11.4: Observed Surface Circulation in the Atlantic

The theories by Sverdrup, Munk, and Stommel describe an idealized flow, but the ocean is much more complicated. To see just how complicated the flow is at the surface, let’s look at a whole ocean basin, the North Atlantic. I have chosen this region because it is the best observed, and because mid-latitude processes in the Atlantic are similar to mid-latitude processes in the other ocean. Thus, for example, I use the Gulf Stream as an example of a western boundary current.

Let’s begin with the Gulf Stream to see how our understanding of ocean surface currents has evolved. Of course, we can’t look at all aspects of the flow. You can find out much more by reading the Tomczak and Godfrey (1994) book Regional Oceanography: An Introduction.

North Atlantic Circulation

The North Atlantic is the most thoroughly studied ocean basin. There is an extensive body of theory to describe most aspects of the circulation, including flow at the surface, in the thermocline, and at depth, together with an extensive body of field observations. By looking at figures showing the circulation, we can learn more about the circulation, and by looking at the figures produced over the past few decades we can trace an ever more complete understanding of the circulation.

Let’s begin with the traditional view of the time-averaged surface flow in the North Atlantic based mostly on hydrographic observations of the density field (figure \(2.3.1\)). It is a contemporary view of the mean circulation of the entire ocean based on a century of more of observations. Because the figure includes all the ocean, perhaps it is overly simplified. So, let’s look then at a similar view of the mean circulation of just the North Atlantic (figure \(\PageIndex{1})\)).

Figure \(\PageIndex{1}\): Sketch of the major surface currents in the North Atlantic. Values are transport in units of \(10^{6} \ \text{m}^{3}/\text{s}\). After Sverdrup, Johnson, and Fleming (1942: fig. 187)

The figure shows a broad, basin-wide, mid-latitude gyre as we expect from Sverdrup’s theory described in Section 11.1. In the west, a western boundary current, the Gulf Stream, completes the gyre. In the north a subpolar gyre includes the Labrador current. An equatorial current system and countercurrent are found at low latitudes with flow similar to that in the Pacific. Note, however, the strong cross equatorial flow in the west which flows along the northeast coast of Brazil toward the Caribbean.

If we look closer at the flow in the far north Atlantic (figure \(\PageIndex{2}\)) we see that the flow is still more complex. This figure includes much more detail of a region important for fisheries and commerce. Because it is based on an extensive base of hydrographic observations, is this reality? For example, if we were to drop a Lagrangian float into the Atlantic would it follow the streamline shown in the figure?

Figure \(\PageIndex{2}\): Detailed schematic of named currents in the North Atlantic. The numbers give the transport in units on \(10^{6} \ \text{m}^{3}/\text{s}\) from the surface to a depth of 1 km. Eg: East Greenland Current; Ei: East Iceland Current; Gu: Gulf Stream; Ir: Irminger Current; La: Labrador Current; Na: North Atlantic Current; Nc: North Cape Current; Ng: Norwegian Current; Ni: North Iceland Current; Po: Portugal Current; Sb: Spitsbergen Current; Wg: West Greenland Current. Numbers within squares give sinking water in units of \(10^{6} \ \text{m}^{3}/\text{s}\). Solid Lines: Warmer currents. Broken Lines: Colder currents. After Dietrich et al. (1980: 542).

To answer the question, let’s look at the tracks of 110 buoys drifting on the sea surface as compiled by Phil Richardson (figure \(\PageIndex{3}\) top). The tracks give a very different view of the currents in the North Atlantic. It is hard to distinguish the flow from the jumble of lines, sometimes called spaghetti tracks. Clearly, the flow is very turbulent, especially in the Gulf Stream, a fast, western-boundary current. Furthermore, the turbulent eddies seem to have a diameter of a few degrees. This is much different than turbulence in the atmosphere. In the air, the large eddies are called storms, and storms have diameters of \(10^{\circ} - 20^{\circ}\). Thus oceanic “storms” are much smaller than atmospheric storms.

Perhaps we can see the mean flow if we average the drifter tracks. What happens when Richardson averages the tracks through \(2^{\circ} \times 2^{\circ}\) boxes? The averages (figure \(\PageIndex{3}\) bottom) begin to show some trends, but note that in some regions, such as east of the Gulf Stream, adjacent boxes have very different means, some having currents going in different directions. This indicates the flow is so variable that the average is not stable. Forty or more observations do not yields a stable mean value. Overall, Richardson finds that the kinetic energy of the eddies is 8 to 37 times larger than the kinetic energy of the mean flow. Thus oceanic turbulence is very different than laboratory turbulence. In the lab, the mean flow is typically much faster than the eddies.

