Marginal sea overflows and the upper ocean interaction

Marginal sea overflows and the overlying upper ocean are coupled in the vertical by two distinct mechanisms— by an interfacial mass flux from the upper ocean to the overflow layer that accompanies entrainment and by a divergent eddy flux associated with baroclinic instability. Because both mechanisms tend to be localized in space, the resulting upper ocean circulation can be characterized as a b plume for which the relevant background potential vorticity is set by the slope of the topography, that is, a topographic b plume. The entrainment-driven topographic b plume consists of a single gyre that is aligned along isobaths. The circulation is cyclonic within the upper ocean (water columns are stretched). The transport within one branch of the topographic b plume may exceed the entrainment flux by a factor of 2 or more. Overflows are likely to be baroclinically unstable, especially near the strait. This creates eddy variability in both the upper ocean and overflow layers and a flux of momentum and energy in the vertical. In the time mean, the eddies accompanying baroclinic instability set up a double-gyre circulation in the upper ocean, an eddy-driven topographic b plume. In regions where baroclinic instability is growing, the momentum flux from the overflow into the upper ocean acts as a drag on the overflow and causes the overflow to descend the slope at a steeper angle than what would arise from bottom friction alone. Numerical model experiments suggest that the Faroe Bank Channel overflow should be the most prominent example of an eddy-driven topographic b plume and that the resulting upper-layer transport should be comparable to that of the overflow. The overflow-layer eddies that accompany baroclinic instability are analogous to those observed in moored array data. In contrast, the upper layer of the Mediterranean overflow is likely to be dominated more by an entrainment-driven topographic b plume. The difference arises because entrainment occurs at a much shallower location for the Mediterranean case and the background potential vorticity gradient of the upper ocean is much larger. 1. Overflow and upper ocean interaction Marginal sea overflows enter the open ocean as dense, bottom-trapped gravity currents (Fig. 1). The Denmark Strait, Faroe Bank Channel, Mediterranean Sea, Red Sea, and Filchner Bank overflows are five major marginal sea overflows that are known to play an important role in supplying deep and intermediate water masses to the global ocean (Warren 1981). Observations indicate that overflows also affect their overlying water through entrainment and eddy formation (Saunders 2001; Candela 2001, and references therein). a. Mass and vorticity balances The importance of overflows on determining the deep ocean properties led past studies to focus on how entrainment affects overflows. Understanding how the dynamics of overflows evolve as they descend the continental slope progressed from so-called stream-tube models (e.g., Smith 1975; Killworth 1977; Price and Baringer 1994). Stream-tube models assume an inactive upper layer, which may have been appropriate to explain the basic dynamics of the overflow. However, this assumption certainly cannot be appropriate for the * Current affiliation: Earth Simulator Center, Japan Agency for Marine-Earth Science and Technology, Yokohama, Japan. Corresponding author address: ESC, JAMSTEC, 3173-25 Showamachi, Kanazawa-ku, Yokohama 236-0001; Japan. E-mail: [email protected] FEBRUARY 2009 K I D A E T A L . 387 DOI: 10.1175/2008JPO3934.1 2009 American Meteorological Society upper oceanic layer because the upper ocean also needs to balance mass lost to the overflow somehow. The Faroe Bank Channel overflow, for example, entrains about 1.5 Sv (Sv [ 10 m s) of overlying North Atlantic water near the shelf break while descending the continental slope (Fig. 1) (Mauritzen et al. 2005). Localized entrainment also implies vortex stretching in the upper ocean. So, there must be a convergent flow in the upper ocean that balances both the mass and vorticity fluxes induced by entrainment. b. Time variability associated with overflows Satellite altimetry revealed regions of high variability of the sea surface height about 50–100 km downstream from the marginal sea strait of the Mediterranean, Faroe Bank Channel, and Denmark Strait overflows (Høyer and Quadfasel 2001; Høyer et al. 2002). Satellite infrared imagery and floats have also shown intense cyclones forming above the Denmark Strait overflow, suggesting active upper ocean and overflow interaction (Bruce 1995; Krauss and Käse 1998). For the Faroe Bank case, which we will emphasize here, snapshots of sea surface height show fluctuations of 610 cm and eddies having a radius of roughly 50 km (Ezer 2006). In situ observations of the overflow show the corresponding mass and current variability: the temperature of the Faroe Bank Channel overflow has been observed to fluctuate with a 3–4-day period about 140 km downstream from the Faroe Bank Channel (Fig. 2; Høyer and Quadfasel 2001; Geyer et al. 2006). These latter observations appear to show more or less discrete eddies moving along the bathymetry. At a given point, the associated temperature fluctuations are up to 48C, comparable to the temperature difference between the Faroe Bank Channel source water and the ambient North Atlantic water (Mauritzen et al. 2005). The temperature of the Mediterranean overflow has also been observed to fluctuate with a 7–9-day period about 200 km downstream from the Strait of Gibraltar with no such oscillations observed near the strait (Stanton 1983; Chérubin et al. 2003). The Denmark Strait overflow has also been observed to be associated with significant time variability (Käse et al. 2003). Observations indicate large time variability both in the overflow layer and its overlying oceanic layer downstream from the strait. There has been some progress in understanding the time variability associated with overflows and its generation mechanisms. Laboratory experiments have shown steady and laminar overflows developing variability, such as waves and eddies downstream from straits (Cenedese et al. 2004). Multiple regimes of eddy formation associated with strong cyclones and anticyclones in the upper layer have also been found (e.g., Whitehead et al. 1990; Etling et al. 2000). Primitive equation models support development of such variability downstream from the strait as a result of interaction with its upper layer through entrainment, baroclinic instability, and vortex stretching (Jiang and Garwood 1996; Spall and Price 1998; Jungclaus et al. 2001). One-and-a-half-layer models of overflows, however, have also shown that overflows may become a chain of eddies even in the absence of upper-layer motion when the transport of the source water varies with time (Nof 1991) or when a steady solution does not exist for the overflow layer FIG. 1. Schematic of an overflow and its mass balance: The transport values are roughly based on the Faroe Bank Channel overflow. Dense water that forms in the marginal sea spills over the sill as an overflow. This overflow (1.5 Sv) descends the continental slope, entrains overlying upper oceanic water (1.5 Sv), and reaches its neutral buoyancy level or the bottom. Figure adapted from Price and Baringer (1994). FIG. 2. Time–latitude plot of the near-bottom temperatures (contoured) and bandpassed (2–8 days) currents (arrows) of the Faroe Bank Channel overflow observed about 140 km downstream from the Faroe Bank Channel (Høyer and Quadfasel 2001). The temperature fluctuates with a 3–4-day period. Figure reproduced by courtesy of D. Quadfasel. 388 J O U R N A L O F P H Y S I C A L O C E A N O G R A P H Y VOLUME 39