Australia: The Land Where Time Began

A biography of the Australian continent 

Antarctic Dry Valleys Formation of Thermokarst in the McMurdo Dry Valleys

Thermokarst is a land surface that has been disrupted and lowered by the melting of ground ice (permafrost). Thermokarst is known to be a major driver of changing landscape in the Artic, though in Antarctica is has been considered to be a minor process. In this study Levy et al. used ground-based and airborne LiDAR coupled with timelapse imaging and meteorological data to demonstrate that:

(1)   The formation of Thermokarst has accelerated in Garwood Valley, Antarctica;

(2)   The Thermokarst formation rate is at present about 10 times the average rate in the Holocene; and

(3)   Increasing insolation and sediment/albedo feedbacks are the strongest drivers of the increased rate of formation of Thermokarst.

It is suggested by this that sediment enhancement of melting that is driven by insolation may act in a similar manner to increasing Antarctic air temperature, which is occurring along the Antarctic peninsula at present, and may serve as a leading indicator of imminent changes of landscape in Antarctica that will generate Thermokarst landscapes which are similar to those in periglacial terrains in the arctic.

It has been uncertain whether the thermokarst landforms in Antarctica are in equilibrium (Schenk et al., 2004; Marchant & Head, 2007), or whether the development of thermokarst is being accelerated by changes to modern thermal boundary conditions (Levy et al., 2013; Denton & Marchant, 2000; Campbell & Claridge, 2003; Shaw & Healy, 1977; Bockheim, 1995; 1995; Swanger & Marchant, 2007; Schenk et al., 2004; Marchant & Head, 2007), though thermokarst landforms in Antarctica have previously been mapped, including within the McMurdo Dry Valleys  (MDV) (Levy et al., 2013; Denton & Marchant, 2000; Campbell & Claridge, 2003; Shaw & Healy, 1977; Bockheim, 1995; 1995; Swanger & Marchant, 2007). Garwood Valley (78oS, 164oE), is a coastal valley in the McMurdo Dry Valleys of southern Victoria Land, Antarctica is a natural laboratory where competing models of the erosion of Antarctic erosion can be tested. The valley is located in a zone of continuous permafrost, having a mean annual temperature of -16.9oC, and is partly filled with a remnant of the Ross Sea Ice Sheet, an ice mass covered with debris that lodged in the valley during the Pleistocene (Levy et al., 2013; Denton & Marchant, 2000; Campbell & Claridge, 2003; Shaw & Healy, 1977; Bockheim, 1995; 1995; Swanger & Marchant, 2007; Healy, 1975; Stuiver et al., 1981).  The remnant debris-covered ice extends up the valley for about 7 km from the modern coast of the Ross Sea, draping surfaces that range in elevation from sea level to about 200 m above sea level. The ablation till that overlies the buried ice is typically about 10-20 cm thick, though it thickens in places to several metres (Levy et al., 2013; Denton & Marchant, 2000; Campbell & Claridge, 2003; Shaw & Healy, 1977; Bockheim, 1995; 1995; Swanger & Marchant, 2007; Healy, 1975; Stuiver et al., 1981; Pollard, Doran & Wharton, 2002). The till has been subdivided into older, up-valley and younger down-valley units, both dating to the Pleistocene (Levy et al., 2013), and both are composed largely of a sand-silt matrix that is overlain by pebble, cobble, and boulder desert pavement (Levy et al., 2013; Pollard, Doran & Wharton, 2002). The buried remnant of the Ross Sea Ice Sheet in the centre of the valley and at the ice cliff study site is overlain by thin glacial till and capped by fluvio-deltaic sediments that were deposited into Lake Howard, the Palaeolake dating to the Pleistocene to Holocene (Levy et al., 2013; Healy, 1975; Pw, 1960), a small lake of about 0.7 km2 that formed within and atop the Ross Sea Ice Sheet ice dam (Levy et al., 2013). In the Garwood Valley the active layer is typically about 20 cm thick, and the base of the active layer is typically marked by the sharp contact between the buried ice and the overlying till (Levy et al., 2013; Pollard, 1960). In thin tills the base of the active layer is typically wet, 5-10 % water by volume, and in thicker tills that do not contact the buried ice is typically dry.

