Product has been added to the basket

Sunless seas

The Laybrinth, Upper Wright Valley, Antarctic Dry Valleys in East Antarctica. Deep (>100 m) channels incised into hard bedrock by the action of massive outburst flooding from one or more subglacial lakes. . Courtesy, David Sugden and George Denton.


Martin J Siegert* reports on the exploration of subglacial Lake Ellsworth in West Antarctica

Geoscientist 17.10 October 2007

Ever since the discovery, ten years ago, of Lake Vostok as a large deep freshwater body beneath the Antarctic ice sheet, scientists have been fascinated by the potential of subglacial lakes as habitats for unusual microbial life and recorders of unique climate records. While several plans exist to penetrate into and explore these environments, none is more advanced that a UK-led project to investigate Lake Ellsworth (West Antarctica).

The notion of substantial volumes of liquid water in Antarctica seems incongruous when one considers that the coldest temperature on Earth, -89°C, was recorded on 21 July 1983 at Vostok Station in central East Antarctica, and that temperatures here are regularly around -60°C. However, geothermal heating from the Earth’s interior, at a normal level, is sufficient to keep the base of the 3.7km thick ice sheet at Vostok Station warm, and permit the existence of Lake Vostok; a body of water, the size of Lake Ontario1. In fact, the Antarctic ice-sheet base is warm in many places. Where subglacial water forms, it flows under gravity and pressure of ice into topographic hollows, where it ponds to form lakes of various sizes. While Lake Vostok is the best known and largest lake in Antarctica, glaciologists have discovered over 150 of them scattered widely across the continent2 (Figure 1).

Figure 1. The locations of Antarctic subglacial lakes. Different colours refer to the nation whose geophysical survey was responsible for lake identification. Lake Ellsworth is circled in red.

Subglacial lakes can be identified using ice-penetrating radar. VHF radio-waves propagate well in cold ice and reflect of boundaries where there is a change in electrical properties. While the most significant boundary occurs at the ice-sheet base, the nature of a radar reflection from a subglacial lake (e.g. a flat, bright echo of constant strength due to the uniform reflections off a mirror-like interface) is distinct from an ice-rock contact (e.g. an undulating, weak echo of variable strength due to scattering at a rough interface; Figure 2). While radar is a good method to determine the existence and location of subglacial lakes, it cannot be used to measure water depths as radio-waves are absorbed in water. Instead, seismic sounding is needed. No lake other than Lake Vostok has been measured by seismic sounding (revealing in one place a water depth of over 500m). Hence, the depths of subglacial lakes, and therefore the volume of water stored beneath the ice sheet, have yet to be calculated with certainty.

The under-ice landscape of Antarctica involves a series of well defined mountains and highlands, as well and extensive regions of troughs and lowlands. Much of the landscape has been eroded by the action of ice sheets, which have grown and decayed in Antarctica since the earliest Oligocene, through the Miocene and, possibly, Pliocene, although many believe the ice sheet may have been stable in the last few million years. During the Pleistocene the ice-sheet has certainly been very stable, and rests on a bed developed by these former dynamic ice sheets. Lake Vostok, which is over 250km long, 80km wide and has an estimated water volume of over 5000km3, sits in a basin that probably has a pre-glacial tectonic origin. However, this basin was probably also scoured out by the action of past ice sheets over the last 34 million years. Several other lakes fill entire subglacial basins of varying sizes. For example, one lake in West Antarctica, named Lake Ellsworth (Figure 2), is housed in a former fjord at the foothills of the Ellsworth Mountains3. By contrast to Lake Vostok, Lake Ellsworth is only 10km long and 3km wide.

Figure 2. Radio-echo sounding transect across the foothills of the Ellsworth Mountains in West Antarctica, revealing the surface of Lake Ellsworth within a 1-2 km deep trough (probably a former fjord).

Deep-water subglacial lakes have inspired biologists to plan their exploration, as unique extreme environments in which unusual micro-organisms might adapt and survive. While this aspect of subglacial lakes research has been given considerable scientific and media attention, we should not underestimate the importance of the two-fold geoscientific rationale behind the future exploration of subglacial lakes.

