RealClimate logo

The Greenland melt

Filed under: — eric @ 23 January 2013

Eric Steig

Last July (2012), I heard from a colleagues working at the edge of the Greenland ice sheet, and from another colleague working up at the Summit. Both were independently writing to report the exceptional conditions they were witnessing. The first was that the bridge over the Watson river by the town of Kangerlussuaq, on the west coast of Greenland, was being breached by the high volumes of meltwater coming down from the ice sheet. The second was that there was a new melt layer forming at the highest point of the ice sheet, where it very rarely melts.

A front loader being swept off a bridge into the Watson River, Kangerlussuaq, Greenland, in July 2012. Fortunately, nobody was in it at the time. Photo: K. Choquette

I’ve been remiss in not writing about these observations until now. I’m prompted to do so by the publication in Nature today (January 23, 2013) of another new finding about Greenland melt. This paper isn’t about the modern climate, but about the climate of the last interglacial period. It has relevance to the modern situation though, a point to which I’ll return at the end of this post.

The new paper in Nature, Eemian interglacial reconstructed from a Greenland folded ice core (NEEM Community Members, (2013)) is the culmination of many years of work in Greenland led by Dorethe Dahl-Jensen and her team from the Centre for Ice and Climate at the Niels Bohr institute in Copenhagen, with substantial involvement of scientists from around Europe, the U.S., China, Japan, and Canada.*

The big news is that this group has managed to obtain and use the information in ice from the Eemian — the peak of the last interglacial period, about 125,000 years ago — in Greenland. Getting usable Eemian ice from Greenland has been a Holy Grail of ice core research for the better part of two decades. We thought, back in the early 1990s, that we had obtained Eemian ice in the GISP2 and GRIP ice cores drilled near the ice sheet summit. It turned out that the lowermost part — anything older than 100,000 years — was messed up by ice flow, making it impossible to learn anything much about climate from it. The Danish group then led a project further to the north at “North GRIP” that, based on radar-echo-sounding data, should have had an intact Eemian period. But the temperature at the base at NGRIP was higher than expected, and the Eemian ice had melted away.

The latest attempt was the “North Eemian” (NEEM) site in northeast Greenland. Here too, the initial results were disappointing. As at GISP2 and GRIP, there are folds in the ice, and some of the layers containing the ice of Eemian age are repeated several times. However, in this case the folds are very large, and there are continuous sections that are not scrambled; they are just a bit out of order. It took significant work, but the group has unfolded the data from the folded layers and it is now evident that the goal of the NEEM project– having an interpretable section of Eemian ice — has succeeded after all.

The findings are spectacular. In the Eemian ice, there is clear evidence of significant melting of what would then have been snow at the surface. The amount of air trapped in the ice undergoes rapid fluctuations, resulting from the fact that ice that melts and then refreezes generally winds up with fewer air bubbles in it than the original porous snow. There are also strong fluctuations observed in soluble gases such as N2O whereas variations in the oxygen isotope concentration — both in the molecular oxygen (O2) in the air and in the ice (H2O) itself — are small. The isotope concentration of the O2 can be matched to that in undisturbed ice from the same time period in ice cores from Antarctica, providing a way to date the ice, showing unambiguously that non-disturbed layers are preserved from the peak of the Eemian period, about 125,000 years ago.

Qualitatively, the evidence for melt in the NEEM Eemian ice shows that it was warm at the time. Obviously. But more interestingly, the last year of the NEEM project was 2012, and researchers were able to witness first hand what the formation of melt layers mean at NEEM in terms of the ambient conditions. In July 2012, the NEEM saw above-freezing temperatures for six consecutive days (10 to 15 July), with rain events on 11 and 13 July. When the water refroze, it formed several distinct, clear layers of ice (which we call a “melt layers”) between 5 and about 60 cm down in the snow, about 1 cm thick. This is a rare event. It was so warm over Greenland in that week that a significant melt layer also formed up at the Summit; in fact, the entire surface of the ice sheet was melting.

That hasn’t happened — not once — in the entire satellite record (see Jason Box’s excellent blog, for more on this, and Marco Tedesco's paper.). In fact, examination of melt layer records from ice cores at Summit shows that a melt layer like the one that formed in 2012 was the most significant Greenland melt event since at least the late 19th century. If you drill about 100 m down into the ice and recover an ice core, you invariably find that layer, shown in the photo below (the bright line at which the person’s thumb is pointing).
.Greenland ice core from ~80 m depth. E. Steig photo.

