RealClimate logo

Putting the recent Antarctic snowmelt minimum into context

Filed under: — eric @ 27 October 2009

Guest Commentary by Andrew Monaghan and Marco Tedesco

Our study published in mid October in Geophysical Research Letters (Tedesco and Monaghan, 2009) documents record minimum snowmelt for Antarctica during austral summer 2008-2009 and lower-than-normal melt for several recent years, based on a 30-year satellite microwave record. Numerous blogs have cited the results as a challenge to previous studies reporting Antarctic warming, while also steadfastly ignoring other studies with similar results (e.g. Barrett et al., 2009). They have overlooked that these studies show that Antarctic warming has occurred mostly in winter and spring, whereas melting of course occurs in summer. And they oversimplify the causality and hence confuse our prediction for the future. We found that the same mechanism that has primarily caused low snowmelt in recent years will likely change in a manner that will enhance snowmelt in forthcoming decades. A brief summary follows.

Map of Antarctica showing number of melting days in summer 2008-2099.
Map of Antarctica showing number of melting days in summer 2008-2009.

Our study demonstrates that low melt years during the 1979-2009 satellite record are related to the strength of the westerly winds that encircle Antarctica, known as the Southern Hemisphere Annular Mode (SAM). When the SAM is in a positive phase – meaning that the belt of winds is stronger than average – it has a cooling effect on Antarctic surface temperatures. The SAM was especially strong in austral spring and summer 2008-2009, and subsequently the 2008-2009 snowmelt was lower than normal. During the past 30-40 years, the SAM has gradually strengthened during austral summer (Marshall 2003), due mainly to human-caused stratospheric ozone depletion. In turn, the increasing SAM has weakened longer-term summer warming over Antarctica. The SAM index is not strongly positive every year of course, and particularly when combined with other atmospheric circulation changes (e.g. a strongly positive Southern Oscillation Index (SOI) – indicative of La Nina conditions) may contribute to anomalously high or low summer temperatures in any given year. The figures shown in our Supplementary Material section in our original paper illustrate this point nicely (below):
1998 Summer1999 Summer
Monthly averaged December-January surface temperature anomalies (K) for 1998 (left, strong negative SAM and SOI) and 1999 (right, strong positive SAM and SOI).

The ozone hole is projected to recover significantly during the next 25 – 50 years due to the Montreal Protocol, which limits ozone-depleting substances used in industrial and household applications. As the ozone hole ‘heals’, the increasing summer SAM trends are projected to subside. As this happens, it is likely that summer temperature increases over Antarctica will become stronger and more widespread because the warming effect from greenhouse gas increases will no longer be kept in check by the dynamic
cooling impact of the SAM.

Therefore, the linkage between the SAM and snowmelt leads to our key conclusion: that enhanced snowmelt is likely in Antarctica as the SAM trends subside during the 21st century and summer temperatures become warmer. Our results agree with studies that have noted cooling and/or slower warming during the past three decades due to increasing SAM trends over the same period. Additionally, our conclusions do not contradict findings showing strong regional warming on the Antarctic Peninsula and in West Antarctica for the past 50 years, and warming over the entire continent for the past century. Our record is limited to the satellite era only, during which ozone depletion has dominated Antarctic summer temperature trends, and as already noted above, the observed warming in the last 50-100 years has occurred mostly in winter and spring. This context is important.

92 Responses to “Putting the recent Antarctic snowmelt minimum into context”

  1. 1
    Randall W. Parkinson says:

    We must get better at marketing our data and corresponding interpretations to the general public. Otherwise, we will continue to loose the publicity battle despite the growning physical evidence in support of climate change.

  2. 2
    PB says:

    Quick comment… your first image in the post has the following typo:
    “Map of Antarctica showing number of melting days in summer 2008-2099”


    [ Fixed, thanks!–eric]

  3. 3
    Alan Robock says:

    I don’t understand the second figure. The legend says both figures are for the same year, 1998, and it refers to left and right, and I see one above the other.

