Summary

A specially designed battery powered vibracorer was developed and built at VUW over a period of 4 years to take cores up to 6 m long from the Antarctic continental shelf in water depths to 1000 m. The cores were to be used for research into the region's climate and glacial history over the last 20,000 years. The vibracorer, which weighed 1.5 tonnes, was designed to be deployed by a small mobile crane through 1.5 m diameter holes in the fast ice fringe around McMurdo Sound as well as from ships (Figure 1). For sea ice deployments a polyester winch line was chosen for lightness and ease of handling, with a breaking load of 6.75 tonnes, giving a safety factor of 4.5.

On 24 November 1993 the vibracorer was lost when the winch rope failed at the spliced eye used to connect winch rope to corer. Lowering into the water had just begun and the vibracorer free-fell 350 m to the sea floor. A remotely operated vehicle (ROV) operated by US colleagues nearby was used to look for the vibracorer, which was found upright on the first dive and with no visible damage. Two subsequent dives with a recovery line attached to the ROV did not find the corer because the ROV could not manoeuvre properly with the line attached and the dives were endangering the ROV. Further recovery attempts were abandoned. The site is located precisely using GPS at 76° 56.631′S and 162° 48.116′E, and a position surveyed independently to within 2 m will be available at a later date.

No reason for failure of the winch rope has been ascertained at this time. It had performed satisfactorily the previous season for 2 deployments in Antarctica and had been tested off Scott Base 7 days prior to 24 November with the full corer load lowered to a depth of 50 m.

The prospect of the corer's recovery from a ship in late January/February when the fast ice is weak or broken out is considered low. The option favoured in terms of least support requirements and most chance of success is to make another recovery attempt from new fast ice in October 1994 with the assistance of the US ROV, which will be working in the area, and a purpose built tool for attaching a cable to the corer.

Equipment Concept and Design

Logistic operation and constraints

The first part of the programme, to core from the fast ice at about 10 sites in Granite Harbour, would not only provide near shore scientific data but also a safe and stable platform from which to gain experience with the corer before ship deployment, with the considerable time pressures involved, in subsequent seasons. The party moved equipment and personnel with a D5 bulldozer, 2 five tonne cargo sledges, a sledge mounted accommodation/ laboratory wannigan, a steel sledge and a RN75 Nodwell vehicle with HIAB crane for lifting and drilling ice holes. This mode of operation has been used successfully for some years, and allows about 3 weeks field time after getting to Antarctica, testing equipment at Scott Base, travelling to the site and returning by around December 6, when summer warming cracks the sea ice.

The style of operation constrained the concept and design of the corer, putting a premium on lightness to reduce weight for the cargo sledges and for air cargo to and from Antarctica. Also because the equipment was developmental and had potential uses in back in New Zealand it was not considered feasible to send it to Antarctica the previous season by ship so it had to be air transported at the beginning of the season.

For operation from the fast ice up to 2.5 m thick the vibracorer needed to fit down a reasonably small hole, say 1.5 m in diameter. Also it was designed to operate from batteries to avoid an expensive and heavy power umbilical and large generator (15-20 KVA?) at the surface. It weighed 1.6-1.8 tonne in air and 1.2-1.5 tonne in water, though the exact weight is not known because of modifications to the legs in early 1993.

page break

The winch was designed to have a line pull of 2-2.5 tonne and again be kept as light as possible. It was powered from the Power Take Off of the hydraulic system on the Nodwell tracked vehicle. The line chosen was a polyester 20 mm diameter double braid rope with a breaking load of 7.5 tonne driven by a capstan onto a separately driven take up drum. A thimble eye splice at the end reduced the breaking load to 90% of the rated value, down to 6.75 tonne. With the corer weighing a maximum in water of 1.5 tonne the safety factor is 4.5. Polyester rope is low stretch and has the best resistance to UV, freezing temperatures and abrasion of the commonly used synthetic fibres.

Operating Procedures and Checks

Loss of Vibracorer in Granite Harbour

Normal checks and operating procedures were followed in this, the first deployment for the season. We arrived on the evening of November 22 and made the access hole in 2.3 m thick fast ice the following day. The vibracorer was assembled and operating systems checked. A small leak in the hydraulic system was stopped. On November 24 the weather was fine but overcast with an air temperature estimated at −5°C in the afternoon. Initially we had a leak in the air compensation system when the corer was first immersed, so it was returned to the surface and the air regulators washed and dried. On the second immersion of the corer the air compensation system functioned correctly, the weight of the corer was transferred to the winch rope and the chain hoist disconnected. Lowering of the corer proceeded until the winch rope eye was 3 or 4 m below the surface, when the rope parted, letting the corer free-fall about 350 m to the sea floor, The time was 1550 hrs.

Attempted Recovery of Vibracorer

Within 4 hours we were in contact with Dr Ross Powell, Northern Illinois University, USAP event S027, whose group had deployed their ROV (remotely operated vehicle) at the site the previous day (23 November). The next day their ROV was back and searching by late afternoon. The vibracorer was located and recorded on video about 20 m to one side of a point on the sea floor directly below the access hole and was sitting upright on the sea floor which sloped between 15 and 25°. The corer appeared intact with no visible structural damage. The feet were penetrating up to 15 cm into a surficial mud blanket over harder sea floor, indicated by scattered boulders covered with mud and encrusting organisms. The air compensator water trap appeared to be clear of water, though it probably would not stay like this for long because of pressure cycling due to tidal action and because once the air bottle pressure is equal to the water pressure the system is then open to the sea.

