The Oleander has been collecting velocity data since the Fall of 1992. Due to novel technical challenges working on a commercial vessel the quality and quantity of data available for analysis were limited at first, the principal difficulty being that of electrical noise in the shipboard power system. Considerable editing of the raw data was required. Starting in 1996 after the ship had been in drydock for routine biannual service and inspection, the depth penetration and volume of data increased substantially. There may have been some change in shipboard machinery. At that time we also changed the orientation of the ADCP such that both pairs of beams point diagonally down to the sides of the vessel instead of fore-aft and athwartships. Whatever the reason, we now obtain better data in heavier (but not heavy) weather than in the past, mostly on outbound transits. In heavy weather bubble entrainment under the vessel all but obliterates the acoustics of the ADCP. For the same reason the data return on inbound transits is less due to the ship's light load factor. With the data now in hand, we seek to examine in greater detail the structural stability of the current, and to identify the corresponding dynamical factors responsible for this. These sections provide an extensive data set with exceptional resolution of the upper ocean velocity field: 8 meters in the vertical and 2.4 km in the horizontal (5 minute averaging at 16 Kt vessel speed). We also have surface temperature from all transits (recorded by the ADCP), and once a month XBTs and surface salinities are collected. The temperature profiles will serve as proxy for density. This is of course a limitation near the surface where the T/S relationship breaks down.

The ADCP data have been processed and archived using the CODAS software package (Firing, 1991). The velocity accuracy depends very heavily upon how well the ship's motion and heading are known. Since 1995 a differential GPS receiver has corrected for the effects of deliberate degrading of GPS navigation (selective availability; a.k.a. dithering). The differential GPS transmission covers the continental shelf, the Slope Sea and usually the entire Gulf Stream. But even without differential GPS, the effect of dithering is quite limited provided one allows for longer averaging in time, 20 minutes, say (Flagg et al., 1998). Since 1995 we also operate a heading GPS system (3D-GPS) which allows us to correct the gyro compass error directly to a heading accuracy of ±0.1°. This matters since an error of 1° introduces a 0.14 m/s error in cross-track velocity at a vessel speed of 16 Kts. In case the 3D-GPS malfunctions, we employ an empirically established statistical correction from recent cruises, but the uncertainties can become significant in heavy weather. Under good operating conditions we think the velocities are accurate at the ±0.03 m/s level.

At the beginning of every month a volunteer observer joins the ship for a one-week roundtrip to Bermuda. They take XBTs across the shelf, the Slope Sea and the Gulf Stream on the outbound leg and some additional XBTs in the Sargasso Sea on the return leg. The horizontal resolution of the XBTs varies and depends upon location. In the Gulf Stream profiles are taken about every 25 km. The shelf and Slope Sea sampling program has been in operation for over 20 years. The once a month sampling means that only one out of eight crossings will have XBTs and depending upon the weather the concurrent ADCP data may or may not be very useful. Thus the number of sections with simultaneous good XBT and ADCP coverage of the core of the Gulf Stream is significantly less than the total number of sections available for analysis.