Formation rates of subtropical underwater in the Pacific Ocean

Water mass formation rates were calculated for subtropical underwater (STUW) in the North and South Pacific by two partially independent methods. One is based on the World Ocean Circulation Experiment (WOCE)/TOGA drifter array over two periods: 1988-1992, and 1992-1996. Drifter velocities were used to calculate two components of the subduction rate, lateral induction and vertical pumping. The second method used CFC-12 data (1987-1994) from WOCE and Pacific Marine Environmental Laboratory to calculate ages on sigma(o) surfaces. Subduction rates were estimated from the inverse age gradient. The two subduction rate methods are independent, but they share a common identification of STUW formation area based on satellite-derived surface temperature maps. Using both methods, one can put bounds on the formation rates: 4-5 Sv in the North and 6-7 Sv in the South Pacific. The drifter calculated STUW subduction rates for 1988-1992 and 1992-1996 are 21 and 13 m/yr in the North Pacific and 25 and 40 m/yr in the South. The CFC-12 calculated STUW subduction rate in the North Pacific is 26m/yr, and 32m/yr in the South. The South Pacific rates exceed those in the North Pacific. Consistent differences between the two methods support earlier studies, they conclude that mixing contributes to STUW formation in addition to the larger-scale circulation effects. The drifter and tracer rates agree well quantitatively, within 22%, except for the second period in the North Pacific and there are some differences in spatial patterns. Tracer rates integrate over time, and drifters allow analysis of interannual variability. The decrease in subduction rate between periods in the North Pacific is due to negative lateral induction entraining STUW into the mixed layer. The increase in the South Pacific rate is due to an increase in the vertical pumping. Although Ekman pumping is in phase in the North and South, the subduction rate is out of phase. These results confirm that subduction depends on the large-scale circulation and a combination of the outcrop pattern and air-sea fluxes. Temporal differences in rates and partitioning between the hemispheres are consistent with interannual changes in gyre intensity and current positions. (C) 2002 Elsevier Science Ltd. All rights reserved.