First, subglacial lakes are highly likely to contain sediments across their floors, which will have accumulated for as long as the lakes have been in existence. For deep-water lakes, this could be as old as the present-day ice sheet, which is of unknown age. In fact the glacial history of the Antarctic ice sheet throughout the Cenozoic is not known well. Sediment records from the floors of subglacial lakes may be able to provide important information concerning past ice sheet changes.

Second, subglacial lakes are evidence of widespread water beneath the ice sheet, which is likely to have an important, albeit poorly known, influence on ice sheet flow dynamics. Recent satellite remote sensing investigations have revealed that subglacial lakes can periodically discharge large volumes of water across several hundred kilometres (Figure 3). In one case, a flux of water of the order of the flow of the River Thames in London was calculated between lakes in central East Antarctica separated by 300km4. The discharge was of the order of 1.8 km3 and lasted for 16 months during 1996-7. As several subglacial lakes are known to exist at the heads of ice streams2 (the fast flowing rivers of ice that drain the bulk of the ice sheet), changes to the basal conditions of these systems caused by subglacial lake drainage could have serious implications for ice sheet stability and, possibly, global sea-level change. Evidence certainly exists in the morphological record of the ice sheet margin for huge outburst events that left water channels cut >100m deep into hard rock (Plate 1, top). It is therefore likely that substantial volumes of subglacial lake water have previously reached the ocean, and could do so again under the glacial processes identified in East Antarctica.

Figure 3. Sudden flow of subglacial lake water beneath the central East Antarctic ice sheet. Satellite altimetry detects (1) ice-sheet surface elevation lowering of around 3 m, signalling the loss of water from a lake 3.5 km beneath.


While there are clearly good and exiting reasons to explore subglacial lakes, to do so is challenging from technical, logistical and environmental points of view. As subglacial lakes are pristine environments, and as the levels of life and nutrients are likely to be low, direct examination of these systems must be undertaken in ultra-clean conditions. In other words, contamination of the lake by the access technique must be avoided, and equipment used to measure the lake must be sterile. Hot water drilling offers the best opportunity for efficient clean lake access, as opposed to ice coring that uses an antifreeze drilling fluid or thermo-probing which is notoriously unreliable over even short (<100m) vertical distances. Hot water drilling would introduce fluid to a subglacial lake, but only as melted ice, which is what subglacial lakes are made of. Providing that melted ice used for hot water originates from before the industrial revolution, such inputs could not constitute a contamination, as such ice melts into the lake naturally.

Once lake access is achieved, instruments can be lowered down the borehole and into the water column. While high-technology autonomous vehicles have been discussed for use in subglacial lakes, in the instance of first lake access the instruments are likely to be as simple as possible and housed in a probe tethered to the ice surface (Figure 4). The probe would be used to identify the broad physical, chemical and biological properties of the lake water and sediments. Once first access has proved successful, and the results shown to be of scientific value, the sophistication of the measuring techniques in subsequent experiments may justifiably increase.

Given that over 150 subglacial lakes are known to exist, which is best suited for exploration? Six criteria have been developed, which can be used to judge a subglacial lake’s suitability for exploration. They are, in essence, whether a lake: (1) provides the greatest likelihood for attaining the scientific goals; (2) can be characterised in a meaningful way; (3) is representative of other lakes and settings; (4) is located within a setting that is well defined; (5) is accessible; and (6) can be explored within reasonable cost and logistical constraints. Clearly, large lakes such as Lake Vostok, which would take decades to characterise to the level needed to comprehend the hydrological system fully, are least appropriate for exploration in the first case. Smaller lakes, which can be measured to the required level efficiently, and which are located near to existing infrastructure and logistics, are well suited to exploration.

Figure 4. The exploration of subglacial lake Ellsworth. Hot-water drilling is used to access the lake. The hole is likely to be kept open for 24-36 hours, during which time a sterile probe will be released into the lake.

UK team

A UK-led team, involving over 30 scientists from 15 universities and research institutes, and indeed scientists from four other nations, have identified Lake Ellsworth as a prime candidate for exploration5. Lake Ellsworth is located within good logistical coverage in West Antarctica, which offers it three advantages over East Antarctic subglacial lakes. First, there is precedent for accessing the subglacial environment in West Antarctica. Glaciologists have measured and sampled the base of West Antarctic ice streams on numerous occasions using hot-water drilling. Hence both the technique and the logistical arrangements are known to be appropriate in this region.

