Aj. Miller et al., INTERDECADAL VARIABILITY OF THE PACIFIC-OCEAN - MODEL RESPONSE TO OBSERVED HEAT-FLUX AND WIND STRESS ANOMALIES, Climate dynamics, 9(6), 1994, pp. 287-302
Variability of the Pacific Ocean is examined in numerical simulations
with an ocean general circulation model forced by observed anomalies o
f surface heat flux, wind stress and turbulent kinetic energy (TKE) ov
er the period 1970-88. The model captures the 1976-77 winter time clim
ate shift in sea surface temperature, as well as its monthly, seasonal
and longer term variability as evidenced in regional time series and
empirical orthogonal function analyses. Examination of the surface mix
ed-layer heat budget reveals that the 1976-77 shift was caused by a un
ique concurrance of sustained heat flux input anomalies and very stron
g horizontal advection anomalies during a multi-month period preceding
the shift in both the central Pacific region (where cooling occurred)
and the California coastal region (where warming occurred). In the ce
ntral Pacific, the warm conditions preceding and the cold conditions f
ollowing the shift tend to be maintained by anomalous vertical mixing
due to increases in the atmospheric momentum flux (TKE input) into the
mixed layer (which deepens in the model after the shift) from the ear
ly 1970s to the late 1970s and 1980s. Since the ocean model does not c
ontain feedback to the atmosphere and it succeeds in capturing the maj
or features of the 1976-77 shift, it appears that the midlatitude part
of the shift was driven by the atmosphere, although effects of midlat
itude ocean-atmosphere feedback are still possible. The surface mixed-
layer heat budget also reveals that, in the central Pacific, the effec
ts of heat flux input and vertical mixing anomalies are comparable in
amplitude while horizontal advection anomalies are roughly half that s
ize. In the California coastal region, in contrast, where wind variabi
lity is much weaker than in the central Pacific, horizontal advection
and vertical mixing effects on the mixed-layer heat budget are only on
e-quarter the size of typical heat flux input anomalies.