RealClimate logo

Is Antarctic climate changing?

Filed under: — eric @ 25 August 2006

Is the Antarctic ice sheet getting bigger or smaller? Is it warming or cooling?

As we’ve reported in earlier posts (here and here), getting accurate answers to these questions is non-trivial, because the available instrumental data remain sparse and generally date back only a few decades, at best. While modern satellite-based techniques such as laser altimetery and gravity anomaly measurements provide important information on very recent changes, to get at the longer term we must rely on less direct methods. In the last 5 years or so, an effort has been under way, much of it under the banner “International Trans Antarctic Scientific Expedition” (ITASE), to do this by collecting many dozens of ice cores from across the Antarctic continent. Two papers out this month represent the first major compilations of results from these efforts. The first, in Science on August 11th, provides a new estimates of Antarctic snowfall changes over the last 50 years. The second, in Geophysical Research Letters (August 30th) provides the first statistical reconstruction of Antarctic temperature change, extending about 200 years into the past.

In the Science paper, Monaghan and others show that there has been no significant change in Antarctic snowfall in the last ~50 years. This is a potentially important result because most calculations suggest that as the globe warms, polar snowfall should increase, somewhat mitigating the sea level rise that is expected to result as the margins of the ice sheets melt and thin. The new results differ from those of Davis and others, who used data from the European Remote-Sensing Satellite (ERS-1) and ERS-2 satellite altimeters to estimate that the Antarctic ice sheet had gained enough mass between 1992 and 2003 to slow sea level rise at a rate of about 1 cm/century. The two sets of results are not necessarily in conflict, but do suggest that the intepretation of Davis et al. that the observed change is due to snow accumulation increases may need revision. [See also our earlier post on the use of gravity measurements to determine mass changes in the Antarctic, here. There are also some new results on mass change in Greenland, reported by Chen et al., out last week in Science Express (here), based on measurements from the GRACE satellites, essentially confirming earlier results of major changes on the Greenland ice sheet.]

Monaghan et al.’s estimates are based on a clever combination of ice core measurements of annual snow layer thickness, and model determinations from the ERA-40 climate model Reanalyses. The reason for using both observed and model data is that, while the sixteen ice cores they use come from all over the Antarctic continent, they nevertheless remain only point sources of information. To obtain useful averages over broad geographic regions, it is necessary to interpolate between the different ice core records. Monaghan et al. use the model results to guide this interpolation. The technique is not dissimilar to other climate field reconstruction methods that we’ve discussed frequently on RealClimate. The chief difference is that Monaghan use the output of a climate model, rather than direct observations, to determine the patterns of covariance in the climate field (in this case, snow accumulation rate). This technique allows Monaghan et al. to extrapolate the model results (which are realiable only for the period 1985-2004) back in time as far as the last International Geophysical Year (IGY), 1957, with a great deal of confidence.

Extrapolated to the future, a possible interpretation of Monaghan et al.’s results is that sea level rise in the future would be even greater than otherwise expected. However, we caution that such extrapolation is probably not warranted. After all, the relationship between temperature and snowfall is based on physical arguments and atmospheric dynamics calculations, not only on data; just because we cannot yet detect an increase in snow accumulation does not mean that it will not happen in the future. Furthermore, while Southern Hemisphere temperature have increased, on average, along with the rest of the globe during the last century, the data are rather sparse at the higher latitudes. And although there is evidence for warming in the mid troposphere over Antarctica from radiosonde data over the last 30 years, it is also clear from satellite data that surface temperatures decreased during the years 1982 through 2002. In fact, one can make the case that the timing of accumulation changes agree rather well with the observed timing of surface temperature change in the Antarctic. Monaghan et al.’s results shows that while there has been no net change, snow accumulation actually increased from IGY until the mid-1980s, and decreased thereafter. Similarly, a simple average of the weather station data show the same pattern for temperature — average warming prior to the mid 1980, and average cooling thereafter. [And as we have reported previously, the recent cooling is quite well understood and not expected to continue over the long term.] It is thus premature to conclude (as the paper does) that we may need to revisit GCM assessments that show increased precipitation over Antarctica in conjunction with projected warming in the future.

The results of the Geophysical Research Letters paper by Schneider and others — on which I am second author — are relevant here. We’ll report on this at greater length once the paper is published. As Monaghan et al. did, we used ITASE ice core records. In this case, our goal was to determine temperature, rather than snow accumulation history, and we used the available Antarctic weather station data (much more complete for temperature than for accumulation) rather than model results. We found that the stable isotope composition of the ice cores mimics the observed temperature pattern — warming between the 1960 and 1980s, cooling since then. Using the stable isotope records to extrapolate farther into the past, we find that Antarctica has warmed, on average, in the last century, along with the rest of the globe. Once the accumulation records have been extended, it will be possible to revisit the relationship between temperature and accumulation trends. I suspect that there will not be any surprises here.

73 Responses to “Is Antarctic climate changing?”

  1. 1
    Tas says:

    Eric, nice post thanks – draws two current results together well. Just a comment or two about the issue of warming, lack of accumulation, and what this might mean for projections (and for the reader’s benefit I should disclaim that I am a coauthor on both papers).

    Specifically, you state that it is “premature to conclude (as the paper does) that we may need to revisit…”. The basis for this remark in Monaghan et al. is the recent work of Turner et al, which showed very strong mid-tropospheric warming since the early 1970s, albeit not at the surface. Now we find that this is not accompanied by accumulation increase. It is this observation that leads to the fairly cautious statement that it “may be necessary to revisit GCM assessments”. Note the word “may”. It is certainly necessary to evaluate whether GCM results are consistent with these new longer and more complete observations, and to understand what lies behind any discrepancies detected.

