Calculations of absolute velocity for the general circulation are carried out on a 10° × 10° grid covering the South Pacific Ocean. The method is founded on a large-scale potential vorticity (PV) conservation that balances vortex stretching with meridional transport between isopycnals. The optimal estimate of velocity is determined by minimizing a square-error measure of model misfit, which is formed from the adjustment to isopycnal slope required to attain an exact PV balance. The measure is compared directly with uncertainty in the large-scale isopycnal slope to obtain a feeling for the validity of the PV dynamics. Based on this test, the model dynamics cannot be rejected at any of the subtropical grid points. They are, however, more in question at the westernmost grid locations near the Tonga-Kermadec Ridge where slope adjustment exceeds uncertainty. The smooth density field, used in the model PV equation, is defined through a least-sqyares regression of bivariate splines. Bivariate splines adequately model the abrupt change in isopycnal slope associated with the Antarctic Circumpolar Current. A break point with a slope discontinuity along 50°S results in the best overall fit. A region where PV is nearly uniform along isopycnals is found in the upper waters of the subtropics. The region is characterized by approximately constant zonal isopycnal slope in conjunction with a β/ƒ variation in meridional slope with depth. The constant vertical gradient in meridional slope is accommodated through a poleward shift in the latitude at which successively deeper isopycnals reach their maximum depth. Poleward shifts of subtropical gyres have been observed in other oceans. Because of the nearly uniform PV, this shallow region cannot contributed significantly to the estimation of absolute velocity. A milder form of degeneracy in the PV field is observed in the Artarctic Circumpolar Current Region, south of 50°S. Although PV is not uniform, the direction of its gradient along isopycnals is nearly constant over depth. Consequently, absolute velocity does not significantly change direction with depth. This is a case where the path of the absolute flow is well defined but its magnitude and sense are left indeterminate. The two major components flow, generally recognized within the subtropical South Pacific, are indicated in the absolute flow patterns. In the upper reaches, the existence of a subtropical anticyclone is well established and is evident on maps of deep reference level steric height. Mid-latitude estimates of absolute velocity, presented here, also exhibit a shallow anticyclonic flow. The absolute flow pattern is a consequence of the relatively weak absolute velocity magnitudes, which are determined at a mid-depth reference level. The velocity estimates are determined without a priori assumptions concerning the vertical structure of the flow. They derive solely from a consideration of PV and the large-scale density field. A deeper flow pattern, recognized in the South Pacific, arises from the northward extension of the Antarctic Circumpolar Current into the deep Southwest Pacific Basin along the Tonga-Kermadec Ridge. Some part of the abyssal western boundary current turns east and south giving the deep flow a cyclonic character. The flow pattern is supported by the distribution of characteristics and simple thermohaline flow theory. The cyclonic flow pattern differs from steric height maps referred to deep isobars, which indicate anticyclonic shear throughout the depth region considered (<3500m). However, the PV field suggested a cyclonic absolute flow field at depth even though the PV and steric height fields are both obtained from the density field alone. The qualitative correspondence between the recognized flow patterns and absolute flow fields lends credence to absolute velocity estimates determined at individual grid points.
Read full abstract
Jan 1, 2015
Giancarlo Sorrentino + 7
Open Access
