Institut fur Meereshunde
University of Kiel
J.Geophys.Res., submitted

Flow and Recirculation of Antarctic Intermediate
   Water across the Rio Grande Rise

by O.Boebel, C.Schmid, & W.Zenk

   During Meteor cruise No. 22, legs 4 and 5, 21 RAFOS floats were deployed in the
subtropical South Atlantic around the Rio Grande Rise to obtain Lagrangian current 
observations. All instruments were built and preballasted  at  the Institut fur 
Meereskunde (IFM Kiel) [Konig and Zenk, 1992]. Since the main objective of this 
cruise was to recover and deploy current meter moorings in the Vema and the Hunter 
Channel, the deployment sites of four sound sources - necessary for the acoustic 
tracking of RAFOS floats - and float launch positions were constrained by logistic 
necessities. Equidistant float launch sites along the cruise track were chosen, 
resulting in an inter-float spacing of approximately 100 km.  All floats were 
launched within 16 days from December 4, 1992 to December 19, 1992 (Table 1) except 
for one float (no. 93) which was launched on January 4, 1993, during METEOR cruise 
no. 22, leg 5. The float weights were adjusted individually to reach neutral buoy-
ancy around 900 m depth, close to the salinity minimum of the AAIW. They recorded 
pressure, temperature and sound signal travel times for three listening windows 
once a day. 

   Due to leaks in both the hydrophone feedthrough as well as the release plug 6 
floats surfaced prematurely within a week or so. Trajectories of the remaining 15 
floats were obtained, resulting in 6.2 float years of data. From these 15 floats, 
10 floats (nos. 67, 70, 71, 72, 74, 86, 87, 88, 89 and 91) suffered from a small 
leakage which made them sink slowly: One float (no. 71) sank to the emergency pres-
sure release level of 1350 dbar and consequently surfaced prematurely after 45 days,
whereas the other 9 showed a self sealing effect at a higher depth, resulting in 
underwater drifts covering a pressure range from 738 dbar up to 1355 dbar. Some of 
the slowly sinking floats did receive the sound sources only weakly (no. 67, day 
100-200; no. 74, day 50-200, no. 86, day 80-130, no. 91, day 0-70), most likely due 
to the technical problems described above. These gaps in the time of arrival data 
were interpolated linearly.  The remaining 5 floats (no. 66, no. 76, no. 84, no. 90 
and no. 93) recorded only small pressure fluctuations on the  order of 50 dbar and 
consequently are considered as tight. The start and end points of the trajectories 
were traceable within 5 km of the launch and first surface position of the float. 
Some floats had to be tracked across the connecting line between two sound sources 
(base line crossing). This results in a reduced local accuracy or unsteady behavior 
of the position data. However, in a statistical sense, these jumps are compensated, 
if a larger data segment around the base line crossing is considered.