These observations suggest that the corer spiralled to the sea floor and landed on its feet without external damage. Internal damage within the four battery housings, air compensated electric motor and vibrator, and data logger instruments could not be determined.

Two further attempts were made to recover the corer with the ROV through the access hole in the fast ice on November 25 and 26. A hook and line were devised for the ROV to carry down and attach to the corer using a hook from the chain hoist and the winch rope. The hook was held by a release mechanism so that if the corer could be caught the ROV could detach itself and return to the surface before the corer was hauled up. The hook and winch line were partly buoyed with 3 kg net floats.

On both recovery dives the ROV reached the sea floor but could not swim and turn properly because the drag of the winch rope was too great for the propulsion unit. The corer was not seen on the first dive, On the second dive the ROV sonar picked up an object behind it and thought to be the corer. However attempts to turn the ROV and swim towards the object caused the ROV umbilical and the winch line to become intertwined, endangering the ROV. At this point attempted recovery was abandoned.

Requirements for Successful Recovery

We consider the best chance of recovering the corer to come from attaching a line by which it can be hauled vertically to the surface. This could most easily be done by a party on the fast ice with an ROV, winch and crane. Dr Powell's ROV would be suitable for attaching a purpose-built device (yet to be constructed) to the corer. We envisage the device to comprise a hook to grab the corer, a spool of doubled line on a pulley and a means of attaching this all to and releasing from the ROV. There would be no trailing line to the surface, which caused the previous recovery attempts to fail. Once the ROV had located the corer and hooked it, the spool would be released but for the ends of the double line which the ROV would carry back to the surface. Then a stronger line could be drawn down from the surface around the pulley, and finally the (pre-tested) winch rope itself to haul the corer to the surface.

We understand Dr Powell's ROV would be available for this purpose next November for 3 to 4 days, as he has scientific work planned in the area.

We also considered the possibility of recovery by trawling from a ship, but believe this has little chance of success. The boulders scattered on the sea floor would make it difficult or impossible to know whether the trawl had caught the corer or boulders before it was brought to the surface. Boulders in the trawl would most likely add to the damage caused by the trawling process. Also, pinpointing the location of the corer would require a ship with dynamic positioning, not available on ice breakers.

Likely Condition of the Corer on Retrieval

The ROV has shown that the exterior structure of the corer was intact about 24 hours after impacting the sea floor. We have no idea of the internal damage caused by the impact but expect that the 288 kg of sealed lead acid batteries will be damaged, although they are contained within high tensile steel housings(< 6 mm thick) designed for water depths of 1000 m. The electro-hydraulic motor housings and vibrator are pressure compensated with air which is likely to leak in the short term and allow sea water ingress.

The vibracorer is made primarily of mild and high tensile steel, coated with industrial grade paints, 316 and 304 stainless steel, marine grade aluminium alloy and high tensile aluminium alloys. These metals are in contact with each other and no sacrificial anode protection was installed because planned immersions were for only a few hours duration for each deployment. The sea water in the Ross Sea is well oxygenated at depth and we expect that corrosion will occur at normal or accelerated rates especially where electrolytic action can occur between dissimilar metals. The presence of any electrical change remaining in the battery packs is likely to enhance corrosion.

We believe that all the electrical components will suffer irrevocable damage in the short term. The main battery packs were probably damaged internally on impact. Instruments in a stainless steel pressure housing with anodised high tensile aluminium (Alumec 79) closures will become damaged in the longer term when the housing is corroded through. Part of the main lifting structure is built of 3 aluminium tubes (100 mm OD, 6 mm wall), which will corrode and weaken the structure with time.

Potential Environmental Impact of Non-retrieval

Environmental impact of the corer remaining on the sea floor depends on rates of corrosion of the metal components, biological response (encrustation, biocorrosion), and on the release of sulphuric acid and lead to the sea some time in the next few years. The main sealed lead acid batteries contain a total of 12.5 kg of sulphuric acid and approximately 240 kg of lead and lead compounds. Rates for all of these reactions are unknown.

The impact on the corer on organisms and sea floor within a few metres could well be significant, but insignificant on a decameter - kilometre scale. Any localised toxic effects are expected to be short term and dissipate within a few years.

An important argument for retrieval, beyond the obligation to endeavour to keep the Antarctic environment pristine, is to provide information on rates of encrustation and corrosion on metals in Antarctic marine waters, and hence obtain realistic estimates of the environmental impact of the loss of equipment in Antarctic waters.

Impact of Loss of Vibracorer on Programme

We must now postpone the ice sheet retreat study approved by RDRC and funded by the University, and withdraw our application for the 1994/95 season. We would nevertheless like to build another corer because the scientific problem of dating the ice sheet retreat remains and is likely to attract increasing scientific interest. We hope this will be possible with funds from the insurance cover. We would also like to build another corer because of the expertise developed over the last three years in such a sea floor sampling system, and the research funds committed to it. We would like to see a tangible product for that time and investment.

We expect drawings, construction and testing to take around two years from the time funding became available. Construction time could also take longer over the next 3-4 years because of our primary commitment to the support of the Cape Roberts Project from 1994 to 1997. Nevertheless we think it realistic to oversee the construction of a new corer in parallel. We therefore do not expect to submit a proposal for pursuing this project until the 1997/98 season.