According to a recent paper on the 2012 melt by Nghiem et al., in Geophysical Research Letters, the 19th century event dates to 1889. One has to go back about 700 years to find the next such event, and overall, these are about once-in-250 year events over the last 4000 years. Prior to that, they occur more frequently — about once per century during the mid Holocene “climatic optimum”, when it was on average much warmer than present in Greenland in summer, due to the peak in Northern Hemisphere insolation due to changes in the earth’s orbit (Milankovitch forcing). Even during the mid-Holocene, though, there is no evidence from the ice cores that there was sufficient melting to create such strong anomalies in the air content and trace gas concentrations in the ice, as was observed in the Eemian in the NEEM ice. Thus, it was even warmer during Eemian than during the mid Holocene.

How much warmer was it? Jason Box estimates from satellite data that the temperature in July 2012 at high elevations over the Greenland ice sheet was a full 10°C (18°F) warmer than the daily average of the 2000′s decade; 1 standard deviation is about 3°C, so this is about a 3-sigma event. If, as the NEEM researchers estimate, the same sort of temperatures were required to produce the EEM melt layers, it suggests that during the EEM in Greenland it was also about 10°C warmer than present in the summer — but not just once per century, but much more often, perhaps every summer. I’m interpreting a bit here: the NEEM group doesn’t actually use the presence of melt layers per se to estimate the summer temperature; rather, they use the observation that the δ18O values of the ice at this time are >>-33 ‰. δ18O is a proxy for temperature in Greenland ice, and the NEEM paper uses this to estimate that the temperature must have been about 8°C (+/-4°C) warmer than present. Not coincidentally, the δ18O values of the snow and rain that fell in July 2012 was also >-33 ‰.

None of this should be interpreted to suggest that we are in “Eemian-like” conditions just yet. After all, there has only been one Eemian-like melt event observed in modern times, and the extremely warm summer of 2012 clearly involved anomalous weather conditions — a particular pattern of pressure anomalies over the northern high latitudes (Marco Tedesco's paper) that may also partly account for the exceptional low sea ice cover that year. The 2012 event, however, gives us a flavor of what the future is likely to bring. It will be very interesting to watch the satellite imagery over Greenland in the next decade and beyond.

What are the implications for the Greenland ice sheet? Possibly, that it is less sensitive to climate warming than some of the higher-end estimates suggest (e.g. Cuffey and Marshall (2000) suggested Greenland could have contributed > ~4 m to EEM sea level), though very much in line with more recent estimates (e.g. Pfeffer et al. (2008)). The estimated temperature change of ~8°C is quite a bit warmer than most previous estimates which are more in the range of 2-5°C (though the uncertainty estimates clearly overlap). Thus, whatever the contribution of mass loss from the Greenland ice sheet to the huge (4-8 m) rise in sea level of the Eemian, it occurred under very strong temperature forcing.

The presence of Eemian ice at the NEEM site itself places constraints on the ice sheet configuration. It obviously rules out any configuration in which this area of the Greenland ice sheet was gone. That typically occurs in ice-sheet model simulations that involve more than about 2 m of sea-level-equivalent mass loss. Thus, the NEEM ice core record suggests both that temperatures may have been warmer than once thought, and and that the ice sheet mass loss was unlikely to have been >2 m of sea level.

The new data from the NEEM ice core may also point to a lower limit on the magnitude of the Eemian sea level contribution from Greenland. Evidently, it can become very warm indeed over Greenland — much warmer than most previous modeling exercises have considered. Combined climate/ice sheet model estimates in which the Greenland surface temperature was as high during the Eemian as indicated by the NEEM ice core record suggest that loss of less than about 1 m sea level equivalent is very unlikely (e.g. Robinson et al. (2011).

There are caveats of course — the new data is just from one site, and estimates of the total ice loss don’t provide information about the rate at which that loss occurred. Still the new data show that Greenland, while evidently contributing significantly to Eemian sea level, cannot have contributed more than half the total — despite the strong forcing. This once again points to Antarctica as the major source of Eemian sea level rise. There are only about 3 m of sea level rise available from West Antarctica, and it remains unclear whether all of West Antarctica may have collapsed. On that subject, look for some more exciting ice core news in the near future, from a core at Roosevelt Island by a New Zealand led team.

Note: There is a nice summary of the implications of the paper on the Nature web site, though note that I’m pretty sure I didn’t say — or didn’t mean to say! — that “We are in a similar climate regime as the world was in the early Eemian,” as I am quoted. A key difference is that CO2 was not as high as today, but insolation forcing was much higher. So the analogy only goes so far. See the paper by van de Berg et al. Significant contribution of insolation to Eemian melting of the Greenland ice sheet for an in-depth discussion about these differences.

