What links the retreat of Jakobshavn Isbrae, Wilkins Ice Shelf and the Petermann Glacier?

Guest commentary from Mauri Pelto

Changes occurring in marine terminating outlet glaciers of the Greenland Ice Sheet and ice shelves fringing the Antarctic Peninsula have altered our sense of the possible rate of response of large ice sheet-ice shelf systems. There is a shared mechanism at work that has emerged from the detailed observations of a number of researchers, that is the key to the onset and progression of the ice retreat. This mechanism is shared despite the vastly different nature of the environments of Jakobshavns Isbrae, Wilkins Ice Shelf and the Petermann Glacier.

We reviewed in a previous post the first mechanism for explaining the change in velocity of Greenland’s large outlet glacier – the Zwally effect – and why it is not the key. This mechanism relies on meltwater reaching the glacier base via moulins and reducing the friction at the base of the glacier. This idea was observed to be the cause of a brief seasonal acceleration of 10- 20 % on the Jakobshavns Glacier in 1998 and 1999 at Swiss Camp 35 km inland from the calving front (Zwally et al., 2002). Examination of recent rapid supraglacial (i.e. on the surface) lake drainage documented short term velocity changes due to such events around 10%, but little significance to the annual flow of the large glaciers outlet glaciers (Das et.al, 2008).

The second mechanism is a dynamic thinning of the terminus zone of the marine terminating outlet glacier reducing the effective bed pressure, allowing acceleration – the Jakobshavn effect. The reduced resistive force at the calving front due to the thinner ice, now experiencing greater flotation, is then propagated “up glacier” (Hughes, 1986; Thomas, 2003 and 2004). If the Jakobshavn effect is the key the velocity increase will propagate up-glacier, there will be no seasonal cycle, and thinning and acceleration would be greatest near the terminus.

That the thinning and acceleration is greatest for marine terminating outlet glaciers has indeed been demonstrated by Sole et. al. (2008). That acceleration began at the calving front and spread upglacier 20 km in 1997 and up to 55 km inland by 2003 (Joughin et al., 2004). On Helheim the thinning and velocity propagated up-glacier from the calving front. Each of the glaciers fronts did respond to tidal variations indicating they had started floating, detached from their bed (Hamilton et al, 2006). This summer, Jason Box and others at Ohio State University observed that Jakobhavns Isbrae retreated again, losing 15 km², and maintaining an accelerated pace from the northern branch of the ice stream as opposed to the greater retreat and acceleration of the southern branch 2001-2005 (Box, 2008). This was accompanied by the second consecutive year of substantial retreat of the glacier just north of Jakobshavn, Sermeq Avannarleq which had been quite stable for much of the last century (Box , 2008b). Sole et. al. (2008) also noted that the recent thinning and acceleration was not limited to just the now more famous Helheim, Jakobshavn and Kangderlugssuaq Glaciers, but included Rinks Isbrae, Equaluit, Cristian IV and all others they observed. Note the greater flow of the southern ice stream in 2000, compare to the northern ice stream in this image from Ian Joughin:

Petermann Glacier is a much different glacier than the others mentioned above. Its velocity of 2-3 m/day (Higgins, 1990) is much lower than 10-30 m/day observed on the other marine terminating outlet glaciers. It is located on the northwest corner of Greenland and certainly experiences less melting and less snowfall. The lower 80 km (in length) and 1300 km² (in area) of the glacier is afloat. This makes it (by area) the largest floating glacier in the Northern Hemisphere. The ice front is not impressive,unlike the faster outlet glaciers. The calving front protrudes a mere 5-10 m above sea level, reflecting the fact that the ice at the front is only 60-70 m thick. Further up-glacier, the ice at the grounding line is 600-700 m thick. The combination of velocity and thickness yield the volume of material calved each year. Petermann Glacier calves 0.6 km³ (Higgins, 1990), whereas Jakobshavns yields close to 40 km³. The thinning between the grounding line and the calving front is mainly via melting as the snowline is at 900 m. The low slope leads to very low velocities, giving the low-lying floating section plenty of time to melt, and surface melt ponds are common.

The Petermann Glacier lost a substantial area, 29 km² due to calving this summer (Box 2008c), and a crack well back of the calving front indicates another 150 km² is in danger. The volume of the ice lost is much less than that from the loss of a comparable area by Jakobshavn because the ice is an order of magnitude thinner. Once again the key to this glacier’s second major ice loss this decade after limited retreat in the last century, is thinning of the floating tongue, which weakens the glacier. The loss of this ice should then lead to acceleration, developing more crevassing and glacier retreat. The crack seen in the image of Petermann Glacier (ASTER image provided by Ian Howat of Ohio State) is more of a rift, like those on Larsen Ice Shelf, than a crevasse. This transverse rift is further connected to longitudinal-marginal rifts. Illustrating the poor connection of the Petermann Glacier to its margin and lack of a stabilizing force this margin has, even 15 km behind the calving front. This is not the only rift of its kind on the glacier. Also note that like on Larsen Ice Shelf the rift crosscuts surface streams.

A series of Landsat images from 2002, 2006 and 2007 illustrate the shift in the terminus and in the position of key rifts A, B and C. The distance back from the terminus has diminished for A and B from 2002 to 2007. In 2006 to 2007 the shift in the position of C is also evident.