    [ Also fixed.–eric]

  4. 4
    pascal says:

    There is a problem with the legend of the second picture.

    “Monthly averaged December-January surface temperature anomalies (K) for 1998 (left, strong negative SAM and SOI) and 1998 (right, strong positive SAM and SOI).”
    I suppose that the second “1998” should be replaced by another year: 2008 for example

  5. 5
    pete best says:

    The question is: Are Antarctica’s ice sheets thinning beyond IPCC formal projections and if so is this due to or scheduled to be due to warming or increased sensitivity is this makes sense?

    [Response: There are no formal projections for thinning in existing IPCC work. Melt on the Antarctic surface has very little impact at the moment, in terms of dynamic thinning of the ice sheet anyway. Except for the Peninsula region, there hasn’t been enough melt to matter (yet).–eric]

  6. 6
    Craig Cogger says:

    Is the figure on the right for 2008?

  7. 7
    the fritz says:

    the recent Antarctic snowmelt minimum

    It’s nice to recognise it

  8. 8

    Ah, context.

    At home at RealClimate–in the denialosphere, not so much.

    Thanks for this informative post!

  9. 9
    CM says:

    The second and third figures are for 1998 and *1999* according to the auxiliary material at GRL.

  10. 10
    sidd says:

    Re: Antarctic Glacier thinning

    A nice paper by Pritchard et al (Nature, v461, pp971 et seq.) shows how the rate of thinning has increased in both Greenland and Antarctica. As for the IPCC, I do not see any very specific predictions of future ice sheet mass balance in AR4, Section 15.3.4 or Sec. of WG II available at

    They do say that there is “medium confidence that at least partial deglaciation of the Greenland ice sheet, and possibly the WAIS, would occur over a period of time ranging from centuries to millenia for a global average temperature rise of 1-4C”

    [Response: See also my comment above.-eric]

  11. 11

    Curious about ozone here. When you say “ozone depletion has dominated Antarctic summer” – does that have the effect of making the summer much cooler than otherwise? Because with a healthy ozone blanket, would not the Antarctic summer be much warmer?

  12. 12

    Of note is that we (over at World Climate Report) were one of the first places to highlight this finding, but never made any mention of how it meshed with Antarctic temperature trends.

    But, your explanation brings up a question: What is the relationship between SAM and Antarctic precipitation? If overly strong SAM is inhibiting warming, is it also inhibiting precipitation?



    [Response: Chip, indeed you did highlight it, but in a highly misleading way. Hence this post (though I don’t know that the authors were specifically responding to you). As for circulation and precip, yes one might expect what you suggest, but at least with available data, there are overall no statistically significant change in snowfall during the last 50 years (see here.–eric]

  13. 13
    Eric L says:

    The deniers I’ve seen referring to your work have given it as a refutation of data from, e.g. GRACE, which show accelerating ice loss over recent years. My understanding is that your ice melt index is an approximation for ice melting at the surface of the sheet but is not a good approximation for total ice loss and it is not in contradiction with data showing accelerated ice loss. This would be a good thing for you to clear up.

  14. 14
    Mark says:

    “My understanding is that your ice melt index is an approximation for ice melting at the surface of the sheet but is not a good approximation for total ice loss”

    It’s about as good (slightly more, but not massively) as sea ice extent is an approximation for total ice loss/gain. The difference being 15% coverage is counted as highly as 100% when it comes to extent, so the more broken up the ice (which is more likely in the summer and more likely still for thin ice) the worse extent is an analogue for ice volume.

    [Response: Huh? Let’s not confuse things! Sea ice is floating ice 1-4 m thick. The melt we’re talking about here is melt on the land ice — the hundreds-of-m-thick glacier ice (the Antarctic ice sheet), and the floating ice shelves (also hundreds of m thick, and NOT the same thing as sea ice. The point is that melt on the surface of the ice sheet is almost all refrozen, so the ice melt index has little or nothing to do with mass loss. These are totally separate things (except when lots of melting on ice shelves leads to collapse of those shelves, as happened on the Larsen ice shelf in 2002.–eric]

  15. 15
    J Buote says:

    “We must get better at marketing our data and corresponding interpretations to the general public. Otherwise, we will continue to loose the publicity battle despite the growning physical evidence in support of climate change.”

