Arguments
Global warming and the vulnerability of Greenland's ice sheet
Posted on 30 May 2014 by John Abraham
This work is important because in the past decade or so, it has been found that Greenland is losing ice, lots of ice, to the world’s oceans. As a consequence, Greenland is one of the significant contributors to sea level rise. The level our oceans will rise to in the next decades and centuries depends strongly on how fast the Greenland ice sheet will melt.
This study is novel because of the way it combined measurements with mass-conservation calculations. The method allowed far better resolution of the thickness of Greenland glaciers and the shape of the valleys beneath the glaciers. The shape and depth of these valleys is important for moderating the speed of ice flow to the oceans. In fact, in the paper, authors state, “the overall state of mass balance of the ice sheet is affected by considerable uncertainties in bed topography and ice thickness.”
While measurements of ice sheet thickness have been made for decades using airborne radar systems, by NASA’s Operation IceBridge in particular, these measurements are not sufficient in number and quality in important ice sheet regions – near the edges and in the South. There, the use of radar is more challenging because of the presence of water in the ice. It is the edges of the ice sheets, particularly the portions that have glaciers that terminate in oceans, which dominate the flow of Greenland’s frozen water into the ocean.
Sukkertoppen glacier. Photograph: Michael Studinger
In order to overcome the paucity of high-quality radar measurements, the authors included ice motion information from satellite imagery. This motion, when input into a mass-conservation model, allowed the researchers to extrapolate “sparse ice thickness data to larger areas with few or no data.”
When they compared their results with other studies, the authors found, “widespread presence of well-eroded, deep-bed troughs along the ice-sheet periphery, generally grounded below sea level, coincident in location and spatial extent with fast-flow features and extending over considerable distance inland.” These features were not previously known.
These findings allow a few conclusions. Aside from the importance of deep troughs to ice motion, the extension inland means that glaciers will have to retreat further than anticipated inland in order to reach a position above sea level. “Some of them will stay in contact with the ocean for centuries, when we thought that in a couple of decades they would stabilize.” said Mathieu Morlighem.
The ice sheet is therefore more vulnerable than predicted, and existing projections of sea level rise contribution from Greenland are too conservative and need to be revised. The research also shows that also means that these troughs are old – it takes 10,000 to 100,000 years for these troughs to be created through erosive action. Also startling is that while only 8% of these regions correspond to ice-grounding below sea level, they are responsible for 88% of the total ice discharge. These are the parts of Greenland that really matter.
As the authors state in the paper, “Our findings imply that the outlet glaciers of Greenland, and the ice sheet as a whole, are probably more vulnerable to ocean thermal forcing and peripheral thinning than inferred previously from existing numerical ice-sheet models.”
Dr. Rignot gave a great summary when I contacted him, "Observations published in a number of recent, independent studies in the scientific literature provide unquestionable and timely evidence that climate change has already significantly shaken up the giant ice sheets in Greenland and Antarctica."
Thanks lots for posting this. Do we have any idea about how much this particular dynamic might increase the rate of sea level rise in the next few decades? A little? A lot? Not at all?
If even a little, how does this, plus what we have learned about WAIS change our understanding of likely sea level rise in the next few decades and by the end of the century? This is a questions of vital concern to the hundreds of millions living in directly vulnerable areas, and really, to everyone one on earth, since nearly all will likely be affected one way or the other by the wrenching changes needed to deal with the evacuations of these areas.
So any light anyone can throw on these questions would be most welcome.
I think this shows a few problems with the IPCC reports and using them as THE reference for assessments. First, new research is coming out fast, changing previous conclusions (as above) and 7 years is a long time. Then (and related) the ESLD effect might have affected estimates of ice sheets and sea level rise. IPCC AR4 was quite far off at the low side but also IPCC AR5 gave a lower value than the average expert: www.realclimate.org/index.php/archives/2013/11/sea-level-rise-what-the-experts-expect/ and www.glaciology.net/Home/Miscellaneous-Debris/comparisonofsealevelprojections
OK, that estimates are changing with time is in principle a normal part of the scientific progress and not in itself a proof of any bias. But the tendence has been that new (dynamical) effects are added with time, increasing estimates, which is the problem with bottom-up approaches. And IPCC AR5 dismissed the semi-empirical approaches in favour of the process-based, but they now seems to be the more realistic.
wili@1,
The study in question does not try to quantify the GIS contribution to SLR nor the prediction of its contribution (in the abstract and press releases I only have access to).