As in the case on Jakobshavns, Helheim and others the key is the pre-conditioning phase of thinning, that leads to more calving, that leads to more acceleration, and that generates retreat. In a recent paper in press in the Journal of Glaciology Ian Howat and others examined changes in terminus position, surface elevation and flow on 32 glaciers along the southeast coast of Greenland from 200-2006. Their key conclusion was that the

… ratio of retreat to the along-flow stress-coupling length is proportional to the relative increase in speed, consistent with typical ice flow and sliding laws. This affirms that speedup results from loss of resistive stress at the front during retreat, which leads to along-flow stress transfer. Many retreats began with an increase in thinning rates near the front in the summer of 2003, a year of record high coastal-air and sea-surface temperatures.

This indicates again the importance of pre-conditioned thinning via melting.

Wilkins Ice Shelf (WIS) refused to hibernate this winter. A previous post noted that the recent collapse of Wordie Ice Shelf, Mueller Ice Shelf, Jones Ice Shelf, Larsen-A and Larsen-B Ice Shelf on the Antarctic Peninsula has made us aware of how dynamic ice shelf systems are.

The reasons for Ice Shelf collapse continue to be identified, but one key thread emerges. The decade prior to collapse the Larsen-B Ice Shelf had thinned primarily by melting of the ice shelf bottom (by the ocean) by 18 m (Shepard and others, 2003). Thinning preconditions the ice shelf for failure by weakening its connection to pinning points at the grounding line as the shelf becomes more buoyant. Glasser and Scambos (2008) observed that prior to collapse that rifts and crevasses parallel to the ice front crosscut the meltwater channels and ponds, hence, post dated them. The number and length of the rifts increased markedly in the year before collapse. There was no evidence of relict rifts, illustrating that these rifts are a feature of the last 20 years. After ice shelf collapse the ice front receded to the pre-existing rifts, and the pre-existing rifts defined the area of collapse. In this case the thinning and resultant structural weaknesses preconditioned the ice to rapid breakup, which proceeded to lose only the preconditioned portion of the ice shelf.

The WIS is buttressed by Alexander, Latady, Charcot and Rothschild islands and by numerous small ice rises, indicating that they are touching the ocean floor. WIS was examined by Braun, Humbert and Moll (2008). They found that drainage of melt ponds into crevasses were of no relevance for the break-up at WIS. On WIS the evolution of failure zones is associated with ice rises. In 1993/94, rift formation started to expand at the northern ice front. Today, the central part of WIS is intersected by long rifts formed in and around ice rises. The rifts up to tens of kilometers long evolve and coalesce prior to break-up events. The conclusion for WIS is that preconditioning of the ice shelf by connection of the rifts in the failure zones near ice rises trigger break-up events. The thinning and rifting lead to a cascade of failure.

The Feb.-April break-up left a narrow 6 km wide fractured connection to Charcot Island. Existing rifts formed between already existing fractures, crossed almost the entire northern shelf. This fragile and vulnerable area was expected to collapse further the next austral summer. However, it instead has happened this austral winter with loss of an additional 160 km² of ice. It is the first winter ice loss of an ice shelf ever observed, and so was surprising. However, looking at the image below, from the European Space Agency showing the extent of the rifts as winter began, makes this less surprising. The question is more what can possibly hold this together? The area of interconnected rifts seen is 2000 km². If this is lost an additional 3000 km² of the 13 000km² of WIS, is at risk when this connection to Charcot Island is lost (Braun, Humbert and Moll, 2008).

It appears then that glacier or ice shelf thinning is the key preconditioning factor for collapse, retreat and acceleration, whether you are in Antarctica of Greenland. The mechanisms for ice shelf thinning include basal melting (from warming ocean waters), surface melting, reduction in glacier inflow and rift development. These are interrelated mechanisms that precondition the ice shelves to collapse. On marine terminating outlet glaciers the mechanisms to trigger thinning is surface ablation causing thinning, and potentially basal melting, though not yet observed (though see this recent paper by Holland et al, 2008). Once the process begins thinner less buttressed ice enables acceleration and more calving and more retreat. There is a potential difference between the two, in glacier such as most marine terminating outlet glaciers, where the glacier flow is rapid, acceleration results from retreat and thinning. In the case of ice shelves a glacier buttressed by them will accelerate after the loss, but the slow moving ice shelf may suffer from reduced inflow. Attention will continue to be focused on these rapid responders to climate change;marine terminating glaciers in Greenland and ice shelves in Antarctica. We can look forward to more details from the extensive 2008 summer field season in Greenland and the upcoming view of the Wilkins this fall.

Unlinked References:

Higgins, A. 1990. Northern Greenland glacier velocities and calf ice production. Polar Forschung, 60, 1-23.
Howat, I., I. Joughin, M. Fahnestock, B. Smith,T. Scambos 2008. Synchronous retreat and acceleration of southeast Greenland outlet glaciers 2000–06: ice dynamics and coupling to climate.Journal of Glaciology, 54(187).
Hughes, T. (1986), The Jakobshavn effect. Geophysical Research Letters, 13, 46-48.
Thomas, R. H. Abdalati W, Frederick E, Krabill WB, Manizade S, Steffen K, (2003) Investigation of surface melting and dynamic thinning on Jakobshavn Isbrae, Greenland. Journal of Glaciology 49, 231-239.
Thomas RH (2004), Force-perturbation analysis of recent thinning and acceleration of Jakobshavn Isbrae, Greenland, Journal of Glaciology 50 (168): 57-66.