Figure 1.

Figure 2.

Figure 3.

Figure 4.

H. Thomas Rossby is a professor of oceanography at the University of Rhode Island's Graduate School of Oceanography. His interests include the study of ocean currents and their variability and the development of ocean instrmentation. He likes to go to sea..
George Schwartze, Jr. received his training in electronics and computers while serving in the US Navy for six years. He has been at GSO since 1976, working first with the atmospheric chemistry group and now with the physical oceanography group.

In the early 1980s, a remarkable instrument that can profile upper ocean currents from a ship came into widespread use. The acoustic Doppler current profiler (ADCP) works by transmitting sound down from the ship and recording the return echoes. The transmitted sound consists of short acoustic pulses that are beamed in four directions: fore, aft, port, and starboard. The echoes result from the sound bouncing off layers of zooplankton that normally inhabit the upper ocean. Due to the relative motion between the waters below (on average the zooplankton move with the water) and the ship, the reflected signal's frequency will be shifted a small but precisely measurable amount. By listening to this frequency shift (Doppler) as a function of time delay (the surface water echoes return first and deeper waters later), one can construct a profile of Doppler shifts as a function of depth. These can then be converted into a profile of currents relative to the ship to of 200m to 400m depth. In years past, the lack of precision navigation limited the system's usefulness to shallow waters where the ship's motion could be tracked over the bottom or to coastal waters with good LORAN-C navigation. With the advent of the Global Positioning System (GPS), a ship's movement can be accurately tracked everywhere, and by subtracting the ship's speed and heading (obtained with GPS) we can estimate the currents within a range of a few centimeters per second.
     In 1992, we installed the ADCP profiling instrument the in the hull of the MVOleander and have operated it continuously since then. The freighter Oleander (and its predecessors) operates between New Jersey and Bermuda on a weekly schedule. The ship has been reporting weather observations, temperature, and salinities for decades. The six years of current measurements offer an interesting view of seasonal and annual variability. Figure 1 shows the ship's track and velocity vectors at 52m depth on one such transit from Bermuda, with sea surface temperature obtained one day later from an orbiting satellite. The figure shows the coincidence of the warm waters and high velocities that characterize the Gulf Stream. Indeed, it is these high speeds that bring the warm waters north from the Caribbean Sea. By contrast, the velocity field in the surrounding waters is quite variable, and only by making many trips can one obtain enough information to construct an accurate estimate of the mean field of currents. Figure 2 shows the mean field along the Oleander line from the first six years of operation. The long vectors pointing northeast between 37°N and 38°N indicate the mean position and strength of the Gulf Stream. The ellipses provide a measure of the variability of the currents: greater in the direction of elongation of the ellipse and less in the normal direction. Their orientation in the direction of the current suggests that most of the variability results from the meandering of the current and not from a large-scale turbulent eddy field. Away from the stream, both north and south, we observe a weak, rather uniform mean flow to the west that cannot be discerned from a single transit as in Figure 1. Also, the nearly circular variance ellipses indicate that the currents fluctuate with equal likelihood in all directions. The westward flowing waters between the Gulf Stream and 34°N look like they will rejoin the stream west of the line, whereas the waters south of 34°N have a more southwesterly heading. The 34°N latitude seems to separate two regimes: a recirculating body of water to the north, and a flow to the south.
     Much of the focus of this program to date has centered on the Gulf Stream and how it varies in space and time. In Figure 3, we show a composite plot of all crossings of the current but plotted relative to the velocity maximum of each crossing. That is each crossing, such as the one in Figure 1, is plotted so that only the downstream components, or those parallel to the velocity maximum, are retained. These are then plotted as a function of cross-stream or normal distance from the velocity maximum, which we position at the origin in the figure. Notice the general similarity of the crossings and tight scatter around the velocity maximum: 2.05 plus or minus 0.25ms^-1. This figure tells us that the Gulf Stream maintains a well-defined structure regardless of its position and direction of flow. Significantly, the maximum velocity shows virtually no variation with season or from year to year. Further, if we sum the velocities from each transit across the Gulf Stream, we find that downstream transport at this depth is relatively constant, with little variation between seasons or from year to year.
      The stability and steadiness of the structure of the Gulf Stream means that its mass transport remains quite stable. This observed steadiness of Gulf Stream transport appears to contradict other studies that suggest that the Gulf Stream can undergo significant variations in transport. During the 1990s, major changes in the wind forcing of the North Atlantic circulation have taken place. We have noticed these indirectly through the very mild winters that New England has experienced in the last few years. Will these large-scale changes show up and transport changes in the Gulf Stream or be manifested in some other way? Theoretical arguments suggest a delay of about three years, the time required for a change or disturbance in the large-scale ocean circulation to propagate across the ocean and manifest itself in the Gulf Stream. The Gulf Stream also transports enormous amounts of thermal energy which is, needless to say, of tremendous interest in terms of climate. Were the current to slow down or its average temperature to drop, the amount of heat carried into the northern North Atlantic would decrease, potentially causing very serious consequences to the climate of Europe and perhaps other parts of the northern hemisphere. By combining these observations with many other data sets, both concurrent and historical, we hope to learn more about the stability of the Gulf Stream and its mass and heat transport.
      Curiously, while the structure of the Gulf Stream has remained stable, it has shifted position substantially during this period. In 1994-1995, it crossed the Oleander transit line almost 100km farther north and, since early 1996, it has shifted back south. These shifts show a strong correlation with major changes in temperature and salinity of the Slope Waters. This correlation is evident in Figure 4, which shows the anomaly of sea surface temperature of the waters between the Gulf Stream and the shelf break as a function of time. An anomaly signifies that the mean and annual cycles of temperature have been removed so that only the departures from the annual cycle remain. (The actual sea surface temperature varies from approximately 24°C in summer to approximately 11°C at the end of winter.) Note the very cold waters in early 1997 and 1998 despite the fact that those were very mild winters (especially 1998) along the east coast of the United States. The second panel shows the position of the Gulf Stream during the same period. Notice its southerly position when the waters are cold and more northerly position when the waters are warm. A similar pattern of variability can also be found in the 21-year long record of surface temperature and salinity along the same line. Although that sampling program did not extend south across the Gulf Stream, we can tell from the data that when the Slope Waters turn exceptionally warm and saline, the Gulf Stream shifts north roughly 100km. What causes the Stream to migrate north or south on these long, interannual time scales, and why do the surface water properties change so dramatically? Our working hypothesis is that the cold waters (associated with a southward displacement of the current) result from a stronger flow of cold waters from the Labrador shelf.
      We are now looking for other data sets to help clarify this question. These figures illustrate how repeat transits enable us to investigate questions that would be extremely difficult to pursue by other means.
      Our most sincere thanks and appreciation go to the operators of the Bermuda Container Line for their strong and continuing interest and support of these measurement programs on the MV Oleander.