POSTCRUISE ALVIN SUBMERSIBLE VISIT TO SITE 1253 AND 1255 CORK-IIS

Because of the likelihood that the spool valves had not shifted and the packer had not inflated at the time of the CORK-II installation at Site 1253, a visit to the site was planned with the manned submersible Alvin in November 2002, roughly 1 month after the installation of the CORK-IIs. Data recovery operations at wireline-installed CORKs in Holes 504B and 896A, located on the southern flank of the Costa Rica Rift, had been fortuitously scheduled for this time. Proximity of the Atlantis/Alvin port call to the margin sites, efficiency of operations at the rift flank sites, and assistance from the National Science Foundation made the visit to the margin sites possible. Participants in that program included K. Becker (chief scientist), E. Davis, T. Pettigrew, and R. Meldrum. The site visits are documented in two edited videos that show the seafloor installations and data downloading (Figs. F31, F32).

Two parallel solutions were prepared for resolving the spool-valve problem. The first would make use of a pump mated to the hydraulic connection at the top of the CORK-II assembly that accesses the packer-filling line (the very line that was believed not to have been pressurized when the packer-inflation go-devil failed to reach its landing point during deployment). This pump was designed and built by B. Carson and L. Holloway for formation testing in the CORKed Hole 949C in the Barbados accretionary prism. If this operation were to fail, a second approach would employ a stand-alone pressure sensor and data logger constructed by E. Davis and R. Macdonald that would be coupled to the redundant fluid sampling line from the lower CORK-II formation screen, which was plumbed to the seafloor through a valved port. This sensor/logger unit was equipped with an underwater-mateable electrical connector identical to that on the CORK-II unit itself to facilitate data downloading at the time of future site visits.

As it turned out, visual inspection to the CORK-II head showed that both spool valves had indeed shifted (although the left-hand piston was ~ in above the level of the right) despite the go-devil never properly landing. The pump was coupled and run for 16.5 min as an ultimate precaution (this served to shift the left-hand spool valve fully down to the same level as the right). However, following an inspection of the downloaded data, it became clear that the installation was successful from the beginning. The success was evident from the data showing tidal pressure signals in basement and small, but well-resolved, average pressure differentials (several kilopascals) between the two basement levels separated by the packer.

Time remaining after operations at Site 1253 was just enough to allow a visit to Site 1255 before the end of the dive. The site was somewhat difficult to find using the sonar scanner on the Alvin, as the site is located at the bottom of a local topographic slope or in a local depression. Here two more mysteries awaited. First, despite the T-handles having been secured by rubber bands at the time of deployment, the top-left sampling valve was found to be rotated ~10° from horizontal. This brought bad memories of Leg 196; the cause of this behavior remains unknown. Fortunately in this instance, the rotation was insufficient to cause the valve to leak. The second mystery is that the formation pressure registered by the sensor connected to the deeper level below the packer indicated leakage somewhere in the CORK-II plumbing until ~36 hr before the submersible arrived. Pressures at the sensor remained close to hydrostatic, and the tidal signal remained unattenuated relative to seafloor pressure until a time near the end of this first recording period, when the sensor pressure rose to a level somewhat higher than that observed in the upper interval (~180 kPa) and displayed an attenuated tidal variation. It is with considerable confidence that we now look forward to the first real phase of data recording and first OsmoSampler recovery at both of these sites.