Second, concerns the safety of working in West as opposed to East Antarctica. The ice-sheet surface elevation in West Antarctica is over 1 km lower than in East Antarctica. Consequently, altitude-sickness problems that researchers can face at Vostok Station (~3.5 m a.s.l.) and Dome A (~4.2 km a.s.l.) – medical evacuations have been needed at both locations in the past few years – are far less of an issue at the site of Lake Ellsworth (~2 km a.s.l.), for example.

Third, the nature of the science that can be achieved in West Antarctica is potentially of great significance from a geoscience perspective. One of the most important debates in glaciology concerns the stability of the West Antarctic Ice Sheet. Many believe it is subject to rapid large-scale fluctuations in size, which has huge implications for global sea level (if all the ice in West Antarctica melted, sea level would rise by ~6m). In contrast, East Antarctica is thought to be far more stable. Gauging the risk requires knowledge of when the ice sheet last broke up, and the associated environmental conditions. Records of such an event have proved difficult to obtain, as they are masked by the overriding ice sheet of the last glacial cycle. However, sediments on the floor of a West Antarctic subglacial lake may provide a unique record of West Antarctic ice-sheet history that can be used to quantify the present day risk of collapse.


Subglacial lake research is in its infancy. While outline plans to explore subglacial lakes exist, no lakes are likely to be explored purposefully within the next 5 years. Russian scientists plan to use the existing ice core to drill into Lake Vostok, but only to retrieve frozen lake water rather than introduce a probe to sample and measure the lake’s water column. This controversial plan, debated widely by members of the Antarctic Treaty in June 2003, relies on ice coring as a means to extract ice close to the melting temperatures. The core was reactivated in 2006/7, but several problems have been experienced since. Core retrieval has proved difficult, and a drill bit was lost at one stage. As the core penetrates deeper, and the ice gets warmer, it is likely that the experiment will become more challenging, which could make the utility of ice coring for this purpose questionable. If the Russian team is successful, however, frozen Lake Vostok water would be extracted in 2007/8.

At the same time, UK scientists will travel to the centre of West Antarctica to perform a geophysical survey of Lake Ellsworth. The results, forming the most comprehensive understanding of a subglacial lake ever made, are eagerly awaited by the Ellsworth Consortium as they will determine whether a full proposal to explore the lake is warranted. If it is, the exploration of Lake Ellsworth ( could take place, at the earliest, in 2012/13 – 100 years since the ‘golden age’ of Antarctic exploration and, of course, the centenary of Scott's expedition to South Pole!


  1. Kapitsa, A., Ridley, J.K., Robin, G. de Q., Siegert, M.J. & Zotikov, I. Large deep freshwater lake beneath the ice of central East Antarctica. Nature, 381, 684-686. (1996).
  2. Siegert, M.J., Carter, S., Tabacco, I., Popov, S. and Blankenship, D. A revised inventory of Antarctic subglacial lakes. Antarctic Science, 17 (3), 453-460. (2005).
  3. Siegert M.J., Hindmarsh, R., Corr H., Smith, A., Woodward, J., King, E., Payne, A.J., Joughin, I. Subglacial Lake Ellsworth: a candidate for in situ exploration in West Antarctica. Geophysical Research Letters, 31 (23), L23403, 10.1029/2004GL021477. (2004).
  4. Wingham, D.J., Siegert, M.J., Shepherd, A.P. and Muir, A.S. Rapid discharge connects Antarctic subglacial lakes. Nature, 440, 1033-1036 (2006).
  5. Siegert, M.J. and the Lake Ellsworth Consortium. The exploration of Ellsworth Subglacial Lake: a concept paper on the development, organisation and execution of an experiment to explore, measure and sample the environment of a West Antarctic subglacial lake. Reviews in Environmental Science and Bio/Technology. doi: 10.1007/s11157-006-9109-9 (2006).

* School of GeoSciences, University of Edinburgh, Grant Institute, West Mains Road, Edinburgh EH9 3JW, UK