Further work by Richardson (1993), based on subsurface buoys freely drifting at depths between 500 and 3,500 m, shows that the current extends deep below the surface, and that typical eddy diameter is 80 km.

Figure \(\PageIndex{3}\): Top: Tracks of 110 drifting buoys deployed in the western north Atlantic. Bottom: Mean velocity of currents in \(2^{\circ} \times 2^{\circ}\) boxes calculated from tracks above. Boxes with fewer than 40 observations were omitted. Length of arrow is proportional to speed. Maximum values are near \(0.6 \ \text{m/s}\) in the Gulf Stream near 37\(^{\circ}\)N 71\(^{\circ}\)W. After Richardson (1981).

Gulf Stream Recirculation Region

If we look closely at figure \(\PageIndex{1}\) we see that the transport in the Gulf Stream increases from \(26 \ \text{Sv}\) in the Florida Strait (between Florida and Cuba) to \55 \ \text{Sv}\) offshore of Cape Hatteras. Later measurements showed the transport increases from \(30 \ \text{Sv}\) in the Florida Strait to \(150 \ \text{Sv}\) near 40\(^{\circ}\)N.

The observed increase, and the large transport off Hatteras, disagree with transports calculated from Sverdrup’s theory. Theory predicts a much smaller maximum transport of \(30 \ \text{Sv}\), and that the maximum ought to be near 28\(^{\circ}\)N. Now we have a problem: What causes the high transports near 40\(^{\circ}\)N?

Niiler (1987) summarizes the theory and observations. First, there is no hydrographic evidence for a large influx of water from the Antilles Current that flows north of the Bahamas and into the Gulf Stream. This rules out the possibility that the Sverdrup flow is larger than the calculated value, and that the flow bypasses the Gulf of Mexico. The flow seems to come primarily from the Gulf Stream itself. The flow between 60\(^{\circ}\)W and 55\(^{\circ}\)W is to the south. The water then flows south and west, and rejoins the Stream between 65\(^{\circ}\)W and 75\(^{\circ}\)W. Thus, there are two subtropical gyres: a small gyre directly south of the Stream centered on 65\(^{\circ}\)W, called the Gulf Stream recirculation region, and the broad, wind-driven gyre near the surface seen in figure \(\PageIndex{1}\) that extends all the way to Europe.

The Gulf Stream recirculation carries two to three times the mass of the broader gyre. Current meters deployed in the recirculation region show that the flow extends to the bottom. This explains why the recirculation is weak in the maps calculated from hydrographic data. Currents calculated from the density distribution give only the baroclinic component of the flow, and they miss the component that is independent of depth, the barotropic component.

The Gulf Stream recirculation is driven by the potential energy of the steeply sloping thermocline at the Gulf Stream. The depth of the 27.00 sigma-theta \((\sigma_{\theta})\) surface drops from 250 meters near 41\(^{\circ}\)N in figure \(10.5.1\) to 800 m near 38\(^{\circ}\)N south of the Stream. Eddies in the Stream convert the potential energy to kinetic energy through baroclinic instability. The instability leads to an interesting phenomena: negative viscosity. The Gulf Stream accelerates, not decelerates. It acts as though it were under the influence of a negative viscosity. The same process drives the jet stream in the atmosphere. The steeply sloping density surface separating the polar air mass from mid-latitude air masses at the atmosphere’s polar front also leads to baroclinic instability. For more on this topic see Starr’s (1968) book Physics of Negative Viscosity Phenomena.

Let’s look at this process in the Gulf Stream (figure \(\PageIndex{4}\)). The strong current shear in the Stream causes the flow to begin to meander. The meander intensifies, and eventually the Stream throws off a ring. Those on the south side drift southwest, and eventually merge with the stream several months later (figure \(\PageIndex{5}\)). The process occurs all along the recirculation region, and satellite images show nearly a dozen or so rings occur north and south of the stream (figure \(\PageIndex{5}\)).

Figure \(\PageIndex{4}\): Gulf Stream meanders lead to the formation of a spinning eddy, a ring. Notice that rings have a diameter of about 1^{\circ}\). After Ring Group (1981).

Figure \(\PageIndex{5}\): Sketch of the position of the Gulf Stream, warm core, and cold core eddies observed in infrared images of the sea surface collected by the infrared radiometer on NOAA-5 in October and December 1978. After Tolmazin (1985: 91).