Update: The website looks like a great resource for those interesting in following the modern melt progression in Greenland.

*There was also significant logistical support from the U.S. Air National Guard, who contract with the National Science Foundation to supply C130 Transport plane support for these kinds of projects.


  1. . , "Eemian interglacial reconstructed from a Greenland folded ice core", Nature, vol. 493, pp. 489-494, 2013.
  2. M. Tedesco, X. Fettweis, T. Mote, J. Wahr, P. Alexander, J. Box, and B. Wouters, "Evidence and analysis of 2012 Greenland records from spaceborne observations, a regional climate model and reanalysis data", The Cryosphere Discussions, vol. 6, pp. 4939-4976, 2012.
  3. K.M. Cuffey, and S.J. Marshall, "Substantial contribution to sea-level rise during the last interglacial from the Greenland ice sheet", Nature, vol. 404, pp. 591-594, 2000.
  4. W.T. Pfeffer, J.T. Harper, and S. O'Neel, "Kinematic Constraints on Glacier Contributions to 21st-Century Sea-Level Rise", Science, vol. 321, pp. 1340-1343, 2008.
  5. A. Robinson, R. Calov, and A. Ganopolski, "Greenland ice sheet model parameters constrained using simulations of the Eemian Interglacial", Climate of the Past, vol. 7, pp. 381-396, 2011.
  6. W.J. van de Berg, M. van den Broeke, J. Ettema, E. van Meijgaard, and F. Kaspar, "Significant contribution of insolation to Eemian melting of the Greenland ice sheet", Nature Geoscience, vol. 4, pp. 679-683, 2011.

114 Responses to “The Greenland melt”

  1. 101
    sidd says:

    I keep coming back to Fig 5b) in Tedesco(2012). This shows the cumulative SMB [Surface Mass Balance –ed.] from Jan thru Sep. Cumulative SMB used to reach a min. near the end of August, and remained substantially above zero for the mean for 1958-2011. The last three years 2010,2011,2012 show cumulative SMB going negative earlier in the year, and the minimum dropping very quickly, and occurring later in the year.

    1) I would like the graf extended to cover the whole year.

    2) individual years plotted separately.

    ( 3) And a pony…!)

    I poked around the University of Liege site but do not see the data for the reconstruction.

    I note that the run off anomaly was 350GT for 2011-2012, in comparison to GRACE mass waste estimate of 574GT. So a larger fraction comes from SMB than dynamic discharge. As has been pointed out in Gregoire and others, dynamic discharge thins the ice sheet and lowers the ice surface, plunging it deeper into the ablation zone. In this connection I see that Fig 6a in Tedesco shows ELA far below the saddle at 67N, but according to Prof. Box at meltfactor, ELA exceeded saddle height last year…


  2. 102
    Steve Fish says:

    Re- Comment by wili — 4 Feb 2013 @ 1:35 PM

    Methane is well mixed in the air.

  3. 103
    David B. Benson says:

    Hank Roberts @98 — Check a bedrock elevation map of Greenland; no inland sea but rather a large lake.

  4. 104
    Steve Fish says:

    Re- Comment by sidd — 4 Feb 2013 @ 3:53 PM

    Acronym finder has 59 entries for SMB including Super Mario Brothers, Super Monkey Ball, and Shanghai Meteorological Bureau. Is it so hard to type out the whole thing just once?

    Should this be in the Real Climate Acronym Index?


    [Response: ;) I think it should! I’ll do that. Though it would be better if people would just write things out. Surface Mass Balance. I mean, learn to type, people. Smrtfone era is mking ppl dum–eric]

  5. 105
    Hank Roberts says:

    David, inland lake of freshwater certainly to begin, but can that persist?

    I was expecting (can’t say ‘anticipating’) a saltwater connection through the area now “0 to 100 meters” above sea level (in this map) thinking the last deglaciation raised sea level up to 130 meters (and assuming isostatic rebound wouldn’t keep up with sea level rise).

    Has the deepest area of the basin been drilled down to whatever sediment underlies the icecap, do you recall? It might have an interesting record, if sediment from several cycles hasn’t been squeezed and scraped away by ice, but I haven’t found that.

    Same of course applies to Antarctica, I think we’re so far mostly looking at the edges.

    I’m well into science fiction scenarios here — if I’m going to stock that basin with fish ….

