Is Pine Island Glacier the Weak Underbelly of the West Antarctic Ice Sheet?

The evidence does indicate that one of the basic underlying principles, proposed by Mercer and Hughes, of what can stabilize or destabilize WAIS was right on the money. The evidence reviewed does not fully confirm the weak underbelly hypothesis, but it provides enough evidence that we had best monitor the situation and expand our attempts to understand it. That is just what the glaciological and scientific community are doing. A number of projects from the British Antarctic Survey, NASA and NSF will continue to expand the research in the area. In January 2008 Robert Bindschadler (NASA) landed on the floating ice shelf of PIG. They found the situation hazardous for plane landing but did leave behind several instruments. NSF has decided to fund establishment of a helicopter camp to safely study the ice-ocean interaction during the 2010-11 summer field season in Antarctica. In 2009 a team of British and American scientists deployed an autonomous robot submarine on six missions beneath the PIG ice shelf using sonar scanners to map the seabed and the ice shelf bottom. This fall NASA’s Operation Ice Bridge has focused much of its energy on the Pine Island Glacier. Seelye Martin of the University of Washington notes that “Pine Island Glacier is a major focus for our mission. We have four flights planned for this glacier. One of our hopes with these flights is to understand the detailed topography under the floating ice tongue. That topography controls the rate of melting there.”


Basal topography of Pine Island Glacier region (from Vaughan et al, 2006).


Bindschadler, R.A., History of lower Pine Island Glacier, West Antarctica, from Landsat imagery, Journal of Glaciology, 48 (163), 536-544, 2002.

Farman, J., B. G. Gardiner and J. D. Shanklin, (1985). Large losses of ozone in Antarctica reveal seasonal ClOx/NOx interaction, Nature, 315, 207-210.

Hughes T. (1981). “The weak underbelly of the West Antarctic Ice Sheet”. Journal of Glaciology 27: 518-525.

Luchitta, B., Rosanova, C., and Mullins, K. (1995). Velocities of Pine Island Glacier, West Antarctica. Annals of Glaciology, 21, 277-283.

Molina, M.J. and F. S. Rowland (1974). Stratospheric sink for chlorofluoromethanes: Chlorine atom-catalysed destruction of ozone, Nature, 249, 810-812.

Rignot E (2008). “Changes in West Antarctic ice stream dynamics observed with ALOS PALSAR data”. Geophys. Res. Lett. 35: L12505. doi:10.1029/2008GL033365.

Rignot, E.J. (1998). Fast recession of a West Antarctic Glacier, Science, 281, 549-551.

Rignot, E.J., D.G. Vaughan, M. Schmeltz, T. Dupont, and D.R. MacAyeal (2002). Acceleration of Pine Island and Thwaites Glacier, West Antarctica, Annals of Glaciology, 34, 189-194.

Scott J.B.T., Gudmundsson G.H., Smith A.M., Bingham R.G., Pritchard H.D., Vaughan D.G. (2009). “Increased rate of acceleration on Pine Island Glacier strongly coupled to changes in gravitational driving stress”. The Cryosphere 3: 125-131.

Shepherd A., Wingham D.J., Mansley J.A.D., Corr H.F.J. (2001). “Inland thinning of Pine Island Glacier, West Antarctica”. Science 291: 862-864. doi:10.1126/science.291.5505.862.

Vaughan D.G., Corr H.F.J., Ferraccioli F., Frearson N., O’Hare A., Mach D., Holt J.W., Blankenship, D.D., Morse, D.L., Young, D.A. (2006). “New boundary conditions for the West Antarctic ice sheet: Subglacial topography beneath Pine Island Glacier”. Geophysical Research Letters 33: L09501. doi:10.1029/2005GL025588.

Wingham D.J., Wallis D.W., Shepherd A. (2009). “The spatial and temporal evolution of Pine Island Glacier thinning, 1995 – 2006”. Geophysical Research Letters 36. doi:10.1029/2009GL039126.

Page 3 of 3 | Previous page

77 comments on this post.
  1. Hank Roberts:


    [Response: Yep, the spectacular scenery does make working in Antarctica something special.–eric]

  2. MapleLeaf:

    Sorry, this is completely off topic, but I think that someone here needs to address PaulHudson’s latest blog at the BBC. See
    Anyhow, people in the know really need to comment on Paul’s blog (which seems to be based on the work by Knight et al. (2009)) and to also present a balanced opinion. Maybe RC could also have a thread on this very topic?
    Sorry for the intrusion.

    [Response: We’ve addressed this silliness before, here, and here.–eric

  3. David B. Benson:

    Mauri Pelto — Most clear, but what are “H” and “PNE” marking the final map, entitled “Basal topography of Pine Island Glacier region”?

  4. MapleLeaf:

    Thanks Eric– still, would not hurt for someone in the know to post some comments on the BBC site.

    [Response: Done. I doubt it will do much good though.-eric]

  5. terran:

    Hi, there has been some mention of volcanic activity in the Antarctic potentially affecting the ice.

    “Ongoing volcanic heat production may have implications for contemporary ice dynamics in this glacial system.” (Corr & Vaughan 2008)

    This article doesn’t mention volcanoes – does this mean that that Antarctic volcanic heat production has been studied and found to be negligible, or is there a lack of data ? I haven’t been able to find much on this on the web, apart from references to Corr’s paper.

    [Response: I think that paper is pretty clear about it, actually. They say that

    “…given such recent eruptive activity, high geothermal heat flux … may be delivering subglacial water to Pine Island Glacier. However, Pine Island Glacier lies in a deep subglacial trough, and this will isolate its subglacial hydrological system from neighbouring glaciers (Thwaites, Smith and Kohler). Therefore, even if continuous or episodic production of melt water from HMSV affects Pine Island Glacier, there is little likelihood that it could affect these neighbouring glaciers. It is thus possible that volcanic activity … contributed to some of the recent changes in velocity of Pine Island Glacier, but it cannot explain the widespread thinning that has been observed across these glacier basins in recent decades. We follow previous authors in favouring an oceanic driver as the likely cause for these changes.”

