Absolute transport of the North Atlantic Current at 42N and the circulation in the Newfoundland Basin

Christopher S. Meinen and D. Randolph Watts
University of Rhode Island
Graduate School of Oceanography

The absolute transport of oceanic currents has been directly measured at few locations in the world, mainly due to the difficulty of obtaining absolute velocity measurements. Even today, most ocean current transport estimates are snapshots and include only the baroclinic transport relative to an assumed level of no motion. Two experiments are reported here measuring the absolute transport of the North Atlantic Current (NAC) normal to the WOCE ACM6 transect near 42.5N: (1) In August 1993, geostrophic velocity shears were determined from a CTD transect and absolutely referenced using POGO floats and ADCP measurements. The NAC absolute transport across this section was 112 Sv +/- 25 Sv. (2) From August 1993 to June 1995, timeseries of velocity shear relative to the bottom were determined using inverted echo sounders along the ACM6 line; these measurements were absolutely referenced using deep current meters and bottom pressure sensors. The temporal mean NAC transport was 146 Sv +/- 14 Sv; the standard deviation was 41 Sv. These transports include a contribution by the Mann Eddy, a large permanent anticyclonic eddy located adjacent to the NAC at this latitude. Nearly all historical estimates of the NAC transport in this area have also included the transport of the Mann Eddy, as the two flows are generally indistinguishable. Because our new transport measurements include the barotropic component, they are 30-60% larger than NAC transport estimates reported by historical studies in which the baroclinic transport was referenced to a level of no motion at the bottom or at 2000 dbars, respectively. Accommodation of this higher transport into the Newfoundland Basin circulation scheme along with other published transport estimates suggests a significant southward recirculation elsewhere in the basin.


Goals of the Study


1993 CTD Section


19-month moored instruments



1993 CTD Section

This first figure shows the location of the study area . Left panel: Locations of the CTD profiles and POGO launches from the 1993 CTD section aboard the R.V. Oceanus. Right panel: Locations of the moored inverted echo sounders (IES) and deep current meters deployed in August 1993 and recovered in April-June 1995. Also shown in the right panel are the locations of other CTD profiles obtained during the mooring experiment.
Three different methods of referencing geostrophic relative velocity profiles were used after Pickart and Lindstrom [1994]: spatially integrating the absolute velocity measurements of an Acoustic Doppler Current Profiler (ADCP) horizontally between CTD stations; temporally integrating ADCP measurements while on CTD sites; and using POGO transport floats to provide the reference (the two ADCP methods were modified somewhat from those used by Pickart and Lindstrom). The biases between absolute velocity sections referenced by the different methods were less than 1 cm/s and the standard deviations of the differences were 5-7 cm/s.

For these three methods, the following sources of error were studied: Ageostrophic velocities due to current path curvature, Schuler oscillations in the ship gyrocompass, inertial oscillations, ADCP misalignment error, GPS accuracy (including dithering), Ekman velocities, scatter in the ADCP measurement, ADCP amplitude coefficient error, heading dependent gyrocompass error, spatial sampling errors, and POGO velocity measurement error. Based on estimates of the sizes of all of these sources of error during the experiment, we determined that the use of the POGO transport float to provide absolute referencing for geostrophic relative velocity profiles resulted in the smallest estimated error at all three CTD-pair spacings tested: 20 km, 40 km, and 60 km. During the 1993 cruise, referencing via the POGO method resulted in absolute velocities accurate to within 4 cm/s.

The next figure shows the resulting velocity sections . Panel A shows the absolute velocity section for the 1993 section using the best available method for each CTD pair (the POGO float failed for two CTD pairings). The green shading and dashed contours indicate southward velocities, the medium blue denotes northward velocities less than 50 cm/s, and the light blue denotes northward velocities greater than 50 cm/s. Contour level is 10 cm/s. Red circles on lower axis indicate CTD sites. Gray shading indicates the ocean bottom. Panel B shows the vertical average of the absolutely referenced horizontal velocity. The error bars indicate the accuracy of the absolute reference velocity. Green shading indicates southward velocity, blue indicates northward.


Interpreting the IES measurements

Hydrographic measurements from the Newfoundland Basin (shown on the site map) were integrated to obtain a value of tau (round trip travel time) for each cast. The corresponding temperature (T) profiles for all of these casts were then smoothed onto a regular grid of pressure and tau. A smoothed field of specific volume anomaly (delta) was similarly produced. These fields show the dominant T and delta profiles associated with any given value of tau. The variation shown in these figures documents a single, dominant mode of variability which we refer to as the "Gravest Empirical Mode" or GEM. The word "mode" here does not refer to an analytical or dynamical mode, but a purely empirical mode.

The scatter about these smoothed temperature and delta fields is quantified in the subsequent plots. The middle panels show the root-mean-square difference between the actual CTD measured T values and the smoothed T field in absolute units. The highest errors are confined to the upper 300 dbars where seasonal fluctuations in temperature are large. The scatter progressively decreases with depth and is uniformly small below the main thermocline.

The total range of T and delta values vary with pressure level, with the largest range in the main thermocline levels and smallest in the deeper waters. Thus the bottom panel shows the same rms differences but normalized by the total (peak--to--peak) T range at each pressure level. The scatter in the thermocline region represents less than 5% of the total signal for both temperature and delta; only in the deepest waters where the actual temperature signal becomes very small does the scatter exceed 10%.


