click image to enlarge               
(Above) potential temperature vs. time and depth from June 1994 to June 1995, at climate mooring Bravo, central Labrador Sea.

From 1994 to 2000 Peter Rhines of University of Washington and John Lazier of Bedford Institute of Oceanography, Nova Scotia, Canada, maintained a moored set of instruments in the central Labrador Sea, to record both the fine scale physics of deep convection and the climate change which is so active in this region. The above figure shows the period June 1994-June 1995, in a depth-time diagram of potential temperature, based on 12 instruments on a mooring at 56.75N, 52.5W. The temperature scale is on the right. Surface waters have actually to cool below the temperature of the deep waters in order to initiate convection, as the surface waters have very low salinity. This is 'cold-driven' convection retarded by fresh-water buoyancy. Eventually the coldness wins out, and low-salinity waters are carried downward and stirred. Notice the high-frequency content of the time-series: small-scale yet tall features are swept past the mooring by the O(20 cm/sec) horizontal currents, which are nearly barotropic (nearly uniform in the vertical). Some of these features are convective plumes with width 200m-1km. Simultaneous salinity records show some density compensation, but not total. The complete description of the this 'theta-S' lateral fine structure is not yet known.
Also shown on the figure are the depths of 6 of the instruments with time (white curves); the mooring is blown over by strong eddies, in one instance instruments were carried down about 600m.

The ocean depth here is 3650m, and beneath the cold, fresh Labrador Sea Water (surface to about 2200m in winter) lies warmer, saline waters known as Northeast Atlantic Deep Water, some of which has entered the Atlantic from east of the Faroe Islands, where a passage opens onto the Norwegian Sea. Still deeper is a colder moderate salinity layer some of which comes cascading down the slope east of Greenland, from the Denmark Strait and the Greenland Sea. This 'Denmark Strait Overflow Water' is the densest water mass originating in the 'Nordic Seas' and Arctic which can make it over the Greenland-Scotland Ridge. All of these events are 'summarized' by the beautiful Erika Dan section of L.V. Worthington and Redwood Wright, carried out in the late winter of 1962. This section, below, is from their publication in the Atlas Series of the Woods Hole Oceanographic Institution (1970).

The winter (1994-5) recorded in the colored mooring record above was the culmination of one of the coldest periods in a century, and since then winters have been much milder. The figure below shows the water column after winter, for the decade of the 1990s. The high-pass filtered salinity profiles show that intense winters 'wipe the slate clean', and then in mild winters a great hash of theta-S fine-structure appears as the central Labrador Sea is restratified from the side. The bold black lines are surfaces of constant potential density, and the light blue curve is the depth to which convection penetrates in winter. Dr. J.M. Lilly has identified eddies ejected from the boundary current along west Greenland as a strong source of warm, saline restratification (Progress in Oceanography 2003; see download section of this website).

Global warming is suspect, in many ways: through release of fresh-water through melting and enhanced precipitation, through increase in the North Atlantic Oscillation activity, which can bring cold winter winds to the region (the 'NAO-+' phase), and eventually through warming and failure of the wintertime convection process.

This region is a headwater for the global overturning circulation of the world ocean; through the Labrador Sea pass the dominant, dense, deep waters which arrive from farther north, having flowed over the ridge connecting Greenland and Scotland. These stack up in boundary currents that circumnavigate the subpolar zone of the North Atlantic, from Canada to Europe. They also interact with the waters of the central Labrador Sea which we see here. These become the intermediate depth circulation of the global circulation, lying between 800 and 2000m depth once they escape the Labrador Sea and move southward along the Americas.

For some literature on this subject see the 'downloads' section on this website. In particular, the 1999 Journal of Physical Oceanography paper by Jonathan Lilly and others.

Experiments in the GFD lab and with computer models have begun to describe the convection process, the larger mesoscale eddies that become involved, and the still larger circulation of the entire ocean basin. Some images from our GFD lab, showing eddies produced by cooling with a circular disk (a thermoelectric cold plate). The disk is at the surface of a rotating fluid in a 1m diameter circular cylinder. shows the time-development. We don't see the fine-scale tornado-like plumes here, but just their mesoscale products, illuminated with fluorescein dye injected with a fine glass catheter at the center of the cold plate.

If the same experiment is carried out in a bowl-shaped basin, here and here, rather like cooling a lake at its edge, the result is a topographic 'beta-plume', a cyclonic gyre that extends cyclonically round the rim of the basin. It meets an anticyclone which fills the remainder of the basin. This is a rapidly rotating experiment in which the buoyancy-flux based Rossby number is quite small. In some experimentsJ the cold disk is transparent, which allows us to see the fine-scale convective plumes as well as the larger eddies and gyres.

The images here show the classic fine-scale cyclones beneath a uniformly cooled surface in a rotating fluid. The side-view (left) is dominated by the larger mesoscale that almost always develops (at least if there are slight lateral contrasts in buoyancy forcing); the fine scale 'tornado' plumes dart downward in between the larger features.