    In short, yes, this has been considered and found to be unimportant, except perhaps locally.–eric]

  6. mauri pelto:

    H refers to the basal high that is separating the northern and southern portion of the glacier. PNE refers to the base camp used by the British Antarctic Survey and University of Texas.

  7. Kevin McKinney:

    So the jury is still out, but signs at present seem rather ominous?

    [Response: Yup.–eric]

  8. Tegiri Nenashi:

    What are those “acceleration” numbers? Did they report that an individual sensor accelerated with time (and its changing location) — the dynamics completely believable without any “collapsing ice shield” theory. Or, they bury new instrument at the same location and that new instrument is moving faster than the predecessor?

  9. Hank Roberts:

    Wow, video:
    one from a series

  10. terran:

    Thank you, eric.

    Am I correct in thinking that we still don’t know the spatial pattern of volcanic heat production in the Antarctic?

    According to Corr’s paper, HMSV is contributing to melting at Pine Island Glacier – I think this fact should at least be mentioned in the article.

  11. Richard Pauli:

    What a great overview of the ice sheet grounding science, I am not sure I understand some basics, I hope you will permit a few questions from a non-scientist:

    Is a grounding line defined as connected to the ocean? What happens when an under ice lake breaches a connection to the sea – no matter how slightly – does that immediately extend the grounding line as the lake boundary?

    If the ratio of above sea level ice to below sea level floating ice is roughly 9/10 — then with ice resting on ground that is 2000 meters below sea level, isn’t there a set ice altitude above sea level where the ice becomes unstable (200 meters above sea level?) – whether or not it is floating? Are there measured areas where the ice surface altitude is less than this ratio? Are these areas of increased instability?

  12. mgx:

    Such an interesting and complex system to investigate.

    Map of glacier thinning

    Map of lakes beneath the ice sheets

    Map of antartica bedrock

    Antarctic circulation pattern

  13. David Horton:

    #2 I have had a go too – there will be a flood of the “even the BBC thinks global warming is over” posts before you can say “nevertheless, the Arctic melts”.

  14. S. Molnar:

    Could someone define the term “trunk” for those of us who are too lazy to use the Google? I can see from the diagram that it lies between the interior and the shelf, but what distinguishes it from those sections, and what is the point labelled “E” (I can guess that “GL” is the grounding line)?

    [Response: Someone may have a more precise definition but it’s basically the relatively straight part of the glacier (think, ‘river’) below the large, usually more circular catchment area (think ‘watershed’) and the ‘toe’ region (typically an ice shelf for the glaciers in question here). If you look at the the third figure above (this one), the trunk begins right about where it says TG and PIG. Put another way, it’s where all the tributaries come together to form one well-defined channel of flow. “E” in the other figure just refers to the edge of a different figure in the paper cited. Of no importance here.–eric]

  15. James Staples:

    I wonder what effect Vulcanism might have on this Glacier; I think it was one of those noted (1980’s? 1990’s?) in a SciAm Item for the fact that a series of under-the-icesheet Volcanoes were causing the emmisions of large quantities of water from under the edge of the glaciers, where they meet the sea, and the fact that these eruptions were causing rapid advances in the sheets march toward the sea.

  16. Michael Sweet:

    Excellent post.

    Is there any data on the temperature of the water in the embayment where the ice calves? Has it changed temperature in the last 50 years? Is data available on the local currents and their rate of flow and temperature?

    If the ice upstream from the calving front warms from whatever it was does that affect the flow rate of the glacier? Are there any temperatures where the deformation rate of the ice changes rapidly? Has the temperature of the glacier ice changed?

    Why can’t the Ross ice shelf start to melt faster? Is it considered to be so massive that it is resistant to change? After all, it is already afloat, a change in currents to bring warm water to the base would start to melt it right away.

  17. Hank Roberts:

    terran says: 9 November 2009 at 5:29 PM
    > According to Corr’s paper, HMSV is contributing to melting

    That’s not correct; Eric answered you the first time you asked, inline: 9 November 2009, 4:01 PM, quoting the paper.

    Did you read somewhere that it “is contributing”? Useful to know where you got the idea if you remember your source.

  18. Edward Greisch:

    WOW! Not the scenery. I mean what you are saying if I am reading you right. I mean, “Goodby Florida.” I hadn’t thought of glaciers calving almost all the way around Antarctica. Nor had I thought of the rock topography as being that rough. The glaciers haven’t made the bedrock smooth and level. Ice must be more fluid than it seems. I see the distance scale is 100 kilometers, but still… More graphics please. Very impressive. Keep up the good work. Could you give us a virtual tour? I am unclear as to where the line is thru the topography.

  19. mauri pelto:

    #11 The acceleration reported is for specific sensors on the glacier in the case of the GPS surveys, or specific features in an area of the glacier in the case of the radar. The grounding line is where the ice stream become afloat, so if a subglacial lake breaches to the ocean, unless it is deep for the ice stream to become floating, it will not alter the grounding line in and of itself. You will note that the depth in the basin shortly upstream of the grounding line is mostly 1200 meters or more below sea level. This requires over 1300 m of ice thickness to support, with a surface elevation exceeding 120 m, as you suggest. The surface altitude is well mapped using altimetry data.
    #10 The thinning and acceleration of PIG are propagating from the terminus-grounding line area inland. This suggests the origin of the acceleration is in this area. The acceleration and thinning do not match the pattern I would expect from sub volcanic activity. This is why the issue was not discussed. It does not seem to fit the dynamics of what we are observing on all three glaciers in the area exiting into the Amundsen Sea.
    #9 Hank Roberts good links as usual. It was not appropriate in the original post to add too many links to the Youtube’s that highlight Bindshadler’s work last year or ice bridge this year. I knew they would come.
    #15 we do not have long term measurements of either the temperature of the water in Pine Island Bay or of the ice itself, so we cannot address this issues as yet. The Ross Ice Shelf is not resistant to melting from beneath. I did work on bathymetry map for PIG in 1985 and at the time it did not appear there was any shoaling to water restrict ocean water movement under the ice shelf. I am not familiar with the constraints to inflow under the Ross ice Sheet, to date it has not been observed that basal melting has increased significantly.