    Of course it has been pointed out by others that the simple thermodynamic argument of warmer=wetter is missing an important part of the story. Particularly in Antarctica, moisture transport, driven by trends in atmospheric circulation, plays a key role, and this is of course why modeling is necessary for understanding. Here also is where another connection can be made with the Schneider et al. paper, which underscores the impact on temperature of circulation changes, particularly associated with the Southern Annular Mode (SAM).

    As you note, the results taken together show reasonable concordance between the surface temperatures and the snow accumulation. I tend to agree that as we see longer reconstructions there may not be surprises, inasmuch as the surface temperature and accumulation link will probably be robust: particularly where cyclonic precipitation dominates.

    But it still leaves me wondering then about the disconnect between surface and mid-tropospheric temperature trends and their respective connections with moisture transport. How does this mesh with the GCMs and what does it mean for projections of snow accumulation in Antarctica? I would be interested to hear comment on this.

    [Response:Hey Tas! Thanks for dropping by RealClimate. I agree with your comments. The major point I’d make in response is that the lack of any statistically significant trend in accumulation in the 50 year record of Monaghan et al. has to be understood in the context of the vary large variability, as Andy [Monaghan] explained very nicely in the NSF press release. The same should be said of the temperature records in our [Schneider et al.] paper. So the lack of any statistically significant trend in accumulation in no way says that tropospheric warming is having no impact on snowfall. It may simply say that other things — i.e. changes in moisture convergence — are having an equally large impact, with a net result of no obvious change. The question then becomes how these things play out in the long term. My suspicion is that dynamics rules the day in the short term, but that over the long term, therodynamics wins out — if it gets a lot warmer, there will be more moisture available and on average there will be more snow. It could happen that the dynamical changes could cancel out this effect, but it seems unlikely to me. Having said all that, it is certainly worth exploring more fully, and I, like you, would be curious what the response will be from the dynamics community to both these papers. — eric

  2. 2
    Rob Miller says:

    In an Antarctica that is moving entropic with respect to a molar heat of fusion in re water there, and the recent appearance of high-atmosphere nacreous cloud forms, very rarely formally observed in Antarctica, it would seem that that continent is radiating a large portion of its heat through a more or less transparent upper atmosphere directly into space, producing a remarkably cold high atmosphere, layering above a putatively cooling but still a relatively much warmer lower atmosphere. The increasing differential of the potential energies between the lower and upper atmospheres of that continent form the crux of these issues, not whether the ice cores are cooling or warming, taken in and of and by themselves.

    [Response: In general, I would agree, and Monaghan et al. were quite justified in emphasizing the upper-tropospheric warming as opposed to the surface cooling, in discussing the snowfall/temperature relationship. I’d note, though, that precipitation in the Antarctic is peculiar, in that much of it actually falls from a clear sky, and the near surface temperature may actually be quite important, especially in the dry continental interior, in determining whether snow actually forms and precipitates or not. A problem that is most certainly not treated very well in GCMs are the details of ice formation at cold temperatures — it is difficult to do because it depends on getting the humidity just right, as well as the aerosol concentraiton and size distribution. These things — and the snowfall itself, since the rate is so low — are very hard to measure, and good analyses are few and far between. — eric ]

  3. 3
    Bryan Sralla says:

    “Average (Antarctic) warming prior to the mid 1980’s, and average cooling thereafter.”

    This is interesting to compare with the Alaska temperature record. A 2005 paper by Hartmann suggests that following a phase shift in the PDO in 1976 (quick warming), there has also been a slight average cooling trend in Alaska.

    It may be significant to note the warming at these high latitudes does not seem to be occuring gradually, but rather in quick pulses (a quick adjustment from one regime to another), and possibly associated with larger scale weather pattern (ocean SST) changes.

    Could this pattern of temp changes in both the high lat SH and NH records suggest that heat uptake in the oceans and exchange with the atmosphere is much more complex than currently represented by the IPCC models? What effect could such have on their ability to project regional climate changes?

    [Response: Linking the 1976 so-called “regime” shift in the North Pacific to anything in the Antarctic is a major stretch. For my part, I consider Thompson and Solomon — attributing recent dynamical changes in the Antarctic to ozone depletion — as the best work on recent Antarctic climate change at the moment. See our earlier post on this (and their paper of course, linked within that post): here. -eric ]

  4. 4
    Eric (skeptic) says:

    The Dec 2004 discussion linked above suggested that stronger westerlies were responsible for preventing Antarctic warming. Is this result still supported in current models? What is the current state of antarctic weather modeling? This paper points out that lowering the mixing ratio has to be lowered beneath the normal model threshold to obtain fewer or thinner clouds in the model to match observations. This to me illustrates the danger of parameterizing weather rather than simulating it in greater detail. But after the model tweak that extra snow has to fall which doesn’t seem to match the reality indicated above.

  5. 5
    Rob Miller says:

    A factoid might putatively be that we will need to make an antithetical and corollary sister for the PETM. Unfortunately, we haven’t yet made a name for what is coming in our supposed future, as sister to our present, so an acronym does not exist like that which exists for the Paleocene-Eocene Thermal Maximum. If a thermal minimum catastrophic climate event were to originate in the Island of Antarctica (no easing landlock of ice to shorten the pain), there might be no time machine in which we could ride that we might be able to find ourselves in any future. And that is why all this science is about.

  6. 6
    Eric (skeptic) says:

    To make my weather question more specific, has the modeling improved from this result where the persistent trough was underestimated until weather was added (see ).

  7. 7
    Sally says:

    Re#5 “thermal minimum catastrophic climate event” as in ice age? I don’t suppose anyone coined the phrase Paleocene-Eocene Thermal Maximum before it happened either.

  8. 8
    Rob Miller says:

    man did not exist during the PETM. no one that i know of ever mentioned, really, a pending thermal minimum catastrophic climate event. besides, we invented the word “catastrophe”. any “age” can go any way “it” wants, inasmuch as me, and everybody else, has got next to zero understanding of the dynamic involved. i just think that ice, catastrophic for humanity ice, is more likely.