  6. 106
    Hank Roberts says:

    Answering one of my own questions, one ice core drilled at the center of the basin.

    The atlas shows two other channels I recall mentioned some time back, on the east and west coasts, where bedrock is also below the 100-meter height that would connect to salt water.

    The basin on the east side of Greenland — gray on the map, below present sea level — was mentioned somewhere as a site to watch for sea water eventually floating the ice off the bedrock and melting from the bottom up.

    Great atlas there, far more info on its website than the last time I looked.

  7. 107
    sidd says:

    Let me try to amend matters. SMB has been defined by the moderator. ELA is equilibrium line altitude, the altitude below which the ice is in net ablation and in net accumulation above.

    The reference to Tedesco is from The Cryosphere discuss,

    this paper is under review for publication.

    More seriously, I made an error in attribution for ELA. Although Prof. Box writes at, the article with present day ELA measurement is at

    My apologies for the carelessness and errors.

    I cannot attribute these to smartphone usage. I have no smartphone, do not text due to degenerating wrists and finger joints, and hence all my typing is on a full size keyboard. I think the discomfort of typing forces undue brevity, and is also clearly leading to errors. I shall try and do better, and confine my comments to times when I am physically more capable.


  8. 108
    Hank Roberts says:

    More and new info, and there go my fiction scenarios for a big lake; it seems the Greenland ice didn’t melt last time ’round:

    “The big surprise from the NEEM core is that the Greenland icecap survived the warmth of the last interglacial quite well. It melted back a lot, and the spot where NEEM was located was about 150 meters lower then than now. In general there was still a big icecap in the Eemian, and it was quite a bit warmer than it is now.

    …. Sea level was MUCH higher in the Eemian, and many researchers believed this was due to the melting of ice in Greenland and in Antarctica. The NEEM core calls this into question, and it appears that the large rise in sea level may have come more from Antarctica.”

    Earlier page of info on the NEEM drilling project here:

  9. 109
    Hank Roberts says:

    So for Greenland — what’s different this time around?

    Rate of change in CO2 — much higher rate of change, we know that.
    How long? Up to the politicians, or the free market, or the people, who knows?

    Black carbon/soot on the ice surface — more? faster? different?
    Diesel smoke is a new factor (and oily stuff as anyone who’s worked around diesel smoke knows, nasty sticky stuff that floats on water — it doesn’t sink and wash away like smoke from forest fires, in my experience around diesel generators in camps.

    Soot from wildfires — I’d guess also more and faster than happened at natural rates of change.

    Dr. Box’s Black Ice will get some info about conditions.

    So this finding that the Eemian Greenland ice didn’t all melt suggests the Eemian sea level rise had to be Western Antarctic meltwater, so the Western Antarctic has to be looked at for faster change than so far expected.

    We’re changing air and ocean temperature — faster than nature ever did — so expect the Antarctic melt again, and faster this time.

    We’ve added black soot — faster this time, and more and different — to the Greenland ice surface, so that’s going to be a new forcing on Greenland. Now what?

  10. 110
    Russell says:

    Watts has outdone himself on the Arctic refreeze front with a post that basically contends that the cyanide doesn’t matter if you pour the koolaid fast enough.

  11. 111
    Dan H. says:

    Yes, using an artificially low point to make a comparison is highly biased, similar to plotting global temperatures starting in 1979. Since Arctic sea ice has shown a much wider divergence in summer rather than winter, it is to no surprise that more winter growth has been observed so far this year. A better graphic is to compare the current winter sea ice with past years.

    As you can see, 2013 is currently 7th lowest of the past 10 years, and still climbing. By the time the melt season starts next month (or the following, if it stays cold longer), this years sea ice area may surpass te other three years. It will be interesting though to see how this one-year ice survives the summer. The one thing his posts does show is the large expanse of new ice.

  12. 112

    #112–Strange, I’d have said that the graphic shows that 2013 is currently the 7th-lowest of the past 33 years.

    And I’ll add for good measure that this year has quite a bit of ‘climbing’ to do to surpass last year’s end-of-winter sprint. I don’t see it happening.

  13. 113
  14. 114
    Hank Roberts says:

    > biased similar to plotting global temperatures starting in 1979.

    — equating Watts’s nonsense to climate papers
    — claiming Russell’s pointer is similar to your claim.

    This is spinning and twisting to post yet again a talking point.
    1979 isn’t picked for bias. Watts’s nonsense is.

    1979 … the starting date for most satellite-based global temperature estimates.