The Antarctic study you're most likely referring to (McMillan 2014), quantifies the current total Atntarctic contribution from recent satelite altimetry as 160Gt/y, equiv. to 0.45mm (central value). Recalling the AR5 number for Antarctic ice sheet contribution: 0.27mm/y, we can clearly see that AR5 has been substantially outdated as (McMillan 2014) increases it by some 70%. We know that IPCC findings about SLR outdate quickly, so no surprise here.
I think the increase of AIS loss rate comparing to AR5 is nothing new, considering e.g. (Hansen 2012) who predicted sustained doubling of icesheet melt rate every 7-10 years with SLR up to 5m by 2010. I think Hansen based his prediction on GIS data but I don't know the recent GIS melt data, apart its contribution is just slightly higher (0.33mm/y by AR5). AIS melt rate in (McMillan 2014) seems to be somewhat slower than (Hansen 2012) prediction, if it can be a bit of consolation. That is just my very rough interpolation: in order to verify if Hansen 2012 was correct, we need to wait a couple decades (if we are young enough).
To put the current IS melt contribution to SLR in perspective, it's worth remembering they are still quite behind the main contributors:
- thermal ocean expansion due to warming: 1.1 (0.8 to 1.4) mm/y
- glaciers: 0.76 (0.39 to 1.13) mm/y
and based on that numbers, IS will not become dominant contributor for another 20y or so, even according to somewhat overestimated prediction by Hansen.
The problem with the IPCC reports is that they are biased to be too conservative. If they were literally, mathematically unbiased, half the time they would hit high, half the time they would hit low. We don't see that. Three obvious reasons are that humans tend to be biased in the conservative direction, the scientific process is biased in the conservative direction (default assumption is that nothing new is happening), and that for reasons of credibility in the face of "skeptics", I think the IPCC attempts to never overpredict. This sort of cascaded filtering is nothing new; there are examples of companies that failed because a CEO inclined to "shoot the messenger" successfully created his own little bubble of misinformation, till reality intruded.
But if the IPCC was literally and accurately unbiased, about half their predictions would fall short as new data arrived. There's things where we can look and say "that won't happen" — we know the ice caps won't melt quickly in place, because physics tells us so pretty directly. But otherwise, it would be nice to see predictions that were based on sound science, yet not artificially muted by pervasive conservatism. (I assume this is what we're getting from Hansen, which is why he sounds so much more alarmed than the IPCC.)
This is like the spectator sport of global warming climate destabilization. A very big deal, everybody should be watching closely.
Thanks
chriskoz @#3: Thanks for your insights. But the last bit leaves me confused: "IS will not become dominant contributor for another 20y or so, even according to somewhat overestimated prediction by Hansen"
Your numbers for GIS and total Antarctic contributions (if I understand them correctly) add up to (.45mm/yr + .33mm/yr =) .78mm/yr, about the same as that from all glaciers, by your numbers, and about 80% of the contributions from thermal expansion.
If we take Hansens fastest doubling time (which, I agree, seems...ambitious) of 7 years, we would be getting about 6mm/year from ice sheets alone in about 20 years. Are you expecting the other factors to grow even faster during this period? If not, how is it that you say IS contributions would not be 'dominant'?
wili@6,
I checked AR5 Table 4.6 and found out that my previous IPCC numbers were the last 18y average IS loss trends. Because IS melt is accelerating, the more recent trends - last 6y - are higher:
Period Ice sheet loss (mm yr–1 SLE)
Greenland
2005–2010 (6-year) (0.63 ±0.17)
1993–2010 (18-year) (0.33 ±0.08)
Antarctica
2005–2010 (6-year) (0.41 ±0.20)
1993–2010 (18-year) (0.27 ±0.11)
Combined
2005–2010 (6-year) (1.04 ±0.37)
1993–2010 (18-year) (0.60 ±0.18)
So, AR5 latest 6-year trend is consistent with (McMillan 2014), no surprise here. The AR5 combined 1.04mm/y value constitutes about 1/3 of current total SLRR of 3.2mm/y.
But the question when we are going to see IS contribution emerging as dominant factor is still too hard to tell. You can check AR5 chapter 4 summary e.g. in Bamber presentation: Look esp. at Figure 4.25 on page 12 and try to extrapolate... Or, with Hansen's fastest doubling every 7y, you'll arrive at 8mm/y by early 2030s... But with more realistic IMO doubling every 10y+, you'll arrive at 2mm/y by 2020s which still less than other contributions, and 4mm/y (a dominant contribution) by 2030s.