  9. 9
    Rob Miller says:

    narwals talk about the north, no ways dream about the south

  10. 10
    Rob Miller says:

    most people understand the idea of “mars” more than they understand the idea of “antarctica”, although it is the fifth largest of seven continents, and mitigates for a large portion of the food and weather available to our species on this planet. to most it might as well be in another solar system. this only points out our own superficiality and a putative argument for our species dissolution or extermination. its really a very bad situation for us.

  11. 11
    Mark A. York says:

    Shishmaref, Alaska is falling into the sea. This seems to me, a long term indicator since the people and their ancestors have never experienced anything like this before.

  12. 12
    Mikko says:

    Thanks for the blog, it’s been my main source of climate information for quite some while now. What I’ve been wondering is, would it be possible in most cases to have a very short ‘abstract’ providing the conclusions your post has in the beginning so that people with limited amount of time could check the conclusions and then decide whether to read the whole post, or just parts of it?

  13. 13
    John L. McCormick says:

    RE #5,

    I am struggling to get your drift, thusfar, but I do offer [an antithetical and corollary sister for the Paleocene-Eocene Thermal Maximum (PETM).]

    It is:

    Paleocene-Anthropocene Intergenerational Nightmare (PAIN).

    Given the information I have processed from exensive reading and listening, I conclude PAIN is in our future.

    [Response:Let me just say that I’m enjoying this discussion, even it is getting a bit silly. -eric]

  14. 14

    Re #13 That should be Pleistocene-Anthropocene not Paleocene, otherwise I completely agree :-(

    See and

    It does leave the Holocene out on a limb, but it should really be put in the Anthropocene because that was when the damage started to be done.

  15. 15
    Steffen Christensen says:

    With respect to Bryan Sralla, comment #3:

    Hartmann and Wendler’s paper ( does seem to show nonlinear stepwise trends in Alaskan temperatures. However, “average temperature of the 6 climate regions”, plotted in their Figure 5, is decidedly NOT the same as average temperature of the land surface of Alaska. Consider their Figure 1, showing the climate regions of Alaska: as you can clearly see, the Interior region is maybe 60% of the land area of Alaska – the other 5 regions are condensed into the remaining area. Averaging numbers with such sampling regions does not an average land temperature make.
    Picking an arbitrary high-water mark, here 1976, raises the possibility of a well-known statistical phenomenon of regression towards the mean. You can always make it look like the temperature is cooling by choosing a nice hot year to start your clock from – and only consider years subsequent to that. So, has the climate in Alaska cooled since 1976? It is important to consider whether more recent data support the hypothesis. The paper you cite uses data up to 2001, their last year of record. We now have data up to 2006. I went to the NASA GISS site for climate change, and served myself up a graph of climate change in Alaska since 1976. The graph is here: . I chose as my reference period the years 1977 to 1987, and as my period of consideration the years 2001 to 2006, the “new data” since the paper was published. If it is true that Alaska is cooling, it must perforce be true that it will continue to cool to preserve the hypothesis. I used the small-scale binning – 250 kilometer radius – so that we get some good spatial resolution on Alaska. Guess what? ALL of Alaska is warmer than the reference period, save the Seward peninsula – which has not cooled; and a region on the border with Canada, overlapping the Porcupine river – which has not cooled. Of course, the GISS data server does not compute probabilities for statistically significant cooling/warming; and the spatial resolution flickers around if you pick different starting or ending years, but the trend is fairly solid – Alaska doesn’t seem to be cooling continuously since 1977.

  16. 16
    Bryan Sralla says:

    Re: #3 Eric, thank you for the response. Just in the way of clarification, I was in no way trying to link the PDO with anything in the Antarctic.

    Instead, it was my intention to point out some interesting contemporaneous data in the high lat NH. I may be wrong for infering this, but it appears to me that the IPCC models are missing many of the salient details in both the Arctic and Antarctic. The spacial and temporal pattern of climate change at both ends seems to be more complicated than advertized. This would seem an important observation worth pondering.

    [Response: Point taken; I agree with you. -eric]

    Re: #15, I have invited Gerd Wendler to respond to your comments. I hope he will.

  17. 17
    John L. McCormick says:

    Thank you, Alastair, I accept your edit and hereby launch the corrected
    [antithetical and corollary sister for the PETM].

    Pleistocene-Anthropocene Intergenerational Nightmare…..PAIN

    It might sound silly to some but it does it for me.

  18. 18
    Sally says:

    Holocene Anthropocene Latent Termination …. or HALT.

    [Response: Sorry, but PAIN is the best one I’ve heard yet. – eric]

  19. 19

    Actually the sequence of Phanerzoic eras, as opposed to the much shorter epochs such as Holocene, Eocene etc, are the Palaeozoic (Old Life), Mesozoic (Middle Life), and Cenozoic (New Life).

    The P-T mass extinction occurred at the Palaeozoic/Mesozoic boundary, and the K-T extinction of the dinosaurs occurred at the Mesozoic/Cenozoic boundary. The megafaunal extinction which began at the end of the Pliestocene (the penultimate Cenozoic epoch)was just the start of the anthropogenic mass extiction which will lead to the Cenozoic/Anzoic mass extinction. Of course, the name Anzoic Era means “No Life” but then no one will be around to name it :-(

  20. 20
    pat neuman says:

    re 3.


    I had discussions with Brian Hartmann about his faulty interpretation of the Alaska climate station data (while he was at the climate-change yahoo group, which no longer exists). My summary of annual climate data (1950-2005) that I made in Feb 2006 is below.