As the buried ice mass in Garwood Valley ablates (Levy et al., 2013) a variety of remarkable thermokarst landforms have been produced which include thermokarst ponds (Pollard, Doran & Wharton, 2002; Shindell, 2004; Arblaster & Meehl, 2006; Chapman & Walsh, 2007), tunnels and thermokarst dolines that have been eroded through the buried ice by the modern Garwood River (Healy, 1975; Lewkowicz, 1987; Lewkowicz, 1986; Burn & Lewkowicz, 1990; Lantuit, 2012; Lantuit & Pollard, 2005), and a large retrogressive thaw feature that is referred to as the Garwood Valley ice cliff (Levy, et al., 2013; Marchant & Head, 2007). Runoffs from the Garwood and Joyce glaciers largely source the Garwood River, flowing about 13 km down-valley from the toe of the Garwood Glacier to the coast of the Ross Sea. At its deepest point the river is less than 50 cm deep. The Garwood River incises the floor of the valley to a depth of about 4 m, and in the steep, V-shaped river channel walls the buried ice is commonly observable  (Shindell, 2004: Healy, 1975).

The ice cliff is at a place where the remnant ice of the Ross Sea Ice Sheet 10-15 m high is exposed at a precipitous break in slope adjacent to the braid plain of the modern Garwood River (Levy, 2013). The main channel of the Garwood River meanders about 1-2 m over about 10 m wavelengths in the vicinity of the ice cliff, undercutting the exposed cliff base in places, and flowing in a sandy channel several metres from the cliff in other places. The cliff face has been moved back towards the south by ablation since 2009 relative to the river channel location. The Garwood River becomes braided, sometimes filling additional small channels to the north of the main channel (away from the ice cliff), at times of high flow.

At the Garwood ice cliff the exposed ice is geochemically identical to the massive ice that is buried by the down valley till, having major ions, δ18 and δD values that are within the range of the variability of the Ross Sea Ice Sheet (Levy, 2013; Pw, 1960). The ice of the ice cliff is capped by about 2 m if glacial till and fluvial sediments that are ice cemented, the upper about 20 cm thawing seasonally. It has an aspect that is generally north-facing, though there are portions of the ice cliff that face east and west as well.

The Garwood Valley ice cliff is indicated to be backwasting rapidly by field observations (Levy, 2013; Pw, 1960). Contrasting with this it is inferred by Pollard et al. (Levy et al., 2013; Pollard, Doran & Wharton, 2002) that the buried ice situated between the ice cliff and the mouth of the Garwood Valley was in equilibrium with modern climate conditions as the thermokarst ponds are largely confined to the mouth of the valley though they did note that small ground temperature changes could lead to thermokarst formation that was more widespread in the valley.

According to Levy et al. there are 2 primary hypotheses that have been proposed to account for the formation of thermokarst landforms in the MDV:

1)      Small summer temperature increases which result in small increases in the thickness of the active layer that drives deeper melting (Levy et al., 2013; Pollard, Doran & Wharton, 2002) and

2)      Warm meltwater advects heat to buried ice by flow from the surface to the subsurface (Healy, 1975; Lantuit & Pollard, 2008; Lantuit & Pollard, 2009; Niu et al., 2012).