  20. terran:

    Hank Roberts, my mistake, I should have said “may be contributing”, as per the Corr paper, “high geothermal heat flux … may be delivering subglacial water to Pine Island Glacier”.

    So what is the current heat input from Antarctic volcanoes ? Do we know? My question about it wasn’t answered by Eric, who merely mentioned that the matter has been considered. It’s not clear to me whether he was referring to the topographical explanation given in Corr for the contained localised effect of HMSV, or if he was referring to other research that exists measuring the subglacial temperature anomalies in the Antarctic caused by volcanic activity.

  21. terran:

    #18 thank you for the explanation.
    Is it not possible that sub-glacial melting from volcanically heated water could melt and lubricate the marine portion of the trunk, facilitating the movement from gravity, and also maybe causing the thinning on-land, where friction is greater and movement slower ?

    [Response: This question comes up a lot and the short answer is that there is a lot of evidence that this source of heat is not very important, though in some areas it may not be negligible. See again the reference given above. As they emphasize, this applies only in those locations where there are volcanoes, which is certainly not everywhere where fast glacier flow and thinning is occurring. See the latest paper on this by Pritchard et al. and the figure from that paper, below. (A good summary of this is here.)–eric]

    Ice sheet thinning in Antarctica and Greenland. Red areas indicate regions of strongest thinning.

  22. sidd:

    Thank you Mr. Pelto

    I note that panel A in the second figure is very similar to the instability depicted in Schoof,Journal of Geophysical Research,v112,F03S28,2007

  23. jerryg:

    Have you seen this yet?

  24. Aaron Putnam:

    Dear Mauri and Eric,

    Thank you for this excellent assessment. This is my first experience posting to RealClimate, however I felt compelled to comment. It seems to me that the foundational intuitions and observations made by the late John Mercer and Terry Hughes often go unrecognized. This is particularly evident in the recent literature pertaining to the problems of West Antarctic Ice Sheet instability and even the general concept of mechanical ice-sheet instability. However NSF-funded Antarctic glaciological research continues today largely thanks to the hard work and forward thinking of those two individuals three decades ago. Thank you for recognizing the pioneering work of Mercer and Hughes.

    With best regards,

  25. Thomas:

    As long as people are bringing up the issue of subglacial volcanism there have been some papers purporting that rapid deglaciation can trigger volcanic eruptions. This is of course the inverse of the usual suspects trying to explain away rapid deglaciation by saying “maybe there’s a volcanoe we can’t see”. I don’t have any of the references, but some handwaving theory is convincing enough to motivate further study.

    Essentially, a simplistic model of volcanism would have magma slowly rising within the crust. When it reaches a shallow enough depth that the confining pressure is too little to contain the volatiles, they start to come out of solution. Presumably an eruption follows soon afterwards. Now add rapidly a rapidly thinning ice sheet, and the rate that contours of hydrostatic pressure would descend could easily exceed the rate of magma rise by severalfold. The result (if this simple theory has any explanatory value) would be that during periods of rapid deglaciation local volvanic activity might increase severalfold. Presumably during ice sheet formation local volcanism would be suppressed by the same mechanism. I think this might make an interesting subject for a future post.

    If indeed this is a real effect, how much increased volcanism might we expect? Iceland is an obvious example. But what about the common snow covered volcanic peaks? Is there enough ice, and is the snowline likely to rise fast enough for the rate of iceloss to significantly decrease the pressure on underground magma?

  26. mauri pelto:

    Our best examples of the impact of sub-glacial volcanoes on glacier flow, come from Iceland. A long terms study of a glacier draining the Grimsvotn volcano under the Vatnojokull Ice Cap indicate that there is an initial acceleration due to increased basal melt water pressure, as the flux increasing from the activity. This was followed by a sustained reduction in flow once a conduit system developed to funnel the enhanced meltwater amounts, reducing basal water pressure.
    Volcanic activity would have trouble sustaining an acceleration as has been observed. An upstream acceleration from such an advent would advect ice downstream, tending to thicken the ice downstream at least temporarily. Enhanced longitudinal stretching would thin the ice. Scott et al, 2009 (see post for link) Emphasize that the acceleration and thinning observations are consistent with the hypothesis that changes in PIG result from changes at the downstream end, which encompasses the grounding area and floating ice shelf. At this time the increasing rate of thinning and acceleration suggest a system out of balance and that
    downstream thinning will cause further ungrounding and acceleration. This would suggest to me that the glacier grounding line will retreat into the deeper basin upglacier of the current grounding line. The cause of the thinning, which then leads to the acceleration is the focus of some of the sub ice shelf studies. Bob Bindshadler of NASA suspects that there is likely about fifty meters per year of basal ice being melted at the grounding line.

  27. Kevin McKinney:

    Thomas, #25–this is interesting to me because I’m looking at Ekholm’s 1901 synoptic paper on “carbonic acid” and climate.

    Part of his view of the recurring ice ages involved an oscillating feedback cycle in which silicate weathering and biological uptake/sedimentary deposition draw down CO2, cooling the globe; this cooling would cause the Earth’s crust to contract by a calculated amount of several kilometers (distributed around the globe), inducing intensified vulcanism, which, prolonged over many millenia, would then raise CO2 levels and hence temperature, starting the cycle over again.

    Not so plausible today, but back in 1901 the crust was understood as a unitary layer. Nobody knew from plate tectonics! (Or, for that matter, from radioactive decay helping to heat Earth’s interior, though I don’t think that would have affected his argument.)

    Ekholm also discusses the orbital obliquity cycle and its influence, presenting some interesting calculations a la his buddy Arrhenius, but he sees it as secondary to CO2/temperature feedbacks.