    Annual temperature data at 18 climate stations in Alaska show 3-6 Deg. F increases from 1950 through 2005. Of the 18 climate stations in Alaska, 6 are located near the coast and 18 are inland.
    Annual temperature plots for all 18 climate stations in AK are at:

    Above summary at climateArchive, message #1800 (Feb 15, 2006).

  21. 21
    Bryan Sralla says:

    Re #20: Thank you for this Pat. The Alaska Climate Research Center is showing a 3.5 degree F warming over this time period. That falls into the same error bar as your estimates. As for the choice of stations, these guys live and work in Fairbanks, Alaska and have no doubt visited many of the field stations personally. I would really encourage you to publish a paper in rebuttal, showing why their choice of stations was inappropriate, and your choices are more robust. Your interpretations of the data are very beneficial for your personal study, but have not been subject to editorial or peer review from the wider community.

  22. 22
    pat neuman says:

    re 20. It should be of the 18 climate stations in Alaska, 12 are located near the coast and 6 are inland.

    [Response: This discussion suggests a RealClimate post on Alaskan climate at some point — and on the PDO — would be useful. We’ll think about this. For now, I’d be delighted if commenters can generally restrict themselves to the topic at hand. I’m not chastising, just politely asking. — eric]

  23. 23
    Jeff Weffer says:

    Antarctica has been frozen over for 35 million years. The average annual temperature at the South Pole is -49.5 C. It is not going to melt.

    [Response: What you say is very true on average, utterly wrong in the important details. Among other thinsgs you should look at our post on Antarctic Peninsula glacier retreat, if you are interested in getting the details right. –eric]

  24. 24
    Deech56 says:

    Hi Everyone,

    This item seems to be making the rounds of the right-wing blogosphere: Greenland’s Glaciers Have Been Shrinking for 100 Years: Study. Any ideas? Does this show anything more earth shattering than on this one island glaciers have shrinking for about a century? An interesting comment that I wonder about is: “The shrinking of the glaciers since the 19th century is ‘the result of the atmosphere’s natural warming, following volcanic eruptions for example and greenhouse gases, created by human activities, which have aggravated the situation further,’ he said.” Somehow, this is interpreted in places like FreeRepublic as proving Al Gore wrong.

    [Response: I’ve not yet looked at that particular study. It is probably a good study, being totally misinterpreted. Perhaps the most relevant response is simply to point folks to our several posts on the general fallacy of using natural climate changes associated with natural forcings as evidence against the responsiveness of the climate to other (e.g. anthropogenic forcing). In particular, here and here.–eric]

  25. 25
    pat neuman says:

    re 21.


    My work on temperature data at climate stations was explained and available for peer review at:

  26. 26
    Steve Bloom says:

    Re #23: Jeff, you need to broaden your sources a bit to include some that focus on science rather than supporting a priori determinations. In addition to Eric’s note, see e.g. this. It seems that the more we find out about past climate, the less stable it seems.

  27. 27
    Steve Bloom says:

    Re #21: Your mastery of Macropielkean is most impressive, Bryan. :) But just out of curiosity, what makes you think that the PDO shift in 1976 wasn’t a global warming signal? I don’t recall any peer-reviewed work on the ACRC site demonstrating that. Perhaps you could point me to it?

    [Response: Careful! What makes you think it was a global warming signal? I see a definite need for a “what the heck is the PDO, anyway” post… -eric]

  28. 28
    Bryan Sralla says:

    Eric, is most of the Antarctic tropospheric warming being observed in the winter months, offset somewhat by negative temperature anomolies during the summer months? What about snowfall anomolies broken down by seasons? Any significant patterns?

    [Response: We unfortunately can’t get at sub-annual snowfall variations with ice cores. We can only get mean annual data. I wish it were otherwise, but this is generally the case, simply because of the way annual layers are detected in the ice. It may be possible at the very highest snow accumulation rates, but there are not many such sites. Most of the Antarctic tropospheric warming is indeed in the winter. I’m not deeply immersed in those data though — I refer you to the original paper noted in our past, here — eric]

  29. 29
    Gareth says:

    Re: #15/17

    Paleocene-Anthropocene No Ice Climate suggests itself.

    A hint of PANIC about our future?

  30. 30
    Bryan Sralla says:

    Re: #27. Thanks for the compliment Steve. RP1 seems like a good guy. I am glad you think that I am catching on. As far as the PDO, I have not a clue whether the shift in Alaska temperatures after 1976 was related to such. The Alaska folks (scientists and fisherman) seem convinced though(admittedly a little antecdotal).

    My reason for pointing out this paper was mearly to observe that the pattern of warming seemingly has taken place in short, big gobs, followed by fairly lengthy periods of more steady climate. Maybe we will eventually observe the same type of pattern on the Antarctic continent?

    To me, it intuitively seems possible this observed pattern may be a bi-product of complicated ocean heat storage, and a trigger-like transfer to the atmoshpere. With the current resolution and understanding of ocean circulation, such processes might be exceedingly difficult to model. RP1 has made similar points, so I give him the credit.

  31. 31
    Eric Swanson says:

    Re #27, etc:

    Recent research has shown that there is much greater variability in ocean currents than has been previously believed, thus the climate change community must consider that there is much more to the problem of climate change than just what happens in the atmosphere. The oceans are connected and it’s been claimed that the THC is part of a system of global circulation, why not consider another alternative for the shift known as the PDO?

    There is evidence that the THC in the Greenland Sea was shutdown for some period in the late 1970’s and early 80’s. Given that fact, is it not logical to connect the evident change in the THC with the PDO shift? Furthermore, the so-called Great Salinity Anomaly (GSA) circulated around the Sub-Polar Gyre of the North Atlantic. I suggest that it’s plausible that these oceanic events were connected, although I can offer no support for that notion. Other GSA type events have been noted since the first was found, perhaps indicating a natural cycle which would change as global warming progresses. Perhaps these GSA events are like a ball bouncing on a hard surface, a system which oscillates while it slowly loses energy as a new rest state is approached. As the yearly cycle of sea-ice in the Arctic may be expected to continue toward a minimum of zero extent in future, we must be very concerned about the resulting changes in ocean circulation.

