South Georgia Glaciers

By Richard Alley

Among its many charms, the island of South Georgia is known for glaciers.  Over half of the island is covered in “permanent” snow and ice (more on this soon!), especially on the southwest side facing the wind, but with spectacular glaciers extending down the leeward northeast side.  The recent scholarly summary by Gordon et al. (2008) provides an outstanding overview, and I’ll borrow heavily from it, as well as from Cook et al. (2010), for recent changes in the glaciers.


The British station King Edward Point, South Georgia Island, is the cluster of red-roofed buildings in the center of this photo. The abandoned Norwegian whaling Grytviken Station is seen in the far distance. South Georgia Island is governed by Great Britain. Ernest Shackleton is buried here.
Photograph Credit: Julian Race/National Science Foundation
Date Taken: September 27, 2011

Weather data are available more-or-less continually from King Edward Point, near Grytviken, starting in 1906 and maintained by the stalwart crew of the British Antarctic Survey.  This relatively sheltered low-elevation site on the northeast coast had a year-round mean temperature of 2.0 C from 1951-1980, with the three summer months averaging 4.8 C, and wintertime “plunging” to -1.2 C (36 F mean, 30 F winter and 41 F summer).  For those of us who live in continental climates, this may sound like maritime boredom. But, the storms of the south make the more-exposed and higher sites much more interesting, with windier, colder and snowier conditions, while föhn winds blowing down the northeast coast can push temperatures above 20 C (68 F).

Precipitation down at King Edward Point averaged 1.6 m of water per year (62 inches) from 1951-1980.  Snow persists year-round and glaciers have accumulation zones above about 300 m elevation facing the winds, but above 450-600 m on the leeward side (roughly 1000 feet, rising to 1500-2000 feet).

The island in the past was heavily glaciated, with ice extending to the edge of the continental shelf, leaving moraines, drumlins and other features that now are below sea level but clearly visible with good sonar.  Ages of these features are poorly constrained but likely date to multiple ice ages including the most recent one (Graham et al., 2008), which peaked roughly 24,000 years ago.  The oldest moraines near the modern coast (Bentley et al., 2007) then may represent the short-lived Antarctic Cold Reversal roughly 13,000 years ago during the overall warming from the ice age, although again with some uncertainty in the ages.  The ice has remained behind those coastal moraines since then, with some interesting wiggles in response to smaller climate changes.

Permanent ice masses today include numerous small cirque glaciers and others ending well inland, many glaciers that make it to the coast but without advancing into the sea to calve, and some that calve icebergs in fjords.  Not surprisingly, the glaciers that tap snowfall from larger drainage areas extending higher up the mountains tend to have more ice and extend farther toward or into the sea.

Based on historical maps, satellite images, and other data, Gordon et al. (2008) traced the history of the glaciers.  Of an available sample of roughly 160 glaciers on the island, they mapped ice-front positions of 36, reaching as far back as the German First International Polar Year Expedition of 1882 but with most of the data much more recent.  They found from the more-recent parts of the records that 28 are retreating, 6 stable, and 2 advancing.  Cook et al. (2010) extended the data set to 103 coastal glaciers, looking at the 1950s to “present” (shortly before the 2010 publication), and found that 97% had retreated over the period of observations.  Those glaciers with the most-extensive high-altitude accumulation-zones have been least affected; those with long tongues at low elevation, or in low cirques, have been especially prone to retreat that may even approach disappearance.   This points clearly to warming as the culprit, melting ice in low-altitude regions.

Indeed, the temperature data show a general warming trend since about 1930, which was a relatively cold time that is marked by small moraines in front of many glaciers indicating a small, short-lived readvance).  Precipitation data show much variability but a slight upward trend paralleling the warming.  As discussed in the previous post, rising temperature and rising precipitation generally lead to smaller glaciers, as observed.

Furthermore, as summarized by Gordon et al., the South Georgia glacier changes are broadly matched by those on other subantarctic islands including Heard Island, Kerguelen, and the South Shetlands, in the Antarctic Peninsula, and in southern South America.  And, I might add, in much of the world.  One cannot blame the changes on some shift in storm tracks, because so many areas are affected, nor can one blame soot from fires because there is no relation between melt and proximity to soot.  Warming is responsible, with high confidence.

Not every glacier is advancing, and it is interesting to look at the exceptions.  For the calving Harker and Novosilski glaciers, Gordon et al point out that geographic factors that give them large accumulation zones and relatively less area at low elevation favor advance.  I have not studied these glaciers, but they might also be exhibiting the tidewater-glacier behavior so well known from Alaska and elsewhere.

A heavily crevassed, warm glacier has difficulty forming a floating ice shelf, and so must calve icebergs where it reaches deep water and starts to float.  Suppose that such a glacier reaches the sea and calves icebergs, but could extend many kilometers down its fjord if calving stopped.  If the glacier is transporting enough sediment for long enough, it will build a moraine shoal so that the glacier terminates in shallower water, reducing calving.  The glacier will then advance down the fjord behind its “shield” of sediment that keeps the deep water away, recycling the sediment in the fjord and adding more to it.  By keeping the sediment near the front, the fjord behind the moraine shoal remains deep but filled with ice.

After advancing many kilometers down the fjord, the glacier may nearly stabilize.  But, suppose that a small warming then melts enough ice to shorten the glacier by one kilometer.  The glacier starts to retreat, but finds its terminus in deep water behind the moraine shoal.  Icebergs break off, and usually they can get over the moraine shoal or melt in the sea water flooding into the growing fjord.  The front of the glacier retreats rapidly.  And, it cannot “stop” at its equilibrium point after retreating one kilometer—it calves away until retreating many kilometers to the shallow water at the head of the fjord.  Then, it begins building a new moraine shoal and advancing to the equilibrium length.  If you happen to observe this glacier during the retreat, you may be greatly alarmed when you see not a small, slow retreat but a huge, rapid one.  If you first observe this glacier just after the retreat, you will see it advancing while neighboring glaciers retreat, and you may wonder what the heck is going on.  I don’t know how closely this explains the recent behavior of Harker and Novosilski Glaciers, but it is such a well-known phenomenon that it likely does apply at least partly—the advance you are seeing may really be retreat with a long detour.

There is still lots of spectacular ice on South Georgia, with a fascinating history that is still being sorted out.  Have fun!

Bentley, M.J., D.J.A. Evans, C.J. Fogwill, J.D. Hanson, D.E. Sugden and P.W. Kubik, 2007, Glacial geomorphology and chronology of deglaciation, South Georgia, sub-Antarctic, Quaternary Science Reviews 26, 644-677.

Cook, A.J., S. Poncet, A.P.R. Cooper, D.J. Herbert and D. Christie, 2010, Glacier retreat on South Georgia and implications for the spread of rats, Antarctic Science 22, 255-263.

Gordon, J.E., V.M. Haynes and A. Hubbard, 2008, Recent glacier changes and climate trends on South Georgia.  Global and Planetary Change 60, 72-84.

Graham, A.G.C., P.T. Fretwell, R.D. Larter, D.A. Hodgson, C.K. Wilson, A.J. Tate and P. Morris, 2008, A new bathymetric compilation highlighting extensive paleo-ice sheet drainage on the continental shelf, South Georgia, sub-Antarctica, Geochemistry Geophysics Geosystems 9(7) Q07011.