On the deep western-boundary current in the Southwest Pacific Basin

The principal system of deep western-boundary currents in the subtropical S outh Pacific is. that in the Southwest Pacific Basin, which transports Circ umpolar Deep Water northward and Pacific Deep Water (at mid-depths) southwa rd. The WOCE PCM9 current-meter array was placed across this system at Lat. 32 degrees 30'S in order to measure the mean transports of those component s and their variations. The array, consisting of 60 current meters on 20 mo orings, extended 1000 km eastward from the Tonga-Kermadec Ridge. and remain ed in place for 22 months, between February 1991 and December 1992. The ins truments were situated approximately at 2500 m, 4000 m, and close to the bo ttom. CTD sections (including dissolved oxygen and nutrients) were occupied along the array during its deployment and recovery, and, in between, by th e WOCE transpacific section P6. Density sections were used to construct obj ectively-mapped fields of geostrophic velocity, which were adjusted using c urrent-meter data as integration constants to provide snapshots of the full velocity field. The resulting adjusted transports sometimes differed subst antially from relative geostrophic transports, but agreed quite well with t ransports calculated from current records alone. A time series of volume tr ansport was derived from objectively-mapped three-day-averaged currents. Th e boundary-current system at PCM9 was essentially 700 hm wide, with flow mo st intense on the flank of the Tonga-Kermadec Ridge, where the maximum mean velocity vector had a magnitude of 9.6 cm s(-1). The time-averaged transpo rt. integrated horizontally across the array and from 2000 m to the bottom, was 16.0 x 10(6) +/- 11.9 x 10(6) m(3) s(-1) northward. Of this roughly 15 .8 x 10(6) +/- 9.2 x 10(6) m(3) s(-1) was northward flow of Circumpolar Dee p Water. and 0.2 x 10(6) +/- 5.1 x 10(6) m(3) s(-1) was northward flow of P acific Deep Water. Even for the 22-month mean, however, the velocity field was strikingly banded vertically, and there was little impression of a zero -velocity surface following the demarcation between Circumpolar Deep Water and Pacific Deep Water across the section; only in the horizontally integra ted sense was there a correspondence between water masses and the variation of transport with depth. The very large variability in transport is associ ated with prominent oscillations of periods near 50 days. 20 days, and 10 d ays, as well as with strong events distributed irregularly across the array that lead to a concentration of spectral energy in a band between 40 and 2 00 days. The origins of these disturbances are not known. While unexpectedl y large changes in the density field near the Tonga-Kermadec Ridge were obs erved from one cruise to another, the huge fluctuations in transport seemed to be connected more with velocity signals varying only slowly with depth. No measurable changes in water-mass properties were detected by the cruise s during the 22 months of deployment, but the salinity was about 0.01 lower at the salinity maximum in the Circumpolar Deep Water than it had been 25 years earlier. The direct, long-term transport measurement suggests that th e total upwelling at 2000 m north of 30 degrees S is 13 x 10(6) m(3) s(-1), corresponding to an areally-averaged vertical velocity of 1.0 x 10(-5) cm s(-1). This is substantially smaller than earlier values, and it helps to r educe estimates of the global deep upwelling closer to those of the global deep downwelling. The small value of Pacific Deep Water transport in the boundary-current sys tem, relative to that of Circumpolar Deep Water, implies, within the framew ork of the Stommel-Arons dynamics, that little of the deep water entering t he Pacific from the Antarctic returns southward at mid-depths. If so, then some present-day circulation schemes and budgetary constructions need to be re-assessed. (C) 1999 Elsevier Science Ltd. All rights reserved.