    [Response: Before making such links between the North Atlantic and the Pacific, it is very worth while to read Clara Deser’s work on the PDO, which in my view greatly de-emphasizes (and correctly de-emphasizes) the significane of the 1976 “regime shift” in the Pacific. I’ll track down the reference and put it here when I get a chance. -eric ]

  32. 32
    Eric Swanson says:


    You wrote: “it is also clear from satellite data that surface temperatures decreased during the years 1982 through 2002”. I presume you refer to the UAH data. As William pointed out back in March, the MSU data of Spencer and Christy may be giving incorrect trends over the Antarctic. There is a strong surface influence on the MSU over the Antarctic, due to the high elevations there. The RSS team does not even report results poleward of 70S.

    As I reported in a paper in the GRL, the MSU annual cycle does not agree with that found in the sonde data (doi:10.1029/2003GL017938). Furthermore, the S & C analysis depends on an algorithm which is intended to remove the known stratospheric component in the data, but this algorithm may be optimized for temperate climate and thus fail to remove the well known stratospheric cooling trend reported over the Antarctic, where the tropopause appears at a lower level in the atmosphere.

    Another issue that S & C completely ignore is the fact that the MSU/AMSU instruments scan cross track, which means that the scans are mostly meridonal over the polar regions. The S & C algorithm combines data from these scans, the result being that there is no way to detect local trends at the nadir positions, even though they produce maps showing such results. S & C also report data poleward of latitude 82.5, even though their algorithm does not provide this data. They use interpolation to fill in these locations.

    There’s more, but, in short, I don’t trust the UAH results over the Antarctic.

    [Response: Eric, thanks for dropping by. I’m actually referring to the surface temperature data from the thermal infrared band, from Comiso (2000, in J. Climate). This is very reliable, as shown by the high correlation with the instrumental data, and I don’t think there is any question about the 1980s, 1990s trends. Much of the Antartic surface really was cooling (and may still be). Whether these trends are meaningful, relative the to the natural variabilty is another question — the trends are simply not big enough to demonstrate unequivocally that they must be due to external forcing (e.g. ozone depletion). (In addition to Schneider et al,. 2006 out next week in GRL, see our papers on this in J. Climate (Schneider et al., 2004), as well as the Annals of Glaciology paper which is some earlier work we did with the ice cores (Schneider et al., 2005). I’m much less familiar with the above-surface data that you refer to, but I really ought to familiarize myself with it. As Tas van Ommen pointed out in his first comment (above), the difference between the surface and upper troposphere temperature trends is interesting and deserves further examination. I’ll write at greater length on this in a future post, after having read more of your work, and others you refer to. Best wishes, — eric]

  33. 33

    Re #31 Where Eric Swanson wrote “I don’t trust the UAH results over the Antarctic.”

    I would be interested to know if you trust the sonde data in the Antarctic, and if so do you trust the sonde data in the tropics?

  34. 34
    pat neuman says:

    Eric indicated that expecting a greater sea level rise for having less precipation over Antarctica than the global climate models have indicated is an extrapolation which is probably not warranted for reasons explained in his RC entry above.

    Based on historically large variation of precipitation over space and time, I tend to agree – that’s probably not warranted.

    However, I think that if average temperature over Antarctica is warming less than indicated (by global climate models), then regions other than Antarctica may be experiencing greater warming than indiated (assuming over-all global average temperature is tracking with global model expectations).

    The summer (July-Aug) temperatures in the southern US in 2005 and 2006 have warmer than might be explained by temperature variation alone. Thus, it may be warranted to suggest that a warmer southern US in summer is the price being paid for psosibly having a slower rate of Antarctica warming.

    For example (55 million years ago),

    Excerpt (Aug 26, 2006 – Minneapolis Star Tribune):

    A blizzard of scientific papers “reflects the community’s excitement
    at discovering an extraordinary perturbation in biogeochemical systems
    that was unimaginable 10 years ago,” James Zachos, an earth scientist
    at the University of California, Santa Cruz, declared on his website.

    “The evidence for dramatic warming during the event is overwhelming,”
    Dickens said. “It is witnessed in all the oceans and continents.”

    For example, digging in Wyoming’s Bighorn Basin, Scott Wing, a
    paleontologist at the Smithsonian’s National Museum of Natural History
    in Washington, found fossilized leaves from ancient bean plants that
    he said had migrated 1,000 miles north from the latitude of Louisiana
    to escape the heat.

    Many species of mammals arose during the PETM and spread to new areas
    of the world, altering the course of evolution.

    But the unusual warmth also caused the loss of many deep-sea
    species. “It was the most severe extinction in the last 90 million
    years,” said Gabriel Bowen, another Purdue geologist.

    Climatic rerun, only faster this time
    Global warming today may signal the return of a monster heat wave that
    altered the course of evolution.
    Robert S. Boyd, McClatchy News Service

    Also see:
    Fossil Butte is helping us to understand climate change

    Some studies have indicated that the tropical regions were only slightly warmer (or about the same) than today when in previous global warm episodes, suggesting that future global warming in the tropics may not be large (certainly much less warming than polar regions).

    However, there seem to be signals being sent (less Antarctica warming, greater warming in the southern US and Upper Midwest) that, although global averages are increasing as indicated by models, the regional warming expectations may have been off (i.e. faster rate of warming in non polar regions than anticipated … compensated for in global averages by perhaps a slower rate of Antarctic warming).

  35. 35
    pat neuman says:

    Having seen Explorer Will Steger present his photos of before and after the collapse of the Larson B ice shelf and other material, it seems that if Antarctica is loosing ice mass it would be more likely to be doing so mainly from an increase in melt runoff following the collapse of ice shelves than from less precipitation over the shrinking continent.