  28. Aaron Lewis:

    The survival of the Antarctic ice shelves including these is dependent on the temperature and salinity of the subsurface ocean water. Warmer or saltier ocean currents could melt WAIS ice from the bottom up regardless of specific local bottom topography. I suggest that this excellent summary should be put in the context of trends in subsurface ocean currents. Has anyone see warmer water in the southern oceans?

  29. Steve Wickham:

    Re #16: I was in Pine Island Bay in March 2006 on a Polar Stern cruise, and maximum water temperatures in the bottom layer were around 1.1 °C. I was doing biology, but colleagues were deploying moorings to get a better picture of the annual temperature/salinity cycle. Don’t know whether they’ve published it yet, though.

  30. Chris G:

    Novice question here, but I’m running the plasticity of ice, the higher density of sea water, and the lower melting point of salt water through my head and I’m wondering what keeps sea water from knifing under the ice and rushing downhill into the basin. I imagine the mechanics would be similar to those that allow meltwater above a certain depth to cut a channel clear to the bottom of a glacier.

    Or rather, presuming the weight of the ice on top and the limited deformation of the ice itself where it meets the terrain forms a seal at the grounding line, is there a simple ratio, something like / that would serve as a rule-of-thumb as to where the ground line is stable?

    Actually, I guess you could look and where ice shelves meet ice sheets and if there is some, relatively consistent, ratio like above for where these lines occur. If the slope inland is positive, then there is no increase-in-pressure-with-depth mechanism and the line is stable where the ratio is 10 to 9. However, if the slope inland is negative, and if the upslope of ice sheet is less than the downslope of the bottom terrain, and the difference causes the ratio line to be crossed, doesn’t the sea water rush in?

    This could happen as the grounding line moves into a steeper downslope area or the outflow of the glacier lowers the height above, or both.

    I think I’m asking the same question as #11. Kind of wondering if we might be in for some major geologic event, like the filling of the Mediterranean, or the North American glacial damn burst events.

    Hoping someone will respond if I’ve managed to produce an interesting question.

  31. Terence J. Hughes:

    Readers may be interested in the genesis of the phrase, “the weak underbelly of the West Antarctic Ice Sheet.” Here it is.

    In the 1980s George Denton and I were charged with reconstructing past ice sheets for CLIMAP (Climate: Long-range Investigation, Mapping, and Prediction, a project of the 1970-1980 International Decade of Ocean Exploration). As a glacial geologist, George was responsible for the areal extent of these ice sheets and, as a glaciologist, I was responsible for their vertical extent. There were two CLIMAP experiments to reconstruct past climates, one at the Last Glacial Maximum (LGM) 18,000 years ago when sea level was about 120 meters lower than today, and one at the Last Interglacial Maximum (LIM) 125,000 years ago when sea level was about 6 meters higher than today.

    John Mercer, in 1968 and 1970, called attention to the fact that the West Antarctic Ice Sheet (WAIS) was grounded mostly below sea level, calling it a “marine ice sheet” that was therefore inherently unstable, and proposed that the higher sea level resulted from its LIM collapse 125,000 years ago. George and I determined to model that collapse for the CLIMAP experiment at the LIM. Along with David Schilling, I had developed a model to reconstruct former ice sheets with ice elevations based on the strength of ice-bed coupling determined by glacial geology. Uncoupling was most complete along liner troughs that I concluded had been occupied by ice streams, fast currents of ice that drain most of the interior ice from ice sheets past and present. Present ice streams draining the Greenland and Antarctic ice sheets have characteristic concave surface profiles produced when slow interior sheet flow is downdrawn by fast stream flow near ice-sheet margins. I reduced ice-bed coupling along ice streams to generate these profiles. For the LIM experiment, we enlisted James Fastook to incorporate the marine instability mechanism published by Robert Thomas in 1977 to make grounded ice margins retreat when ice downdrawn by ice streams became thin enough to become afloat along the WAIS marine margins.

    Large floating ice shelves had formed in the Ross Sea and Weddell Sea marine embayments after WAIS had collapsed in those sectors after the LGM, and I had concluded that these ice shelves, being confined and pinned at places to the sea floor, were now buttressing these sectors of WAIS, preventing further collapse. However, upon inspecting the large map of Antarctica published by the American Geographical Society in 1970, George Denton and I noticed that one-third of the remaining WAIS were drained by two large ice streams that entered Pine Island Bay, an ice-free embayment in the Amundsen Sea sector.
    We decided to run our model for the CLIMAP LIM experiment without allowing an ice shelf to form as the grounding line retreated and enlarged Pine Island Bay. Sure enough, the remaining WAIS collapsed. This collapse scenario was published in the Antarctic chapter of The Last Great Ice Sheets (LGIS) published by Wiley-Interscience in 1981, a book we edited that presented our CLIMAP work.

    George and I were interested in “marketing” our result for Pine Island Bay. We needed some catch-phrase. Winston Churchill had called the Mediterranean coast “the soft underbelly” of Nazi Europe and recommended launching an
    invasion there. In World War I, however, that strategy had been tried by the British with disasterous results at Gallipoli (using Aussies and Kiwis as cannon fodder), and the American/British march up the Italian boot was our longest and bloodiest campaign during World War II. The underbelly wasn’t so soft. Likewise, the Amundsen sea sector of WAIS has remained after the Ross and Weddell sectors have collapsed, so maybe that sector isn’t so “soft” either. To avoid ridicule, George and I didn’t use that imagery in LGIS.

    However, the imagery intrigued me so I replaced “soft” with “weak” and used in in a letter to the editor of the Journal of Glaciology in 1981. Right or wrong, the imagery “took” and the rest is history, as recounted above.

    [Response: Terry, thanks for stopping by! –eric]

  32. Eachran:

    Mr Pelto and eric, all very interesting.

    If I have understood your grounding lines correctly : is there any significance in the ridge in the grounding line which looks to me like a knee waiting to break some sticks for the fire?

  33. RichardC:

    “That topography controls the rate of melting there.” (see last inset)

    Yes, ice melting at the grounding line freshens and so lightens the sea-water. The negative slope of the seabed creates a halosiphon (salt-driven) loop, with salty, warmer sea-water replacing cold fresh water from the melt.

