizontal advection, force thermal-anomaly changes in the mixed layer. Each of these components was apparently involved in the 1976-1977 decadal SST shift.
Our purpose is to describe monthly to decadal variations in surface variability over the Pacific basin. The results presented combine a historical set of monthly marine observations and a two-decade numerical simulation (Miller et al., 1994a,b) that uses the Oberhuber (1993) isopycnic ocean general-circulation model.
The first part of the paper is an analysis of observed monthly surface marine atmospheric variability in relation to sea surface temperature (SST) over the Pacific Ocean. This observational evidence provides background to the second section of the paper, in which monthly wind stresses and heat fluxes are used to force an ocean general-circulation model (OGCM). In the observational section, we concentrate on the variability of the latent and sensible heat fluxes, since they provide a good example of the characteristics and problems involved in bulk formulations using the marine data. The fluxes exhibit monthly variability with large-scale organization, and their effect on the ocean can be detected in the monthly fluctuations of the SST anomaly field.
The second part of the paper reports on applying the marine data parameterizations to a simulation over two decades (1970-1988) of the Pacific Ocean basin using the OPYC OGCM. This run is forced by monthly mean wind stress and fluxes derived from surface marine observations, which are introduced in the first section. In assessing this run, we compare the model SST anomalies with observations, particularly their low-frequency variability. Further insight is provided by the model upper-ocean heat budget, which is not well described by observations.
Aside from diagnosing mechanisms causing monthly SST variability, one motivation for this simulation was to better understand a strong regime-like change in SST that occurred in the North Pacific basin during the mid-1970s (Douglas et al., 1982; Nitta and Yamada, 1989; Trenberth, 1990; Miller et al., 1994a,b; Graham, 1994; Trenberth and Hurrell, 1995, in this volume). Although thorough observational documentation of decadal-scale variability is lacking, glimpses from recent historical episodes suggest that this "gray area" (Karl, 1988) of the variability spectrum contains important climate effects. The actual shift was identified in fall and winter of 1976-1977 when SSTs in the central North Pacific cooled markedly and SSTs along the west coast of North America warmed (Venrick et al., 1987; Trenberth, 1990; Ebbesmeyer et al., 1991). In the atmosphere, the shift involved a basin-scale deepening of the wintertime Aleutian Low System, and appears to have been at least partially instigated by forcing from the tropical Pacific (Graham, 1994; Trenberth, 1995, in this volume), although an alternative theory involves sea-air feedback and advection of thermal anomalies by the North Pacific subtropical gyre (Latif and Barnett, 1994). Accompanying changes in many other physical and biological variables in the North Pacific basin and around its margin were also noted (Venrick et al., 1987; Ebbesmeyer et al., 1991).
The bulk formulae parameterized heat fluxes, calculated from standard marine surface meteorological observations, are the only means of estimating a multi-year time history of the heat exchange over broad regions of the oceans. Similarly, SST has been measured routinely by merchant ships for several decades, and, in the absence of a comprehensive temperature-versus-depth set in the upper ocean, SST is used to infer the variability of the upper-ocean heat content. Although the ocean heat-content structure can be complex, during winter in the extratropics the upper ocean is quite well mixed and SST is a good indicator (White and Walker, 1974).
As is brought out in the section below, there is a rich variability, superimposed upon the climatological mean, that involves both atmospheric and oceanic fluctuations. Three tests of the surface heat budget are discussed here. The first two, which study the data directly, are to relate monthly anomalies of the latent and sensible fluxes to the anomalous atmospheric circulation and to anomalous tendencies of the SST field. The third is to incorporate the monthly bulk-formula wind stress and heat flux as forcing of an extended OGCM run and to compare the simulated versus the observed SST fields. These results represent the first hindcast of which we are aware that uses both observed total-heat-flux and wind-stress anomalies as forcing for such a long time interval.
The primary data employed in the observational section of this study, as well as in the model forcing, are gridded