    Bryan, What you said about my study on AK data (in #21) warrants more discussion, but at a later time. However, I think your comment that my interpretations of data have not been subject to editorial or peer review from the wider community warrants a reply from me at this time.

    Supervisors with my previous employer (NOAA NWS) refused to allow me to get peer review on studies of climate stations and climate-hydrologic change in the Upper Midwest. NWS supervisors claimed that climate change was too controversial and politicized to deal with, and well beyond the time frame of NWS hydrologic modeling and forecast responsibility. I was well aware in 2000 that hydrology was experiencing change along with the climate in the Upper Midwest. Me saying so at that time was very detrimental to my career as a hydrologist. I was removed from government service in 2005. I don’t think going through a peer review process would be worthwhile at this time, given my professional reputation with NOAA. I’ll spare Hank the details this time.

  36. 36
    kary says:

    This is a no brainer, we need to stop the coal consumption and push the solar and water fuel techonologies before it is to late and there is no stopping the climate change.

  37. 37
    Eric Swanson says:

    Re #33: Alastair,

    Having looked intensely at some of the BAS sonde data from the Antarctic, which I presented in my GRL report, I know there are many problems with that data. I think it’s well known that there are many problems associated with all sonde data sets, an issue which has been often discussed in the scientific literature. I hope you’ve seen the CCSP SAP 1-1 report, but there is also Seidel et al., “Uncertainty in Signals of Large-Scale Climate Variations in Radiosonde and Satellite Upper-Air Temperature Datasets”, J. Climate 17, 2225 (2004), if you are really interested.

  38. 38
    pat neuman says:

    Response in 32. says … Whether these trends are meaningful, relative the to the natural variability is another question — the trends are simply not big enough to demonstrate unequivocally that they must be due to external forcing (e.g. ozone depletion). …

    I think scientists need to be careful not to downplay the meaning of regional changes to just natural variability. Besides, there is no natural climate variability anymore due to anthropogenic global warming happening. Moreover, in evaluating change in a single region such as in parts of Antarctic, scientists should be looking at the latest globally averaged data and climate change in other regions as well. For that reason, in order to better understand what’s happening with temperatures in the Antarctic, scientists need to also consider what’s going on in the Arctic, the Upper Midwest, Europe and southern US summer abnormally warm temperatures in 2005 and 2006, and other regions of the world.

    Having a ‘big enough’ cooling trend in Antarctic would not demonstrate unequivocally that the cooling is due to external forcing (e.g. ozone depletion). Besides, based on the theory of polar amplification we should be seeing significant warming happening in Antarctic and not significant warming in low latitudes of the US. Scientists need to be monitoring regional temperatures. If trends are developing that don’t fit the big picture modeling (polar amplification) then scientists need to try to understand that and give their best shot at what’s likely to happen on a regional basis. Let us know ASAP if you think polar amplification should be tossed out the window for now.

    I sited a study as an example in comment #34. In Wyoming’s Bighorn Basin, Scott Wing … found fossilized leaves from ancient bean plants that he said had migrated 1,000 miles north from the latitude of Louisiana to escape the heat.

    If polar amplification near the time of the PETM (55 mya) was large, why would bean plants migrate 1,000 miles north because of the heat? Do scientists think that heat was less near the latitude where the foothills now exist in WY, CO and MT than the heat in Louisiana? It seems more likely to me that the the warmth near the PETM resulted in a large increase in sea level again (similar to 90 mya), was what caused the bean plant to migrate northward, not the heat. However, if it was the heat down south that caused the migration, then I question the significance of polar amplification.

  39. 39
    Jeff DeLaune says:

    Please excuse the naive question, but given the size of Antarctica, does most of the moisture in the atmosphere precipitate out near the edges of the continent, so that less moisture is available to build snow mass in the interior parts of Antarctica?

    [Response: Yes, that’s right. Take a look here for maps of Antarctic average snowfall. The most-easily understood figures is actually this one. Units are mm/year, I think (!). In any event, obviously much higher at the edges than the interior. -eric]

  40. 40
    Brian Allen says:

    I haven’t seen a reference to the recent op-ed’s out there on the the “tropical Arctic climate” of 55mya. The op ed states that “Study Suggests that Humans not a Major Cause of Global Warming”. See Does anyone know what this study is or why it would have any relevancy to global warming? Thanks, BA

    [Response: We do have a post on the actual results of the original studies in the Eocene in the works (somewhere). But as far as actual relevance to present day changes go – there are none. (But it’s still a fascinating piece of climate history). -gavin]

  41. 41
    Eric Swanson says:

    Re# 38
    The issue I was discussing was whether or not the measurements were accurate, not what these measurements show. I did not intend to imply that I was ignoring regional changes, only that I think the TLT product from the analysis by UAH of data derived from the MSU satellite instruments is faulty. I hope you are able to see my GRL paper, as I tried to show where I believe there may be a problem.

    The other side of the question you pose regarding the apparent lack of the projected polar amplification is directly related to the question of whether or not the data accurately represents what is happening over the Antarctic. If the TLT data for the Antarctic showing a cooling trend is wrong and there is actually a warming trend, then that becomes a rather large issue. The data from the ground and sondes do not provide a clear picture, as there are many areas without data. Most of the station data comes from the coastal areas, with only a few mid-continent locations providing long term records. I noted a small warming trend in the TLT during the colder months, as did the study that William Connolley worked on. Take a look back at WC’s post here:

  42. 42
    Brian Allen says:

    Gavin thanks to your reply to my post in #40. I did see reference to the Paleocene-Eocene Thermal Maximum that implied a doubling or tripling of CO2 and other greenhouse gases contributing to a 10-12C increase in global temperatures. This would seem to support concerns with AGW rather than suggest otherwise. BA

  43. 43
    pat neuman says:

    re 41. I understand that. Proceeding, in looking at the article by William Connolley (Mar 2006 at RC) as you suggested, I end up following a link back to the article by Cecilia Bitz (Jan 2006) which says the term polar amplification should therefore be reserved to describe the amplification of surface temperature changes. Thus, cooling or no change in surface temperatures in Antarctic in winter, discounting temperature variability and errors in measurement, imply that polar amplification is not taking place in the southern hemisphere. However, surface temperature measurements northern hemisphere (Alaska and Upper Midwest climate station data), suggest that polar amplification is occurring, but average Jul-Aug min temperature data at climate stations in the southern US in 2005 and 2006, and probably Europe in 2003 and 2006, suggest that more than just polar amplification is occurring.