    Just a quick note re: the quote above. This site, and many others, are getting the evidence out there and I for one can’t thanks everyone here and to all of those contributing updates on the science.

    I’ve personally linked this site many times in the last while to counter the endless assault of skeptic arguments that tend to dominate forums. IT is working.

    Just thought you’d like to know…

    Thanks again,
    J Buote

  16. 16

    Thanks guys for the info. Very worthwhile site which I too send out to others for converting the doubters.
    One question I would like to ask and that is has the cumulative ice melt over the last 50 years been incorporated into the data? By this I mean that 50 years ago the ice cover would have been x % more or less than the % cover of today, or has it been pretty stable with annual increase and decrease on a seasonal basis? It would seem that from evidence that I have been given by a good friend of mine (former base leader at Port Lochroy in the mid 1950’s when they first found indications of an anomaly in the atmosphere became apparent, later called the Ozone hole.) that the ice cover is very similar today as it was way back then. This is supported by photographic evidence showing the ice sheets behind the base and little has changed. Any views on this?

  17. 17
    Glenn Tamblyn says:

    Here is a comment from the paper by Perlwitz et al that you cite
    “But just how influential a full stratospheric ozone recovery will be on Southern Hemisphere climate largely depends on the future rate of greenhouse gas emissions, according to the study. Projected increases in human-emitted greenhouse gases such as carbon dioxide will be the main driver for strengthening the positive phase of SAM.”

    Does that mean that it is a race between ozone recovery winding back SAM and allowing warmer air into the heart of the continent, and general AGW strengthening SAM and maintaining the blockage so that the centre of the continent doesn’t warm as much?

    [Response: Yup, this pretty much sums up the situtation. Of course, it is far more complicated than that, as explained e.g. in our (our paper, and a more recent one by Turner et al. about the non-SAM related variability and its relationship to sea ice changes.-eric]

    If so that is obviously a fairly complex prediction to be having to convey to the broader world. The comment earlier – “We must get better at marketing our data and corresponding interpretations to the general public. Otherwise, we will continue to loose the publicity battle despite the growning physical evidence in support of climate change.” may be particularly appropriate in this situation.

    Thanks for a very informative post guys!

  18. 18
    Vern Johnson says:

    Stronger summer winds cause less melting? ( they might accelerate sublimation). Please also explain why increased exposure to ultraviolet rays due to ozone depletion might cause stronger west winds in Antarctica. Otherwise there will be skeptics galore about this very paradoxical result. Time to move a few hundred polar bears to this locale before they lose their arctic habitat. With larger ice mass at the south pole compared to the north, this could cause some wobble and unpredictability re the earth’s rotation and the stability of the seasons, “summer” being one of these.

    [Response: Who said this had anything to do with ultraviolet? There’s a huge literature on this (start here), and one of our very first posts at RealClimate discusses it (here).–eric]

  19. 19
    Mr Sh says:

    Thank you for good co0ment. I am always irritated by the arguments of deniers without correct understanding of new scientific findings. Academic communities and journalism are, I believe, responsible to providing their correct interpletation.

    Thanks again from Japan

  20. 20
    Mark says:

    “The melt we’re talking about here is melt on the land ice ”

    OK, that wasn’t clear.

  21. 21

    I am still not sure whether you are saying that the Antarctic ice is increasing or decreasing. If the melt is refreezing then there should be no change.

    However it would seem that the ice mass is increasing due to the Ozone hole, and so there is a negative contribution to sea level rise.

    Why can’t you come clean and say what you really think?

