Robots under Ice: In October Last Year, Scientist Martin Doble Joined a Research Expedition to Antarctica to Gather Data beneath the Sea Ice Using State-of-the-Art Underwater Probes. but the Supercooled Seas, Temperamental Technology and Inquisitive Wildlife Made for an Eventful Month on the Ice

Article excerpt

[ILLUSTRATION OMITTED]

'Seal in the hole!' This wasn't the best time for a curious mammal to put in an appearance. Our robot submarine was due back in the ice hole any second now, and we had no idea how this big male Weddell seal was going to react.

Anxious moments followed, but fortunately, our visitor was interested without being aggressive or panicked, and the hole proved to be big enough for both machine and beast.

We soon became used to sharing our space with these placid animals while launching, running and recovering missions; not to the point of being blase about it, you understand--tapping away on a laptop in a tent while a half-tonne seal bobs and snorts in the freezing water at your feet never becomes 'routine'--but comfortable and relaxed nonetheless.

IN EREBUS'S SHADOW

I had come to Antarctica's Ross Island as part of an international team investigating how the ocean current flowing under a floating ice barrier leaves eddies and turbulence in its wake. Small-scale mixing such as this controls heat exchange between the water and the ice, and forms a piece of the puzzle in our understanding of how Antarctica's giant floating ice shelves will respond to warming seas.

The setting couldn't be more spectacular. The 3,794-metre active volcano Mount Erebus dominates the view, rising in gentle curves to its summit crater. A lava lake inside the crater sends a constant plume of steam into the sky, although no violent eruptions have taken place for thousands of years.

Vast glaciers creeping down the mountain's slopes drive the improbably narrow finger of the Erebus Glacier Tongue 15 kilometres out to sea. This floating glacier is around 300 metres thick where it leaves the shore, tapering to just 50 metres at its tip.

Early visitors to the region knew the glacier tongue well: both Robert Falcon Scott and Ernest Shackleton built huts nearby, travelling past it as they set off for the South Pole. That 'Heroic Age' is long gone, but the scientific ethos of those early explorers still has a major focus nearby: although seemingly infinitely remote, our camp was only about 20 kilometres from the USA's McMurdo Station, Antarctica's largest base. Its smaller companion, New Zealand's Scott Base, was our 'home town', from where we staged the experiment and to which we snuck back for the occasional shower.

[ILLUSTRATION OMITTED]

[ILLUSTRATION OMITTED]

The sea itself is frozen over for most of the year with more than two metres of ice, which glues the ice tongue in place and protects it from waves and currents raised by the often hurricane-force winds that howl down from the heart of the continent. The sea ice also allows easy travel around the region and provides a stable surface on which to set up camp.

If you want to study the ocean, however, it makes gaining access to the water difficult and time-consuming. In order to overcome this difficulty, we needed a device that could be launched from a small hole, travel around under the ice taking measurements, and then return to the same hole. Fortunately, just such a device exists: a 'robot submarine' known as an autonomous underwater vehicle (AUV). The AUV is programmed with a mission and, once under way, operates without instruction from the surface: a 'launch and forget' science probe (or, 'launch and worry about' in reality, as they are rather expensive).

TECH TROUBLE

It has only recently become feasible to use AUVs in marine science, as their endurance and reliability have slowly increased. Using autonomous vehicles under ice presents several new problems, however, since the standard 'panic solution' during open-ocean missions is for the vehicle to surface, get a GPS fix and send off a text message with its position--none of which are possible under ice.

Instead, it's best simply not to have a problem in this environment. Although the vehicle can be located acoustically, going and drilling it out is, at best, time-consuming and, at worst, impossible. …