Levy et al. suggest it is not likely that other geological heat sources contribute to the melting of ground ice in the Garwood Valley: it is indicated by mapping (Levy et al., 2013) that there are no active fault-directed hydrothermal features or volcanic features present in the Garwood Valley, and there is generally low geothermal heat flux in the MDV, about 50 mW/m2, (Doran et al., 2002; 24). To test these hypotheses of thermokarst generation, as well as to suggest new hypotheses to test, Levy et al. established an observation system at the Garwood Valley ice cliff, a timelapse camera system, and biannual terrestrial laser scanning (ground-based LiDar) of the ice cliff and vicinity. Air temperature and humidity were measured by the continuous monitoring station, temperature of the ice cliff by infrared radiometry, the range to the ice cliff by ultrasonic distance sensor, wind speed and direction, and longwave and shortwave radiation balance. The orientation of the radiation sensors set up in such a way that they monitor the radiation fluxes into and out of the ice cliff, instead of the traditional incoming (upward-facing) and outgoing (downward-facing) directions. At night the continuous monitoring station is shadowed because the Sun is placed behind the south valley wall by the low angle of the Sun from about 20:00 to 06:00; during these hours the ice cliff and station are illuminated only by diffuse insolation.


The Garwood Valley ice cliff, which is a thermokarst feature of coastal Antarctica, has been found to be losing volume estimated to be at a rate that approaches 10 times its average ablation rate of the Late Holocene. In each of 3 measurement windows the ice cliff ablation rate has increased: 2001-2010, November 2010 to January 2011, and January 2011 to January 2012. The current rate of ablation in the Garwood Valley exceeds the ablation rates of McMurdo Dry Valleys glaciers by a factor of 5 (Bliss, Cuffey & Kavanaugh, 2011), which represents a major departure from typical rates of land surface changes in the Ross Sea sector of Antarctica. Levy et al. interpreted melting driven by insolation to be the main driver for the ice loss in Garwood Valley, the melting being the result of insolation coupled with an albedo-feedback in which the albedo of the ice surface is lowered by the debris that had fallen onto the ice cliff leading to an increase  in conductive heat transfer.

Additional buried ice deposits may be destabilised by future warming of the atmosphere and the ground in the Dry Valleys (Fountain, Pettersson & Levy), which when exposed at breaks in topography would be likely to experience similar sediment/melting feedback to that that has been observed at Garwood Valley. Levy et al. suggest that such ice exposures would experience rapid ablation that is greater than has been predicted by conductive thermal models alone (Pollard, Doran and Wharton, 2002), which may result in retrogressive thaw slumps as well as other thermokarst landforms which exhibit polycyclic periods of rapid melting as a response to variations in the exposure of ice and sediment burial (Shindell, 2002). This process is not unique to the Dry Valleys, as it is similar to interactions between ice and sediment observed on non-polar glaciers that are covered by debris in which a thick mantle of debris isolates the subsurface ice, which allows melt to take place at locations where the ice is not covered where melting can be driven by sensible heat and solar radiation (Benn et al., 2012). Levy et al. suggest the recession of the Garwood Valley ice cliff may be an indicator of landscape change that is more widespread than is expected to transform areas of low elevation and landscapes of the Antarctic coast by the close of the century. The formation of degraded permafrost landforms in Antarctica that is similar to Arctic cold desert permafrost environments (Lewkowicz, 1987; Lewkowicz, 1986; Burn & Lewkowicz, 1990; Lantuit et al., 2012; Lantuit & Pollard, 2008), may result from the acceleration of thermokarst formation in Antarctica by climate warming in the future (Shindell, 2004; Arblaster & Meehl, 2006; Chapman & Walsh, 2007).

Sources & Further reading

  1. Levy, J. S., A. G. Fountain, J. L. Dickson, J. W. Head, M. Okal, D. R. Marchant and J. Watters (2013). "Accelerated thermokarst formation in the McMurdo Dry Valleys, Antarctica." Scientific Reports 3: 2269.


Author: M. H. Monroe
Last updated 
Journey Back Through Time
Experience Australia
Aboriginal Australia
National Parks
Photo Galleries
Site Map
                                                                                           Author: M.H.Monroe  Email:     Sources & Further reading