    I think melting also controls the topography. Glaciers push rock, which would explain the hump and negative slope. That big hump probably was the terminal moraine of the glacier for substantial periods, and could be a potential end to the tipping point of runaway grounding line retreat.

  34. Andy:

    The University of Texas placed an excellent workshop on line a couple years back “Secrets of the West Antarctic Ice Sheet”. Any plans for another workshop to present these and other findings?

  35. Andy:

    There was a paper in Science several years back whose conclusion was that once glaciers left their terminal moraine grounding lines behind, the embayment left behind would freshen and cool with meltwater and provide a negative feedback to more melting. Essentially stopping the runaway process. I’ve looked for it, but no luck.

    Has this negative feedback hypothesis since been discarded?

  36. Chas Webster:

    I Hope somebdy will also post a reasoned comment on Will’s column in Newsweek
    As long as the media gives voice to these views, I don’t think they should go unanswered–as we see they are not giving up.

    It’s understandable some are tired of doing so. That’s what graduate students are for, eh?

  37. David B. Benson:

    Andy (35) — That does not appear to have happened or be happening to tidewater galciers in southeastern Alaska; those retreat back to the grounding line provided by the upward-inland slopes.

  38. David B. Benson:

    Somewhat off-topic, but
    Antarctica Glacier Retreat [should read: Antarctic Penninsula ice shelf collapse] Creates New Carbon Dioxide Store; Has Beneficial Impact On Climate Change

  39. mauri pelto:

    #32 and #33 The ridge that the grounding line rests upon is not a moraine. It is several hundred meters high and as can be seen even better in the basal topography map it is not really a ridge but a very broad area of green colored higher terrain that extends beyond the current terminus. There is a trough of blue terrain upstream of the grounding line and another trough of blue beyond the current terminus. This higher subglacial bench is not a weakness, like a knee for a stick, applying torque it is a pinning point buttressing the ice stream. If the glacier retreats off of it, than imagine what may happen.

    #35 David Benson is correct in referring this idea cited to Alaska. This cannot be plausible in a situation where the basal elevation never reaches sea level.

    To get a better feel for the setting take a look at the Youtube from NASA, narrated by Bindshadler

  40. David B. Benson:

    mauri pelto (39) — As a most amateur student of glaciology, I pleased to be found correct by an expert!

    You wrote “This cannot be plausible in a situation where the basal elevation never reaches sea level.” To clarify, you are referring to PIG where the basal elevatiion is well below sea level?

  41. terran:

    Eric and Mauri, thank you very much for your informative replies.

    The melting in Greenland is fairly evenly distributed about the perimiter (apart from a couple of stretches), but the most significant thinning zones in Atarctica do appear to be in proximity to the areas where there are known volcanoes.

    A set of Anarctic volcanoes is viewable on the University of Texas’s GoogleIce.
    ( , requires GoogleEarth). From a prima facie inspection of the area, the presence of numerous formations in the ice similar to those made by the identified volcanoes suggests that more volcanoes may exist than have been identified in the region?


    May I ask whether the type of drainage system known for the Pine Island Glacier?

    The Magnusson et al 2009 paper indicates that geothermal meltwater will contribute to basal pressure (and so to basal sliding) as long as the drainage system remains a distributed one capable of sustaining the pressure. Once a continuous tunnel flow drainage system forms, basal pressure will drop and glacial motion will decrease.

    Geothermal heat production may contribute most to glacial movement when it is low enough to keep the pressure within capacity of a distributed drainage system.

    In the case of the Grímsvötn lake, it can be speculated that had the 1996 eruption not occured, the ice-dam could have held for an indefinite number of years, continuing its contribution to the basal sliding. We also don’t know for how long the dam had been in place for.

    If I may ask a final question – I remember reading that the increased precipitation in Greenland has caused a downflow of ice from the central highlands, which has increased glacial flow and ice loss through calving. (I can’t remember where I read this, and my immediate search has not found it – I’ll keep looking ).

    Western Antarctica has mountainous topography, so if the above effect is true, might something like this be contributing to the increased glacier movement?

  42. terran:

    Apologies, in the first sentence to Eric in my previous comment, the phrase “(apart from a couple of stretches)” should be “(apart from a couple of stretches with no thinning)”.

  43. Jeffrey Davis:

    re: the video referenced in #39

    from McMurdo to the ice shelf:

    “like flying from DC to Kansas City and then riding a lawn mower to Denver.”

    That’s staggering. I’ve got nothing to add except my admiration for the people who go out there to do science.

    “If the glacier retreats off of it, than imagine what may happen.”

    We can all imagine what might happen, but we think faster than the processes transpire. Do the changes take decades? 100s of years? 1000s of years?

  44. terran:

    And the first “melting” should of course be “thinning”. I apologise for the poor proof-reading.

  45. Thomas:

    26) If it retreats off the grounding line, I suspect it would behave similarly to Columbia glacier in Alaska. That moved of its grounding line in the late seventies, and has been in catastrophic retreat phase ever since.

    27) That is quite a different mechanism. Even if this triggering via ice loss mechanism proves to be valid, I doubt there are enough volcanic areas being rapidly deglaciated to cause a global feedback (climate=>volcanics=>climate). The real concern is that the mechanism could create further climate change hazards that hadn’t been appreciated.

  46. Jim Cassa:

    Hello everyone,
    I know this is a blog for scientist, but does anyone know of any good articles for the layman on climate change?

    The reason I am asking you guys is because you know the story from the top down, where others do not.

    I am very interested in posting a layman article on my blog here,

    If you can help, please get in touch,
    thank you,
    Jim Cassa.

  47. Eachran:

    I have been thinking overnight : it often happens for me on the subject of global heating.

    The question relates to increased glacier flow and major ice sheet collapse.

    I assume that increased glacier flow is insufficient to cause rises of metres in the next 200 years (am I correct, please?)