  44. 44
    Eric Swanson says:

    Re #43

    Climate models as were available some 15-20 years ago showed strong polar amplification due to the snow/sea-ice/ocean albedo feedback. These models tended to be too cold in polar regions, thus the sea-ice extent was too large, extending rather further from the poles compared with the known historical extent. The models also used albedo values for sea-ice which were too large, while that for the ocean was too small, especially in the NH with the high solar zenith angles near the pole. Thus, in the older models, as a little warming melted some sea-ice, there was a large difference in absorbed solar energy by the ocean below. I think the model builders are doing a better job lately, with much better sea-ice models and higher resolutions, which allow more precise representation of the various physical processes.

    It should also be noted that the sea-ice in the Arctic and around the Antarctic are in very different configurations, with the Arctic having an ocean at highest latitudes surrounded by land, while the Antarctic has a continent at highest latitudes surrounded by a very large area of water. The sea-ice cycle in the Arctic leaves a large area of ice at the end of the melt season, while the cycle around the Antarctic results in almost complete melt, except for the ice shelves at the glacier terminations. Most of the low altitude land ice melts in the NH summer, while the Antarctic remains mostly ice covered year round. I think that it should not be a surprise that the two polar regions respond differently to AGW.

  45. 45
    pat neuman says:

    re: 44 I understand what you said, that’s helpful information.

    I’ve been wondering if the climate models account for increasing melt rates as the atmosphere becomes more humid. My paper on snowmelt and dewpoints (below) concludes that higher humidity increases melt rates due to latent heat of condensation. During my career with NWS North Central River Forecast Center, I observed faster thaw of ice and snow on days with higher humidity, especially evenings.

    Snowmelt & Dewpoints in Minnesota, Wisconsin, and North Dakota
    September 11, 2003

  46. 46
    Eachran says:

    Good morning all, just to cheer you up a report in today’s Guardian.

    It is almost Antarctica but not quite : you can look at it as connected.

    Cities in peril as Andean glaciers melt
    Ice sheets expected to last centuries could disappear in 25 years, threatening water supplies

    John Vidal, environment editor
    Tuesday August 29, 2006


    Andean glaciers are melting so fast that some are expected to disappear within 15-25 years, denying major cities water supplies and putting populations and food supplies at risk in Colombia, Peru, Chile, Venezuela, Ecuador, Argentina and Bolivia.
    The Chacaltaya glacier in Bolivia, the source of fresh water for the cities of La Paz and El Alto, is expected to completely melt within 15 years if present trends continue. Mount Huascarán, Peru’s most famous mountain, has lost 1,280 hectares (3,163 acres) of ice, around 40% of the area it covered only 30 years ago. The O’Higgins glacier in Chile has shrunk by nine miles in 100 years and Argentina’s Upsala glacier is losing 14 metres (46ft) a year.

    Although a few glaciers in southern Patagonia are increasing in size, almost all near the tropics are in rapid retreat. Some glaciers in Colombia are now less than 20% of the mass recorded in 1850 and Ecuador could lose half its most important glaciers within 20 years.

    The rate of glacier retreat has shocked scientists, says a report on the effects of global warming in Latin America by 20 UK-based environment and development groups who have drawn on national scientific assessments. Their study says climate change is accelerating the deglaciation phenomenon.

    “The speeding up of the … process is a catastrophic danger,” says Carmen Felipe, president of Peru’s water management institute. In the short term, the president says, it could cause overflows of reservoirs and trigger mudslides, and in the longer term cut water supplies.

    According to the Colombian institute of hydrology, back in 1983 the five major glaciers in El Cocuy national park were expected to last at least 300 years, but measurements taken last year suggest that they may all disappear within 25 years. Meanwhile, the ice sheet on the Ecuadorean volcano Cotopaxi and its glacier has shrunk by 30% since 1976.

    “The [drastic melt] forces people to farm at higher altitudes to grow their crops, adding to deforestation, which in turn undermines water sources and leads to soil erosion and putting the survival of Andean cultures at risk,” says the report by the Working Group on Climate Change and Development, which includes the International Institute for Environment and Development, Christian Aid, Cafod, WWF, Greenpeace and Progressio.

    Their report, Up in Smoke, says snow and rainfall patterns in South America and the Caribbean are becoming less predictable and more extreme. “East of the Andes, rainfall has been increasing since about 1970, accompanied by more destructive, sudden deluges. Meanwhile, the last two hurricane seasons in the Caribbean rim have caused $12bn (£6.3bn) damage to countries other than the US. Tropical storms are expected to become more destructive as climate change intensifies. Climate change models predict more rainfall in eastern South America and less in central and southern Chile with a likelihood of greater and opposite extremes. The 2005 drought in the Amazon basin was probably the worst since records began.”

    Rises in sea level are expected to be especially severe in the region over the next 50 years, with 60 of Latin America’s 77 largest cities located on the coast. The first hurricanes have recently hit south of the equator line in Brazil. “The net effect … is to reduce the capacity of natural ecosystems to act as buffers against extreme weather.”

