Starting in 1983 French oceanographers conducted a very beautiful Lagrangian study of flow near the mid-Atlantic ridge south of the Azores. Their objective was to study exchange of fluid motion across a major topographic feature on long time scales. By deploying them clusters they could also study dispersion processes. They deployed a total of 30 SOFAR floats at roughly 700 m depth in two comparable groups at two sites about 300 km to either side of the ridge crest. The impact of the ridge striking. Since this area is east and south of the Gulf Stream in what we might envision as the wind-driven Sverdrup gyre one would expect a mean flow to the east and south at a cm/s rate. What did they find?
These two figures show the spread of 25 floats over the first two years. The left panel shows their day-by-day trajectories while the right panel shows the start-to-end vectors. Of the 14 red and 11 blue floats deployed west and east of the ridge only one from group each crossed the ridge. They are shown in black and green. Due to the shoaling ridge topography, floats cannot drift toward the axis, only away. This gives the impression there is a mean flow away from the ridge, but this is misleading due to their uneven distribution to begin with. A close examination of the trajectories will reveal a tremendous amount of detail, and I strongly recommend reading the two papers by Ollitrault and Colin de Verdière (references below). Here I do something very simple: giving equal weight to the two clusters, what is the ensemble mean 2-year velocity of all 25 floats? It is surprisingly small.
mean u and v for group west: -0.29, 0.17 km/day N = 14
mean u and v for group east: 0.45, -0.23 km/day N = 11
As is evident from the vector plot these number indicate that two groups move in nearly opposite directions. When combined the net mean velocity of all floats over the two years is:
<u,v> = 0.15 ± 0.19, -0.06 ± 0.09 km/day or 0.18 ± 0.23 , -0.07 ± 0.11 cm/s.
The means are only borderline significant but taken together they suggest a very weak mean flow across the ridge at 700 m depth with just a hint of an anticyclonic pattern: NE west and SW east of the ridge. I had not expected the mean flow to be this small. These numbers are comparable to the ~ 3 mm/s mean flow at 2 km depth over and west of the East Pacific Rise I wrote about last fall.
Why is the ensemble mean velocity so small? The dynamical impact of the ridge is the obvious reason even though it crests at ~1500 m depth or ~800 m below the floats. We’ve seen the impact of the mid-Atlantic ridge in other Lagrangian studies, notably the one by Bower et al. (2002). Another reason for the weak mean flow may be due to vertical shear. The wind-driven Sverdrup gyre, so far as we know, is limited to the top kilometer of the ocean.
The mean velocities may be small, but when multiplied by large cross-sections, say a length of the ridge, they can add up to sizable transports. These two studies show the effectiveness of floats to determine weak flows accurately. In fact, if it is mean flow over long time that is sought, it may not be necessary to track them – it will suffice to get their start and end points with what might be called deep drifters. They would give us a simple easy-to-use tool for probing longer timescales in the ocean. You could liken such floats to a ‘geochemical’ tracer with the advantage that you can choose location and duration of study.
What do we really know about the interior circulation of the North Atlantic? We rely heavily upon geostrophy, but as always this requires accurate reference information. We deploy current measuring moorings, but their high cost limits their use. They may also subject to subtle biases due to the impact of topography in their vicinity. We use satellite-derived sea surface height to estimate upper ocean velocities, but it is useless at depth where we need to determine mean velocities to sub-mm/s accuracy.
Almost certainly the mean abyssal circulation will look different at depths above and below ridge crest depths. While eddy-dispersive processes may filter out all but the largest topographic features (long ridges), topography will likely shape mean flow at various depths such that on multi-year timescales flow may not have a simple 2-mode pattern. Suitably deployed floats (deep drifters) would give us a cost-effective tool with which to address these questions.
Bower A.S., B. Le Cann, T. Rossby, W. Zenk,, J. Gould, K. Speer, P. L. Richardson, M. D. Prater, and H.-M. Zhang, 2002. Directly measured mid-depth circulation in the northeastern North Atlantic Ocean. Nature 419:603–7
Ollitrault, M, and A. Colin de Verdiere, 2002. SOFAR floats reveal midlatitude intermediate North Atlantic general circulation. Part I: a Lagrangian descriptive view. J. Phys. Oceanogr. 32, 2020–33
Ollitrault, M, and A. Colin de Verdiere, 2002. SOFAR floats reveal midlatitude intermediate North Atlantic general circulation. Part II: An Eulerian Statistical View. J. Phys. Oceanogr. 32, 2034–53