    If one takes a vertical slice through the horizontal plane of the WAIS then the vertical axis shows depths, from top to bottom, air, ice, water (maybe?) and ground, whilst the horizontal axis shows distance, from ice covered ground to open sea.

    Through this vertical slice the ground has a profile of peaks and troughs which, translated to the relief for the ice covered terrain and sea, give ridges and valleys where the ice sheet is less or more thick.

    Vertical and horizontal forces operate on the below sea level grounded ice sheet from tides, ocean currents, thermal expansion and volcanic and other geological (?) activity.

    The ice sheet moves vertically but not uniformly, even if only slightly, creating stresses somewhere in its mass.

    The question relates to a major collapse of the sheet sufficient to free many cubic kilometers of ice into the sea where it would melt more rapidly.

    Where would the fracture most likely take place? On an underwater ridge?

    Is sufficient known about the relief of the ocean bed to model likely locations for fracturing. I appreciate that one fracture effects the likely appearance of other fractures but I have no feel for this.

    What would be the likely total of cubic kilometers freed from the sheet itself and how does this translate to SLR and when?

    Or is fracturing not considered likely in the next 2 hundred years, say, because of the sheer thickness of the sheet at very low temperatures?

  48. Barton Paul Levenson:

    OT but vaguely relevant to climate studies–I have corrected two erroneous figures (for He and Kr) on my gas database web page, with an explanation as to why:

  49. terran:


    Some relevant references, following up my uncited mention of increased highland precipitation in Greenland in post #41 :

    “The Greenland Ice Sheet and Global Sea-Level Rise” (Julian Dowdeswell 2006)

    “Recent Ice-Sheet Growth in the Interior of Greenland” (Johannessen et al. 2005)

    The two papers don’t mention any velocity changes due to increased slope, but they do record that at elevations above 1500-2000 m, the ice has been growing in thickness by about 5-6cm/yr, whereas below the 1500m level there’s been a thinning of about -2 cm/yr.

    This would indicate that the slope has increased, and with it the gravitational stress, leading to an enhanced downhill flow of the ice.

  50. Kevin McKinney:

    Jeffrey Davis, I think the point here is that we don’t know the answer to your question yet; hence the intense research focus. However, the observed changes seem to suggest that the answer may turn out to be biased more toward the low end of the range than hitherto thought.

    This PDF “power-point” by DeConto et al comparing models with the ANDRILL data might be of interest.

  51. Kevin McKinney:

    #46, Jim Fassa–

    Start with the “Science Links” on the RC sidebar, to the right of your page.

    The very first one, “AIP: The Discovery of Global Warming,” is universally recommended. It’s got a ton of information, organized historically, but with hyperlinking that lets you go deeper, or pursue an interesting point, and is well-written and well-researched.

    There are also links to the IPCC reports, and the Summary for Policy Makers is succinct and clear. That might be another one to try early on.

  52. mauri pelto:

    #41- The bed beneath the main trunk of PIG in the basin upstream of the grounding line is thought to be water saturated till. This provides a different hydrology than a conduit system. It is more distributed. In terms of greater accumulation in the upper basin driving greater movement. This could happen on GIS but has not, the changes in thickness at the summit have been minor compacted to ice sheet thickness and such a change would take time. Why is this not the scenario here? Again look where the thinning and acceleration are most pronounced, at the terminus. Where did the process first become apparent, near the terminus. This suggests a downstream not an upstream mechanism driving this change. In addition the upglacier area would have to thicken appreciably before any acceleration occurs. This would be visible in the altimetry record.

  53. Chris G:

    Followed up on the Columbia Glacier references. Posting the links I found in case other amateurs are not already familiar with that situation.

    in particular

    Figure 9. Changes in the surface elevation and bed profile of Columbia Glacier.

    Looks like a really good example, albeit on a smaller scale than the PIG.

  54. Chris G:

    Hit the Submit button too fast, the relevant quotation is,

    “Research at Columbia and other
    Alaska glaciers can be applied to the Antarctic.
    where most of the world’s ice exists.
    The West Antarctic Ice Sheet is also a tidewater
    glacier, but more than a thousand
    times the size of Columbia Glacier. A similar
    instability may exist there, and scientists
    say a rapid retreat could cause a rise of
    sea level of several meters in a few centuries.”

  55. Jesse:

    Question: while the heat input from local volcanism is currently negligible (am I understanding this right?) would that change if there was some major activity?

    Also: would that major increase in volcanic activity (if there were one) be something that, under ordinary circumstances would be negligible, but because of the human-induced warming have a greater effect?

    Another way to put it: minus AGW, the sub-ice-sheet volcanoes would not matter. With it, do they matter now?

    My impression, given what has been posted here, is possibly no. But intuitively I was thinking that if you had a set of eruptions — basically if we were all really, really unlucky — then you could have a more serious problem.

    And how much of the ice sheet has to go before Florida is underwater? At what point do we get the “tipping” where it becomes catastrophic? That is, should we be building very big dikes around New York right now?

    Sorry if this seems obvious, I just want t make sure I am understanding what I am reading.

    [Response: There’s no question that volcanic eruptions under the ice could make a difference, but they’d have to be in the right place, and they aren’t. And keep in mind that volcanoes erupt for a while and then stop. In contrast, the warm water that is under the floating ice (ice shelf) is a persistent source of heat under just the right place.
    As for tipping points, I personally think this term is greatly overused, and is inappropriate in most place (Arctic sea ice for example). But the tidewater glacier situation is one place where it is entirely appropriate, and where a tipping point can be well defined, and in two different ways. On the one hand, continued melting from below — or increased melting from above (this isn’t imminent, but will happen at some point in the next century or two) — will eventually lead to a tipping point in time where rapid retreat begins, regardless of what ocean or air temperatures do from that point on. On the other hand, the point at which the grounding line retreats from its current high ground is a clearly-defined physical location — hence another tipping point.–eric]

  56. Hank Roberts:

    > At what point do we get the “tipping”
    There’s no single point. Think “slippery slope.”
    > should we be building very big dikes around New York right now?