  22. 22
    Lawrence Coleman says:

    May I take this opportunity to drive yet other nail into the ACC sceptics coffin. Take a look at science daily. Two links — : and this one —
    They show highly accurate sediment data from a Greenland lake not subject to glacial erosion from the past 2 ice ages and 3 interglacial periods spanning roughly 200k years. They indicate very strongly that since about 1950 ( and I would say that date is pretty much spot on with all the reports I have read) there have been marked changes in the marine ecology specifically diatoms and algaes that have not occured ever during the past 200k years, despite ice ages and cyclical fluctuation in solar maximum’s, the earth’s tilt and associated wobbles etc. This is the first time that quatic indicators in sediment are showing the unmistakable affects of global warming. Sorry skeptics – but since these are NOT natural indicators, the only other possiblity is antropogenic! And around 1950 is when I have seen most of the worlds glaciers really begin to retreat in earnest as well and the chart of CO2 really beginning to take off like a rocket. Some feedback on this would be most welcome.

  23. 23
    Cardin Drake says:

    re: 22

    The problem is that actual thermometer readings from close to the lake you are referring to show no warming since 1950. That would seem to be conclusive proof that the changes in the marine ecology have another, non-temperature related cause.

    [Response: I haven’t read the paper yet so I can’t comment. I’m familiar with the location though, and I doubt very much that there is no warming since 1950 there! Can you point to some data to backup your claim?–eric]

  24. 24
    Mark says:

    “That would seem to be conclusive proof that the changes in the marine ecology have another, non-temperature related cause.”

    Not really.

    Somewhere nearby gets more rain because of higher water content of the air and this lake gets more runoff into it or less rain because it falls elsewhere that isn’t the catchment.

    Migrant birds now turn up earlier and stay longer because where they leave from is changing its climate. These migrants eat the fora of this lake at different times and for longer.

    Or they may move to another lake where the food is more plentiful and the larger predator species are not culled by the birds.

    That is even if the lake really hasn’t warmed since 1950…

  25. 25

    Re #23;

    Sure, Eric, check out the temperature from nearby Clyde, N.W.T. (as represented in the GISS temperature record), as we show in our World Climate Report story.


  26. 26
    PeterPan says:

    So Antarctica is losing ice (Velicogna 2009) but it is not from snowmelt (Tedesco & Monagan 2009). So where is Antarctica losing it? May it be from ocean-driven ice shelves melting?

    Thanks for the article, it’s very helpful against the constant disinformative campaign from skeptics, specially when it comes from the authors themselves.

  27. 27
    Mark says:


    How come the graph you’ve got has the temperatures of 1-5C when the NASA data you link to as the original data shows temperatures -10.5C down to -14C.

  28. 28
    Mark says:

    Also this:
    “There is a bit of press covering a just-published paper that concludes that the current climate and ecological conditions in a remote lake along the north shore of Canada’s Baffin Island are unique”

    doesn’t really help, does it. If it’s unique, you can’t really draw any conclusions except that every other place doesn’t do that. About all you can say is “It isn’t doing this!”.

  29. 29
    Adam Gallon says:

    Re Comment by Lawrence Coleman — 28 October 2009 @ 8:10 AM
    Have the authors analysed the silt for chemical residues, eg DDT? This was shown to accumulate in lake silt in a Swiss study published in November’s Environmental Science & Technology.
    There is an old DEW Line station near to the the lake, so spraying against insects would be expected.

  30. 30
    Luke T says:

    re: #26/PeterPan
    So Antarctica is losing ice (Velicogna 2009) but it is not from snowmelt (Tedesco & Monagan 2009). So where is Antarctica losing it? May it be from ocean-driven ice shelves melting?
    Exactly. Pritchard et al. 2009 (Nature)infer that grounded ice/glaciers are responding to ice shelf collapses in addition to ocean-driven melting — “a widespread phenomenon that does not require climate warming sufficient to initiate ice-shelf surface melt.”