Results from 19-month mooring deployment

Six sample plots detailing the temperature (shading) and velocity (contours) structures observed on the dates shown. Velocity contours are at 10 cm/s intervals, with bold contours at intervals of 50 cm/s. The weak currents shown in the left panels, with peak velocities of only 60 cm/s, result from oblique crossings of the moored section by the North Atlantic Current. The methods used here only determine the normal component of the velocity. The upper right panel shows a more normal velocity section, with peak speeds of over 100 cm/s. The middle and lower panels on the right show the extremes of the effect of the Mann Eddy on the apparent velocity of the North Atlantic Current. In the middle panel the eddy is nearly separate from the North Atlantic Current while in the bottom panel the two northward flows are completely coalesced, with peak velocities of over 160 cm/s. The North Atlantic Current and the northward flow of the Mann Eddy are generally indistinguishable, like the right top and bottom panels. On brief occasions the eddy moved shoreward enough that the southward flow of its eastern edge was observed by our moored instruments, as in the bottom right panel.

Based on these daily pictures the stream-coordinates mean temperature and velocity sections were determined. Velocity contours are at 10 cm/s intervals, with bold contours at intervals of 50 cm/s. Temperatures are displayed in degrees C. The stream--coordinates origin was defined as the 10 C isotherm at 450 dbars. Nineteen months of daily measurements were averaged; time periods when the North Atlantic Current was crossing the moored section obliquely by more than 20 degrees, approximately 30% of the time series, were excluded prior to averaging.

The next figure shows the absolute transport across the section in different seasons as calculated from the moored instruments. The top panel shows the net transport (northward minus southward) integrated in each gap between moorings individually. The bottom panel shows the cumulative integrated transport (northward component only!), with the integration beginning at the second mooring from shore. The inshoremost mooring was neglected in order to focus on the northward transport associated with the combined North Atlantic Current and Mann Eddy.


Comparison to historical transport estimates

Different estimates of the NAC transport are listed in this table. Asterisk -- Mann's estimate does not include the transport of the so-called ``Mann Eddy'' as it had moved away from the NAC during his study. Reiniger and Clarke [1975] used 24 hour averages from moored current meters to reference three separate geostrophic shear sections from hydrographic sections. Clarke et al. [1980], Mann [1967], and Worthington [1976] worked solely with unreferenced geostrophic shear sections based on hydrographic measurements (Worthington had two sections).

The absolutely referenced estimates of the transport agree to within the accuracy of the various estimates. Because most of the historical transport estimates were made relative to the assumption of a level of no motion at 2000 dbars or at the bottom, these transport values are shown also. There is fairly good agreement between these various estimates as well. Note the importance of the bottom velocity component of the transport however. It represents 17-35% of the total transport!

The reduction in transport when calculated in stream--coordinates was unexpected. Work with a simple analytical model of an eddy indicates that, since the Mann Eddy does move with respect to the core of the North Atlantic Current, the stream--coordinates averaging would result in the averaging of some of the northward flow of the Mann Eddy with some of the southward flow of the eddy. The result would be a lower northward transport, as was observed. For this reason the Eulerian transport estimate was used for comparison to other transport estimates in the northwest North Atlantic.


Discussion of circulation

The combined northward absolute transport of the North Atlantic Current & Mann Eddy was observed to be about 145 Sv at 42.5N. This is significantly larger than historical (non--absolute) estimates at this location, which prompted us to reconsider the overall circulation ideas for the northwestern North Atlantic. This sketch indicates a possible circulation scheme for this region. The numbers indicate transport estimates in Sverdrups (1 Sv = 1000000 m^3/s) made in this and other studies. All of the transports are absolute except for the estimate of the eastward transport across the Mid-Atlantic Ridge, where no absolute estimate has ever been made. Several studies have suggested, however, that the level of no motion assumption at the bottom may be valid crossing the Mid-Atlantic Ridge, so this estimate is treated as robust.

Historical estimates of the recirculation within the Mann Eddy give 50-60 Sv; hence the difference, 90 Sv, is identified as the throughput northward transport of the North Atlantic Current on this transect. With 90 Sv entering this part of the North Atlantic Basin and only 30 Sv leaving over the ridge, significant southward recirculation must occur elsewhere in the Newfoundland Basin in addition to the recirculating Mann Eddy. Potential pathways for this recirculation are shown by dashed lines. A number of float studies have shown no evidence for southward recirculation offshore of the North Atlantic Current. This is a region of high eddy variability, however, and the mean velocities needed to produce our hypothesized southward recirculation are so small (0.5 cm/s) that it is unlikely that they would be easily observed.


Conclusions


For further information please contact Christopher Meinen (meinen@pmel.noaa.gov) or D. Randolph Watts (rwatts@gso.uri.edu) via email. This poster was presented at the WOCE Conference in Halifax, Canada during 24-29 May 1998.

The GEM methodology has also been applied to measurements in the Subantarctic Front south of Tasmania. This highly successful application was the subject of another poster at the WOCE conference.

URI/GSO: Dynamics of Ocean Fronts and Currents Homepage