    [Response: I’m not sure that we should be building dikes around New York, but in point of NYC has plan to address sustainability, and the threat from sea level rise is part of the things being considered seriously. Some might call this a response to lefty brainwashing, but I think it remarkably forward thinking.–eric]

  57. mauri pelto:

    Chris G good point with respect to the Columbia Glacier. This is a much different type of glacier. However, the thinking in terms of why it was going to retreat significantly parallels that of PIG. It terminated on a shallow shoal, really mostly an island. Behind this was a deep water basin. If it retreated into this basin the glacier would begin to calve fueling a calving retreat. The shallow point at the terminus was a pinning point of stability. Similarly if PIG grounding line retreats into the deeper basin, it is hard to imagine a stable position during a resultant calving retreat, in that deep basin stretch. The terminus of PIG is not the key it is the grounding line location.

  58. Ric Merritt:

    Chas Webster #36: Thank you so much for keeping us up to date on George Will. We now know that, given the evidence that each decade is warmer than the last, the obvious conclusion is that we’re cooling. This may well continue for decades. And (about equally important, measured by the space devoted to it) no worries about climate, ‘cuz sometimes people like to get too worked up about sharks.

    I’m feeling much better now.

  59. Hank Roberts:

    Thank you Eric for the pointer to New York’s plan — much smarter than “building very big dikes … right now” — reducing fossil fuel use; increasing efficiency; revising flood plain maps, insurance and building codes. Smart.

  60. terran:

    #52 Mauri, thank you for your patient explanations.

    From Shepherd et al., 2001
    “The glacier develops a driving stress in excess of 100 kPa in surmounting the bedrock trough, and the associated upstream thickening appears to determine the location of the present grounding line.”

    Since the trough is impeding the outward flow of the glacier, doesn’t this suggest that if the grounding line retreats into the trough, the glacier mass will no longer be impeded, and the driving stress which would previously build up would now be able to move the ice forward, and so tend to offset the retreat of the grounding line?

  61. mauri pelto:

    Terran: You are correct a reduction in driving stress would occur, if the grounding line retreats off of the current higher terrain. That is what causes the acceleration, less driving stress holding the ice back. This acceleration is what we are talking about, less friction, less buttressing-less braking. This acceleration in our experience leads to more calving, more retreat and more acceleration. It does not in general lead to a readvance to a point of stability.
    Take a look at the image at
    Notice how the surface smooths out beyond the green line-grounding line-on the floating section, where there is no be interaction to generate surface roughness. Also notice at the top of this image the grounding line swings closer to the terminus. Note the substantial rifting that is apparent at the grounding lines closest point to the ice front. These rifts are zones of weakness as has been noted on Wilkins and Larsen Ice Shelf. If these rifts are enhanced that is a warning sign. Rignot (2002) indicated that they were increasing in this area. This is a 2001 image and the large rift beyond this that spread across the glacier in 35 days or so, led to an iceberg calving event.

  62. John (Burgy) Burgeson:

    To Jim Fassa — take a look at my article, published last summer. A copy of it is on my web site at

  63. Jeffrey Davis:

    re: 50 of Kevin McKinney

    The graphs in the pdf you referenced show many almost “instant” collapses of the WAIS across 5 million years, but the graphs aren’t fine grained enough to say if “instant” is ~10,000 years or less. Still, if “instant” is even 1000 years, the year to year increase in melting for low-lying coastal regions will be dramatic.

  64. SCM:

    Re #46
    Jim you should look at the links in the ‘start here’ section of the website (see menu at the top).

  65. Hank Roberts:

    Another image, after the event Mauri points to above:

  66. Chris G:

    With respect to Mauri #57, Terran #60, and Mauri #61:

    First, thanks!

    On the grounding line, I would imagine it is an indicator of where the balance of forces lies (with some offset for the viscosity of ice) rather than being a key or event driver, if that is what Mauri meant.

    Terran, as I imagine it, there are huge driving forces on a column of ice a thousand meters high. These forces would be high enough to deform the lower ice faster than it could melt wherever the forces were not in equilibrium. So, as the leading edge melts, this deformation and filling would be first observed as a thinning of the sheet and, in particular, the trunk or main ice stream channel, as well as an acceleration of the stream. (Which is what we are seeing now.) I don’t see the GL moving until after any previously built up driving force (height of unsupported ice column) has already been released (deformed to a lower height). I imagine calving as occurring when the loss (ie. from melting) from underneath is faster than the ice above, under less pressure, is willing to deform. Maybe that isn’t how an expert would put it, but it seems to be a working model for this layman.

    Thinking on the rate of collapse, and going back to the Columbia Glacier as an example. That glacier has been retreating about 0.8 km per year, under similar topography conditions as the PIG. The PIG is a little higher and a little deeper; so, larger forces are involved and maybe the ice flows/deforms faster. Let’s just give it a round number of 1 km per year of retreat. It’s about 200 km before the topography starts rising again. So, if you want an intuitive guess as to how long it will take, let’s say about 200 years to clear the main channel and a bit longer for the rest. Kind of mundane for a Hollywood movie. Of course, the calving front will be a spectacular 30 km across the main trunk and, where Columbia has earth on each side, the PIG has ice on both sides of the channel; so, it will be more complicated than the Columbia collapse.

  67. Sam Vilain:

    I’m not sure if they were the same model as DeConto’s, but at Professor Tim Naish’s inaugural lecture at Victoria University of Wellington, New Zealand, he showed the results of a detailed melt model of the WAIS, and if I understood correctly it took something like 200-300 years for most of the sheet to melt. This timescale of melt was also confirmed by the ANDRILL record (Naish headed at least one of these expeditions) and another paleo record showing that the sea level rose rapidly around a corresponding time. There’s no question that the sheet could melt in under 500 years – it’s done it before, and that was without massive CO₂ forcing…

    [Response: Careful here. It isn’t directly about ‘melt’. This part of Antarctica will remain well below freezing for some time to come. It’s about increased flow of the ice into the ocean, and from *there* it melts (and of the icebergs raise sea level whether they melt or not). Note that 300 years is a very firm minimum time frame. Deconto and Pollard’s results actually suggest anywhere from 300 to 6000 years. That’s a sea level rise of about 1 m/century to as little as 5 cm/century. Either is ‘serious’ I suppose but the range is pretty big. We simply don’t know yet which end member is the best estimate of reality.–eric]

  68. Eachran:

    Still all very interesting and I think that I understand the basics but the main point with me is : what is the number?