  31. 31

    Mark (re:27),

    We showed summer (JJA) temperature from Clyde–which was the season most germane to the findings discussed by Axford et al.


  32. 32
    colin aldridge says:

    A very cogent and credible analysis. Not inconsistent, as you say with global warming or indeed the IPCC sea level assessment but inconsistent with the more extreme forecasts of increases in sea level over say the next 30 years.

    Care needs to be taken however when we justify this sort of finding with “natural variability”.. in this case SAM, SOI and La Nina… since this begs the question of what impact does this natural variability have on temperatures. IMO, until we can build such natural variables into our GCM’s and model these impacts of natural variables on historic temperatures we will not have a truly robust basis for predicting future temperature caused by CO2and will have to live with the very unsatisfactory IPCC forecast of 1.5 to 5Deg C for a doubling CO2

  33. 33
    Mike Smith says:

    Re #25, #23, #22:

    First, the general trends across /Canada/ for the 50 years can be seen more clearly in:

    “Temperature and Precipitation Trends in Canada During the 20th Century” Zhang et al. (2000).

    How unusual such a steady temperature is in a global context can be seen in figure 3.9 on page 250 of Chapter 3 of the “IPCC’s Fourth Assessment Report (AR4), The Physical Science Basis”. This is for a trend of over 100 years!!! The only significant slight cooling in the whole of the Earth in the 20th century was in part of the North Atlantic. So picking a station of the North East coast of Canada is slightly cherry picking. Especially the period chosen (i.e. over the whole century).

    The IPCC also looked at 1979-present (FAQ 3.1 p253): Showing the region has warming. Compared to the data you link to, this matches:
    (roughly 0.17 deg C/decade)

    Papers looking at the region, such as:
    “Recent Cooling along the Southern Shore of Hudson Strait, Quebec, Canada, Documented from Permafrost Temperature Measurements” Allard (1995).
    say that the previous lack of warming is due to the Thermohaline circulation.

    In general, “no warming since 1950” isn’t true, it’s warmed considerably over the last 30 years.

    Maybe someone else could explain or link further? :D


    PS I study computational neuroscience, not climate science! Please correct my interpretation :D Also maybe a link to the effect of the Thermohaline circulation on such a temperature record?

  34. 34
    Mark says:

    Chip, germane in what way?

    After all, a study into the temperature differences from 1998 to 2008 is germane to a study of how it’s cooled.

    It is only germane however in that picking the data that way is the only way to get a decade long cooling, therefore picking other data (like, say, 1988 to 1998) would not help a study of how it’s cooled.

    It comes back to the “Why the continued interest” thread: would your thesis have worked if you had used DJF instead? Or an Alaskan site. Or…

    It is also incorrect to label a point to data you used that doesn’t point to the data you used.

  35. 35
    Armand Chase says:

    IS there any merit to these articles denying AGW?



    [Response: No. Same old fudged and misleading graphs – correlations that mysteriously end in 1980, implicit climate sensitivities beyond anyone’s wildest imaginations (when comparing CO2 with temperature over short terms) etc. etc. – gavin]

  36. 36

    Mark (re: #34),

    We picked Clyde, N.W.T., because it was the station closest to the lake that was being studied and thus presumably best represented local environmental conditions. We showed the summer temperature (and our graph was clearly marked as such), because that was the season that Axford et al. discussed (and produced their own reconstruction of). And I linked to the GISS data page for Clyde so if you went there you would know what you were getting (rather than linking to a page full of nothing but data without description).

    Seems reasonable to me!


  37. 37
    PeterPan says:

    #30 Luke T,

    Thanks, Luke! Now I have my three-piece jigsaw puzzle complete! ;-)

    # 33 Mike Smith,
    Baffin Island (where sediments are extracted from in Axford et al., 2009) is not covered in Zhang et al. (2000) (though the warming seems to be widespread in the southern area that they cover). I havent’t read the article, but from their press release, the attribution part seems to be that cold species survived previous warming periods (being even abundant) but not this one. (If that’s all), the reason might be the rate of the current warming, but other (surey anthropogenic) stress may share the responsibility.

