Temporal Variability of the Transport Through Drake Passage

Six years of bottom pressure data have been used for the monitoring of the temporal variability of the Antarctic Circumpolar Current (ACC) transport at Drake Passage (DP). (Figure 1) shows the locations of the bottom pressure recorder (BPR) deployments. The FS1, FS2 and FS3 positions were first instrumented in November 1988, with the BPRs being moved to the DN93 and DS93 positions in November 1992, where monitoring continues. Additional bottom pressure data from the DN92 and DS92 positions were obtained for December 1991 to November 1992. (Figure 2) shows the currently available time series of bottom pressure from the north and south DP positions and north-minus-south pressure difference respectively, with the dashed line being the data smoothed with a 1 cycle/day cut-off low pass filter, and the solid line being the same data smoothed with a 31 day low-pass filter. Geostrophic calculations on these data yield values for the standard deviation in transport of approximately 8 Sv, subject to the assumption that the variability is predominantly barotropic (e.g. Vassie et al., 1994). This is slightly less than the 10 Sv obtained during the International Southern Ocean Studies (ISOS) programme in the mid-1970s to early 1980s (Whitworth and Peterson, 1985).

Pressure data from the FS2 position were found to be unsuitable for direct inclusion in geostrophic calculations, due to the proximity of the FS2 instrument to the Polar Front (PF). A strong correlation between bottom pressure and bottom temperature indicates that meanders/eddies and lateral shifts of the PF (e.g. Legeckis, 1977) have a significant effect on the pressure record. The fact that this was a positive correlation indicates that the change in sea level across the PF has a greater influence on bottom pressure than the change in density across the front. Historical monthly-mean sea-surface temperature (SST) maps were obtained from the Jet Propulsion Laboratory (JPL) for the period 1982-1986, and used to investigate the low-frequency movements of the PF. The maps were filtered with a 2 pixel square gradients filter, and the mean latitude of the PF for each month extracted. A very regular semiannual signal was found in the latitude of the PF, peaking each year in May/November or June/December, agreeing well with the observed semiannual signals in the FS2 bottom pressure and temperature data. The annual signal from the SST maps was less regular, peaking between April and July each year.

Since 1992, the DP BPR deployments have also featured Inverted Echo Sounders (IESs) on the instrument rigs. Data from these have been used to empirically obtain parameters such as dynamic height and vertically-averaged density at the north side of the passage, from which sea level can be obtained (Woodworth et al., 1995). No such derivation is possible for the south-side IES data, due to a non-linear relationship between measured acoustic travel time and the desired parameters. The IES data indicate that there is significant steric contamination in the BPR data from the north side of DP due to activity of the SubAntarctic Front (SAF), and consequently the north-minus-south pressure difference will overestimate the true transport variability. However, the fact that bottom pressure and acoustic travel time from north DP were observed to be negatively correlated implies that bottom pressure will provide a better measure of transport variability than sea level, since the change in density across the SAF partially compensates for the change in sea level.

To investigate the causes of the observed transport variability, gridded surface wind stress data were obtained from the National Center for Atmospheric Research (NCAR). These were derived from the output of the European Centre for Medium-Range Weather Forecasts (ECMWF) global analyses. Following Wearn and Baker (1980), the zonally-averaged eastward wind stress was calculated for the latitude band 37.5 to 67.5 deg.S. (Figure 3a) shows this compared to the bottom pressure data from the south side of DP (here plotted inverted for comparison). It is clear that there is a strong negative correlation between the series. No such agreement was found with bottom pressure from the north side of DP, due at least partially to the steric contamination outlined earlier. Wind stress curl was also calculated for the latitude bands 40-45 deg.S and 60-65 deg.S, after Peterson's (1988) suggestion that changes in the wind stress curl adjacent to the ACC will influence the Sverdrup transport into the ACC band, and hence the ACC mass balance and geostrophic transport. Some agreement was found between these parameters and south DP pressure, but it is believed that this is due to the fact that these manifestations of the wind field strongly resemble the zonally-averaged wind stress, rather than being the cause of the variability in the pressure. This is suggested by the fact that south DP pressure more closely resembles wind stress curl at the north side of the ACC than wind stress curl at the south side.

