JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 101, NO. C9, PAGES 20,711-20,726, SEPTEMBER 15, 1996 Eddies of newly formed upper Labrador Sea water Robert S. Pickart,• William M. Smethie Jr., 2 John R. N. Lazier, 3 E. Peter Jones, 3 and William J. Jenkins • Abstract. A new type of submesoscale eddy has been observed southof the Labrador Sea duringlate winter, embedded in equatorward flow alongthe western boundary. The eddyradius is 20 km, with a weak dynamic signature (swirlspeed of 0.5 cm/s). The center of the eddyis characterized by weak stratification, elevated concentrations of oxygen and anthropogenic tracers, and low tracer ages, all indicative of newly ventilated water. Strong lateral intrusions distortthe shape of the feature.The water mass contained in the eddyis not classical LabradorSeawater (from the centralLabradorSea) but is significantly fresher and hencelighter. It is of the correct density to be the source of the high- chlorofluorocarbon layer of the shallow deep western boundary current observed further south and henceis termed upper LabradorSeawater. Using a combination of hydrographic data sets alongthe western boundary to implementa simplelateral diffusion model, it is shown that such eddies decay of the order of several months and are difficult to observe equatorward of the Grand Banksof Newfoundland. This is in contrast to deeperlenses of classical LabradorSeawater whichpersist further equatorward. Tracer- derived ages of the upper LabradorSeawater eddy rangefrom 3 to 5 years, mucholder than the lifetime deduced from their lateral diffusion. A simple convection model of tracer age shows that this age discrepancy is caused by gasexchange beingunableto maintain equilibrium between the deep convecting mixedlayer and the atmosphere during formation. 1. Introduction Chlorofluorocarbons, or CFCs, were first measured in the North Atlantic Ocean in the early 1980s. On the basis of the understanding at that time of the subthermocline water masses and circulation in the western basin, it was believed that two dominant CFC signals would be revealed near the western boundary. First, it was expected'that near a depth of 3500 m, the Norwegian-Greenland overflow water would contain ele- vated concentrations of CFCs. This water mass originates at the sea surfaceat high latitudes and is quicklytransported southward as part of the deepwestern boundary current (or DWBC, which in this paper refersto the combination of water masses flowing equatorward along the western boundary in the depthrangeof 700-4000 m; see Pickart[1992]).The Norwe- gian-Greenland overflow water is thusquiteyoung relativeto the surrounding water and correspondingly hasa high concen- tration of dissolved oxygen (received during itscontact with the atmosphere). Similarly, it was expected that near a depth of 1800 m, there would be another strong CFC signal correspond- ing to convectively formed LabradorSeawater (LSW) which gets entrained into the DWBC [e.g., Talleyand McCartney, 19821 . The first CFC sections across the DWBC did indeed reveal a well-defined coreof highconcentration corresponding to the •Woods Hole Oceanographic Institution, Woods Hole,Massachu- setts. 2Lamont-Doherty EarthObservatory, Palisades, NewYork. 3Bedford Institute of Oceanography, Dartmouth, Nova Scotia, Can- ada. Copyright 1996by the AmericanGeophysical Union. Paper number96JC01453. 0148-0227/96/96JC-01453509.00 Norwegian-Greenland Sea overflow water near 3000-3500 m [e.g., Smethie, 1993] (Plate 1). However, whilethere wasin fact a second, shallower layer of high CFCs, thislayerwassituated above the LSW (between 500 and 1000 m); curiously, within the depth range of LSW the CFC concentrations were decid- edly low (Plate 1). In hindsight this surprising observation mighthave beenanticipated asearlier tritium measurements in the DWBC revealed this upperwell-ventilated layer [e.g., Jen- kinsand Rhines, 1980].However,the earlier observations did not make the existence of this layer any less puzzling. An increasing number of CFC sections in the North and South Atlantichave begun to map out thismiddepth signal. At this point, high CFCs have been detectedin this layer as far south as19øS in the DWBC [Wallace et al., 1994]. However, the influence of thiswater is by no means confined to the western boundary. Smethie [1993] shows that there are elevated amounts of CFCs in this layer seaward of the Gulf Stream in the subtropical gyre.Such presence of young water is largely due to interaction of the Gulf Streamand DWBC, whereby the separating Gulf Streamentrains a large amountof the upper DWBC [Pickart and Smethie, 1993]. The DWBC water then apparently recirculates as part of the southern Gulf Stream gyre[Spall, 1996]. South of the Gulf Stream crossover the CFC signal along the western boundary deepens [Fine andMolinari, 1988], in part because the upper portion of the layer was pulled off the boundary by the Gulf Stream and replacedfurther southby inflow which is low in CFCs [Pickartand Smethie, 1993]. According to Spall [1996], this inflow is in fact the recirculated upper DWBC water whichis significantly reduced in tracer content via mixingalong its interior path. More of the high CFC water in this layer is againpulled off the boundaryas the DWBC reachesthe equator. Here the CFC core splitsinto two tongues: one continuing southward 20,711