  38. 38
    Ben Lawson says:

    #36: Chip thrashes about for anecdotal weather patterns that counter a specific study result and says: “Seems reasonable to me!”
    What a surprising conclusion from an off-topic uninformed denialist looking for an excuse to reject scientific evidence…

    You have no idea whatsoever as to what is scientifically “reasonable” and what isn’t, do you?

  39. 39
    Hank Roberts says:

    > scientifically “reasonable”

    Now, now, he’s just doing what’s required in his job, “advocacy science.”

    You start with the belief, then pick the facts to support it.

  40. 40
    MarkB says:

    Chip (#36),

    Seems that your single station data would be at odds with most other Arctic station data for that time period.

    Of course, anomalous stations that don’t show significant warming (about 0.1 C in this case) are not questioned by “skeptics”.

  41. 41
    JCH says:

    How’s the ice doing on the island that has not warmed?

  42. 42
    Mr Sh says:

    #36: Chip, why you picked up the summer record in the context of AGW while the guest point out “Antarctic warming has occurred mostly in winter and spring, ..”? Of course, the paper you refered talked about the arctic lake, not the antarctic, but it is quite doubtful to neglect these unless there is an evidence that the arctic warming never occurs in winter and spring.

    [Response: All I can say in response to all this back and forth regarding the Baffin paper is I guess I’d better read it and write something up. None of this speculation for or against Chip’s point is likely to be enlightening. Whatever complaints one might have against Chip’s take on all this, at least he has evidently read the paper! How about readers let this discussion die until I have time to put up a post on it. –eric]

  43. 43

    #31 Chip, I must disappoint I am afraid, I worked Clyde River in 1980-81 at its second station situated
    higher across the Bay the station moved with the village in the 70’s. The original station was much lower at sea level, knowing inversions Clyde’s temperature record should not to be used ! and again, in the late 80’s the station moved to the airport! Another few miles away. However did the research need be a little more thorough… Try again!!!

  44. 44

    Whoever did the research, of course. Clyde is a most beautiful place with mysterious Fjords just near by.

  45. 45
    Lawrence Coleman says:

    28: Mark. It’s unique in respect that the sediment shows an unbroken linear timeline back 200ky. Most other glacial lakes only go as far back as the last ice age 50ky when the weight of the moving ice would just scour clean the lake bed removing any old sediment in almost all cases. Mark have you read the report?

  46. 46
    Lawrence Coleman says:

    24:Mark. Everything you mentioned would be cyclical at least over the last 200ky except when you mentioned migratory birds may be visiting the lake sooner (and presumably eating more midge larvae?) because the climatic conditions are changing in response to CC. There may be grain of possibility there?. Whichever way you cut it..still comes back to climate change doesn’t it!

  47. 47
    Lawrence Coleman says:

    Re 29: Adam Gallon. DDT would indeed affect the midge larvae more than diatoms due to bio-amplification. It seems that the more ‘silica’ in the water the faster the growth rate of diatoms due to thats what their shell is make of. I would imagine that in an enclosed lake the amount of silica in ppm would be reasonably constant.

  48. 48
    Mark says:

    Chip: “Seems reasonable to me!”

    Well, call me skeptical.

  49. 49
    Theo Hopkins says:

    Maybe I missed it, but what has ozone got to, do one way or the other, with antarctic temperatures? Can someone point me to an answer.

  50. 50
    Mangeclous says:

    Hi Eric,

    Greetings in my first comment and thanks for the huge effort put in this site or, at least, in its huge positive role as information source for us non-scientists.

    After reading this post I keep a couple of doubts. First, why is snow melt decreasing if the ozone layer is already increasing? Shouldn’t there be a progressive trend the other way round? And second, shouldn’t we notice a similar effect in the North Pole, annular wind-wise?

    Thanks in advance and sorry for my English and lack of knowledge,