Variability in the Upper Ocean Mass and Thermal Structure of Drake Passage

  

Janet Sprintall, Scripps Institution of Oceanography, UCSD

   


Temperature Variability.

Year-round, near bi-monthly monitoring of the upper-ocean temperature variability in Drake Passage has been undertaken since September 1996 through repeat XBT surveys. The 29 transects completed to date follow three main tracks shown in Figure 1. Approximately 70 XBT probes are dropped per crossing at intervals of ~10-15 km, with higher station spacing of ~6 km when crossing the Subantarctic Front (SAF) and the Polar Front (PF). The mean positions of the SAF and PF from the XBT data are much further south than their historical positions. There is significant variation in the width of the Antarctic Polar Frontal Zone (APFZ) on interannual time scales (Figure 2), although there is no clear seasonal signal nor a dependence on the frequent presence of eddies (E) in the APFZ.


Figure 1. 		  
Figure 2.

The mean temperature structure shows a strong temperature gradient across the passage, with the highest variability found at the surface with the meridional excursion of the PF (Figure 3). At the southern end of the transect, the variability reflects the formation of the sub-zero temperatures of the Antarctic Surface Water formed in JAS and surface-capped to ~100 m each JFM by summer heating. For this reason, the average top 100 m temperature is much cooler in winter than the average temperature at depth, which shows little variation between 200-700 m (Figure 4). Mesoscale eddy-type features populate the APFZ during most winter (JAS) surveys, and their impact is evident in the exchange of Antarctic and SubAntarctic water masses in the northern part of the section. The minimum variability between 200-800 m and 59°S-61°S occurs in the Upper Circumpolar Deep Water, which remains a constant 2°C.


Figure 3. 		  
Figure 4.


Geostrophic Transport Variability.


To determine transport estimates from the XBT temperature measurements we derive an empirical relationship between upper ocean temperature and a baroclinic transport streamfunction. The multivariate regression relationship is determined using over 250 CTD casts from the historical data collected in the Drake Passage. Validation of the relationship is obtained using the ISOS cruises in Drake Passage during 1975-76 and 1979-81, which consist of both full-depth CTD casts and XBT measurements. The rms difference between the actual dynamic height relative to 2500 m, and the value predicted by the regression relationship using both CTD and the XBT temperatures was of the order of 5-10 dyn cm. The difference in dynamic height across the Drake Passage is of order 1.5 dyn m. This suggests that the relationship can be used to determine baroclinic transport (relative to 2500 m) from the recent high density XBT sections with only a small error.

The geostrophic transport for each cruise strongly suggests a banded nature to the flow through the passage (Figure 5). Alternating eastward and westward bands of transport are evident in the APFZ, although in general, strong eastward flow is found at the location of the SAF and PF (triangles). Some, although not all, of the reversing bands of flow can be attributed to the presence of mesoscale eddy-type features. In the mean cross-track transport (Figure 6), eastward transport is associated with the PF (~58S), although substantial eastward transport is also found north of the SAF along the continental boundary, albeit with strong variability. Transport to the south of the PF is typically much weaker, although again, significant eastward transport can be associated with the Continental Water Boundary (CWB) of Antarctica. The 5-year time series of total transport through Drake Passage shows substantial variability on interannual time scales, and commensurate with the positions of the fronts in Figure 2, no real seasonal signal is discernible (Figure 7). The mean total transport value of 78.3 Sv is of the same order as that determined from the ISOS measurements relative to 2500 dB.

Figure 5.
Figure 5 full size.
   
Figure 6. 		  
Figure 7.


Mesoscale Variability.


Mesoscale eddy-type features frequently populate the APFZ, particularly during late winter (September and November) surveys, although in more recent years they have occurred more regularly with no seasonal preference (Figure 2). The XBT temperature, XCTD salinity and TOPEX/Poseidon sea surface height anomaly (SSHa) at the time of the Drake Passage XBT surveys in September 1996 (Figure 8, top), September 1998 (middle) and November 1998 (bottom) clearly show the meridional water mass exchange of the cold, fresh Antarctic Surface Water to the south, and the warm, saltier SubAntarctic Water to the north. The variance of the SSHa shows a bullet in a deep basin through which the transect cross (Figure 9), that is strongest during JAS. Preliminary analysis of the SSHa data found both extreme high and low features congregate in this basin, before weakening and moving north-northwest with speeds 20-30 cm/s. Spectral analysis of the SSHa data suggest time scales of ~40 days and spatial scales of ~150-200 km. Future work will combine the highly complementary high resolution XBT and SSHa data to further examine the time and space scales of mesoscale eddy propagation and heat transport through Drake Passage.

Figure 8. 		  
Figure 9.