BPR data previously collected (1986-1988) at Amsterdam and Kerguelen Islands in the southern Indian Ocean were also compared to the different measures of wind forcing. These data are described by Vassie et al. (1994). (Figure 3b) shows the series of bottom pressure at Kerguelen Island compared to the zonally- averaged wind stress, and again it is clear that there is significant variability common to both series. It is interesting to note that here there is a direct correlation between wind and bottom pressure, whereas there was a negative correlation between wind and south DP pressure. From the geostrophic relationship, this implies that the forcing for the pressure variability at Kerguelen actually occurs south of the islands, and hence the Amsterdam-minus-Kerguelen pressure difference will not be truly representative of the ACC transport variability. The wind stress curl time series were found to agree less well with Amsterdam and Kerguelen bottom pressure, except at very long (i.e. greater than annual) periods. This again suggests that the zonally-averaged wind stress is causing the observed variability, though it is clear that an analysis of the BPR data from the three World Ocean Circulation Experiment (WOCE) choke points is needed to resolve this matter.

Acknowledgements

We are very grateful to Tony Craig at NCAR for supplying the wind data, to the JPL PODAAC for the AVHRR data, and to Bob Spencer and the POL Technology Group for maintaining and operating the BPR/IES instruments.

References

Legeckis, R., 1977. Oceanic polar front in the Drake Passage - Satellite observations during 1976. Deep- Sea Research, 24, 701-704.

Peterson, R.G., 1988. On the Transport of the Antarctic Circumpolar Current Through Drake Passage and Its Relation to Wind. Journal of Geophysical Research, 93(C11), 13,993-14,004.

Vassie, J.M., A.J.Harrison, P.L.Woodworth, S.A.Harangozo, M.J.Smithson & S.R.Thompson, 1994. On the temporal variability of the transport between Amsterdam and Kerguelen Islands. Journal of Geophysical Research, 99(C1), 937-949.

Wearn, R.B. & D.J.Baker, 1980. Bottom pressure measurements across the Antarctic Circumpolar Current and their relation to the wind. Deep-Sea Research,27A. 875-888.

Whitworth, T. & R.G.Peterson, 1985. Volume transport of the Antarctic Circumpolar Current from bottom pressure measurements. Journal of Physical Oceanography, 15, 810-816.

Woodworth, P.L., J.M.Vassie, C.W.Hughes & M.P.Meredith, 1995. A Test of TOPEX/POSEIDON's Ability to Monitor flows through the Drake Passage. Submitted for publication.

Figure Captions

Figure 1 Locations of POL bottom pressure recorder (BPR) deployments at Drake Passage.

Figure 2 Drake Passage bottom pressure time series (a) from the north side (b) from the south side and (c) north-minus-south pressure difference. Dashed line is data smoothed with 1 cycle/day cut-off filter, solid line is same data smoothed with 31-day moving average filter.

Figure 3a Zonally-averaged eastward wind stress between 37.5 and 67.5* (solid) and south Drake Passage bottom pressure (dotted; plotted inverted for comparison). Both series smoothed with 31-day moving average filter.

Figure 3b Zonally-averaged eastward wind stress between 37.5 and 67.5*S (solid), and bottom pressure measured at Kerguelen Island (dotted). Both series smoothed with 31-day moving average filter.

This work has since been published in:

On the temporal variability of the transport through Drake Passage. Meredith, M.P., J.M.Vassie, K.J.Heywood and R.Spencer. Journal of Geophysical Research, vol.101, no.C10, 22485-22494. (Correction vol.102, C2, 3501).

For more information, contact:

Mike Meredith m.meredith@uea.ac.uk