    “What we are seeing are many more negative and cumulative impacts. The larger the rate of [climate] change, the more the adverse effects predominate. Climate change is set to turn an already rough ride into an impossible one,” says the report, which adds that the impact of climate change is “hugely” magnified by existing environmental abuse.

    It proposes that Latin American governments do not repeat the mistakes made by past and present North American and European governments. Several countries in the region are proposing a new generation of mega dams which would displace thousands more people and destroy vast areas of the Brazilian Amazon. The new importance of soya, both as a food and biofuel crop, could also devastate the environment, leading to a battle for land between companies.

    Large-scale coal, oil, and copper mining not only threaten fragile environments, says the report, but in some cases can physically endanger remaining glaciers and greatly increase climate changing emissions. “The Pascua Lama project on the borders of Chile and Argentina intends to move three glaciers that cover gold, silver and copper deposits. The glaciers sustain the mountain and valley ecosystems and there are fears that toxic wastes used in the mining will contaminate land and water,” says the report.

    Yesterday, the groups called on rich countries to urgently reduce greenhouse gas emissions and proposed that Latin America and the Caribbean governments be helped to reduce their vulnerability to extreme weather.

    “The only option we have, apart from demanding that developed countries take responsibility for the damages that climate change is causing, is to try to neutralise the adverse impacts that are [already] upon us. It is time to rethink the model of international aid,” said Juan Maldonado, former Colombian environment minister and president of the UN convention on biological diversity.


    “With each new flood, drought or hurricane in Latin America, precious gains in poverty reduction are lost. Extreme weather is set to cause massive loss of life in developing countries throughout the region. The international community must invest more in helping poor communities cope with the effect of climate change,” said Simon Trace, chief executive of Practical Action.

    The world’s many thousands of glaciers have been stable or in slow retreat for more than 100 years but since around 1980 they have mostly been retreating drastically. The fastest decline is in the Himalayas, the Arctic, the Alps, the Rockies and the tropics. Most glaciologists believe this natural phenomenon is being accelerated by global warming. The effects of glacier melt are expected to be severe. Hundreds of millions of people in Asia and Latin America are dependent on glacier water. A reduction in runoff will affect the ability to irrigate crops and will reduce summer stream flows to keep dams and reservoirs replenished. In Norway, the Alps, and the Pacific north-west, glacier runoff is important for hydropower. If all the ice on the polar icecaps were to melt, the oceans would rise an estimated 70 metres (230ft). But even a small melt will affect coastal life.

    Guardian Unlimited © Guardian Newspapers Limited 2006

  47. 47

    Re #44 Eric, you seem to be saying that the models were overestimating the rate of sea-ice melting in the Arctic but are now correctly estimating a slower melt. In fact the Arctic sea-ice melt is accelerating see and I know of no models which predict that. Do you have any references?

    You have given a very good description of the difference between Arctic and Antarctic conditions. Using that, and what Pat wrote about the rate of the melting glaciers depending on the dew point, then it is possible to explain polar amplification and why it does not apply in the Antarctic yet. Water vapour is the clue. By condensing on the ice, it gives up its latent heat and so melts two water molecules for every one molecule that condenses. The greenhouse effect of water vapour, and the insulating effect of more clouds also helps to increase the melt rate. Once the ice starts melting it produces water and water vapour so causing a positive feedback. So long as the surface of the Antarctic ice is too cold to melt because its altitude and latitude is too high, then it will survive, especially since the polar vortex is bringing down dry air from the stratosphere.

    Even with global temperatures higher than today’s, then it is unlikely that the Antarctic ice will melt completly since it started growing in the Miocene when CO2 and tempertures were higher than today. Because it is self sustaining, then temperatures and or CO2 levels will have to rise above those of the Miocene for melting to commence at the south pole. OTOH, melting has started at lower latitudes on the Penisular, and this melt could spread to the West Antarctic ice sheet which it seems has collapsed into the sea before :-(

    Anyway that’s my 2 cents worth.

  48. 48

    Re #47 I was of course replying to Eric Swanson, although I would be even more interested in any comments Eric Steig might care to make. Sorry for any confustion.

    [Response: Alistair: I will respond at some point; Eric Swanson has raised a bunch of interesting points but it’ll take me a while to digest the discussion he has started.
    Best — Eric Steig]

  49. 49

    #47, Alastair, My thinking likewise, its more the edges of Antarctica that need to be looked at. With interaction between warmer low altitude air and the standing ice walls or slopes. Since coastal stations have reported a warming, it is only logical that some degradation must be found near the sea.

  50. 50
    Gerd Wendler says:

    Comments on Steffen Cristensen remarks #15 by Gerd Wendler

    In our paper (Hartmann and Wendler ( the sudden warming which occurred in 1976 is being discussed and related to the shift in the PDO. We compared the 25 years prior to this event to the 25 years thereafter. We used all first order climatological stations in Alaska operating for the time period, hence, the best data available taken by professional meteorologists with calibrations carried out regularly. With the exception of the Arctic, which is based on a single station (Barrow), a fairly flat temperature trend is observed for all other regions. Nowhere we state that the mean, which we do not present in our Figure 5, is the mean for the land surface of Alaska.

    Christensen’s critique is arbitrary in that using only a 10 year base (1977-87) compared it to a 5 year mean temperature (2001-2006) at a different time period is quite dubious. Why do we have 30-year means in climatology? Choosing short term temperature intervals says little about trends as the annual variability is large.

    Finally, I agree with Christensen insofar that we had some warm periods since 2001. For example, the summer of 2004 was the warmest ever observed in Fairbanks, and 2005 was the 3rd warmest. On our website we have information on trends for the period 1949 to 2005 and 1977 to 2005 ( Note that for the period 1977 to 2005 there has been a slight increase in mean annual temperature (taken as the average of the first order weather stations), but some stations show a cooling.