    We have the IPCC report on SLR which no scientist believed (me neither and I am not a scientist) when it was published and has now unofficially been updated by the experts to 1m by the end of the 21C.

    We now need to fill in the missing bits like can the increased speed of galciers themselves lead to a number higher than 1m by the end of the century or do we need a big chunk of the sheet to fall off to achieve anything like 3m or maybe much more by the end of the following century.

    As I understand it GIS is stuck by virtue of its topography (am I correct?) and limited to 2m max.

    So where does this lead with the WAIS?

    Does the increasing speed of glaciers and subsequent calving (which I understand can be significant when the grounding line (?) is so deep) provide sufficient floating ice to affect SLR significantly?

    Or do we need a fracture somewhere?

    The paleological record shows maybe big changes. They surely cant come just from the speed of glaciers.

    If I were a practising scientist I would lend a hand but I am not.

    [Response: That 2 m max number is actually Tad Pfeffer’s estimate for *EVERYTHING* including Antarctica and Greenland in the next 100 years. So much less than 2 m from Greenland alone. However, 2 m is a lot! And Pfeffer’s calculations don’t go beyond 100 years. In 200 years, it may be a very different story (i.e. upwards of 3 m or more, easily).–eric]

  69. Eachran:

    that should read paleontological (I think).

  70. mauri pelto:

    Examining the specific area around the grounding line in more detail shows how close to the basin the Grounding line is. This article is an examination of observation of rapid changes of a significant WAIS glacier, and the implications for this glacier. Further to note that proposed mechanics were well assessed 30 years ago, and the scientific community is launching a significant effort to monitor and quantify this glacier. It is the results of this that could lead to the big picture answers many seek.

  71. Lynn Vincentnathan:

    Thanks for the understandable-to-a-layperson post.

    I have a question. Would the rising sea level have any impact on this glacier & ice shelf? I’m thinking that ice floats (esp in salt water, I suppose), and since this glacier bed is below sea level, and if sea water were to get into it (or even at front edge points where it meets the sea), a rising sea level might put even more upward pressure on the glacier.

    I know glaciers are heavy, so that might not be the case, but the sky is big and CO2 molecules small, and yet we have the natural GH effect and AGW….

    [Response: Perfectly reasonable layperson question. Yes, of course ice floats! It doesn’t matter that they are “heavy”. What matters is that they are less dense than liquid water. Indeed, one can measure the up and down of the ice shelves and even the grounded ice upstream with the rising and falling tides on a daily basis. So there is no question that rising sea level can ‘unground’ these glaciers. The question though, is how fast. That’s a much harder problem to answer — hence all the research going into this.-eric]

  72. mauri pelto:

    Sea level rise of the rates we are seeing now, have a greater influence on larger, thinner ice shelf systems. The PIG is not dominated by its ice shelf, hence the potential for rapid ice stream response. However, the grounding line of ice of the thickness of PIG is not as responsive as say the thin ice at the margin of Petermann Glacier in Greenland or Wilkins Ice Shelf.

  73. chris:

    re #49 terran

    Today’s Science has a paper that is relevant to your discussion of accelerating Greenland ice mass loss in relation to its components:

    M. van den Broeke et al. (2009) Partitioning Recent Greenland Mass Loss Science 326, 984-986 (November 13, 2009)

    abstract: Mass budget calculations, validated with satellite gravity observations [from the Gravity Recovery and Climate Experiment (GRACE) satellites], enable us to quantify the individual components of recent Greenland mass loss. The total 2000–2008 mass loss of ~1500 gigatons, equivalent to 0.46 millimeters per year of global sea level rise, is equally split between surface processes (runoff and precipitation) and ice dynamics. Without the moderating effects of increased snowfall and refreezing, post-1996 Greenland ice sheet mass losses would have been 100% higher. Since 2006, high summer melt rates have increased Greenland ice sheet mass loss to 273 gigatons per year (0.75 millimeters per year of equivalent sea level rise). The seasonal cycle in surface mass balance fully accounts for detrended GRACE mass variations, confirming insignificant subannual variation in ice sheet discharge.

  74. Hank Roberts:

    Not only tides, but also waves from distant storms, may move the floating ice enough to crack pieces loose; this research used summertime solar-powered instruments so reports the correlation with distant N. Hemisphere winter storms; they comment that S.H. winter storms would too but were not recorded (no solar power for their instruments).
    GEOPHYSICAL RESEARCH LETTERS, VOL. 33, L17502, doi:10.1029/2006GL027235, 2006

  75. Kevin McKinney:

    Re #74, Hank–

    Solar power in high latitudes. . . it’s a sometime thing. (H/T to the Gershwin brothers.)

  76. Terran:

    Mauri, Eric, Chris G, and all, thank you for an enlightening conversation (from my perspective, at any rate :).

    If I may add one more speculative question : are the portions of glacial sheets formed during periods of high ice flux less stable, and more prone to calving, than those formed during slow flux?

  77. mauri pelto:

    Terran: An examination of a map of glacier velocity for either the Pine Island presented in this post or of the ice streams feeding the Ross Ice Shelf indicate that most of the ice sheet region is not a fast flow region.
    Compare the entire basin area to the magenta fast flow regions in either. Most of the deformation in an ice stream is in the bottom portion and thus most of the thickness of the ice stream is just carried along. Take a look at the internal layering of PIG in Figure 3 of the following paper. Notice the continuity of most of the layering, some of which is quite old. This illustrates two things, that PIG flow regime has not changed much for quite awhile and two that all the action from topography and changing basal conditions is happening lower in the ice stream.