RealClimate logo

The Arctic Climate Impact Assessment

Filed under: — stefan @ 5 December 2004

In early November 2004 the results of the Arctic Climate Impact Assessment (ACIA) were published, a uniquely detailed regional study compiled by 300 scientists over 3 years. The study describes the ongoing climate change in the Arctic and its consequences: rising temperatures, loss of sea ice, unprecedented melting of the Greenland ice sheet, and many impacts on ecosystems, animals and people. The ACIA is the first comprehensively researched, fully referenced, and independently reviewed evaluation of arctic climate change and its impacts for the region and for the world.

Sadly, in recent years we have become accustomed to a ritual in which the publication of each new result on anthropogenic climate change is greeted by a flurry of activity from industry-funded lobby groups, think tanks and PR professionals, who try to discredit the science and confuse the public about global warming.

An example of this is the article “Polar Bear Scare on Thin Ice” by industry lobbyist Steven Milloy , featured on Fox News. Milloy claims in his polemic that the ACIA “debunks itself” based on one graph of the 1,200-page study. This graph shows the evolution of Arctic temperatures over the past century, and the fact (well-known to climatologists) that in the 1930s similarly warm temperatures where reached in the high Arctic as at present. Milloy concludes from this that both warmings are due to a natural cycle.

Scientifically, this argument holds no water: it is simply not possible to draw conclusions about the causes of climate variations by just looking at one time series. Only considering the time series of Arctic temperature, it is impossible to tell what the cause of the 1930s warming was, what the cause of the recent warming is, and whether both have the same cause or not. Milloy’s specious argument is a characteristic example for a method frequently employed by “climate skeptics”: from a host of scientific data, they cherry-pick one result out of context and present unwarranted conclusions, knowing that a lay audience will not easily recognise their fallacy.

In fact, the conclusion of the ACIA study that the recent warming is due to anthropogenic greenhouse gases is of course not based on one particular time series, but on a host of further scientific data. For example, looking at all the temperature data rather than just one time series reveals that the pattern of warming of the 1930s was very different from the recent warming. In the 1930s, warming was localised to the high latitudes, consistent with this warming being the result of a natural oscillation (the so-called “Atlantic Multidecadal Oscillation”). Very similar natural oscillations are also found in climate models. The recent warming, in contrast, encompasses most of the planet; this is consistent with it being the result of a global forcing. A very similar pattern of warming is found in climate models as a result of rising greenhouse gases. (For full details, see the publication of Johannessen et al., Tellus 2004). Many other lines of evidence demonstrate convincingly that anthropogenic forcing was very likely the dominant factor in the warming of recent decades.

Milloy further claims that the observed global warming of 0.6-0.8 C over the 20th Century is “well within the natural variation in average global temperature, which in the case of the Arctic, for example, is a range of about 3 degrees Centigrade”. This is another misconception frequently promoted in skeptics articles: it confuses global-scale with local changes (note Milloy’s delicate phrasing of this), and hence compares apples with pears.

Local climate variations are generally much larger than global ones. The reason is simple: it is easy to generate large localised temperature changes simply by changing the atmospheric circulation patterns (as happens for example in the North Atlantic Oscillation, NAO) – this will steer the winds along a different track, causing some regions to warm and others to cool. In a global or hemispheric average, in contrast, this kind of redistribution of heat cancels out. To get global-scale variations, you need to add heat overall, not just shift it around to a different place. Global-scale variations are therefore much smaller, and they reflect changes in global climate drivers, for example in greenhouse gas concentrations or in solar activity. For this reason, an anthropogenic warming trend can only be clearly identified in hemispheric or global averages or in pattern studies. It can neither be demonstrated nor debunked by looking at individual local time series. Even as the global average temperature is rising, some regions have been cooling in recent decades (e.g., the Labrador Sea region or the Antarctic).

19 Responses to “The Arctic Climate Impact Assessment”

  1. 1
    Steven D says:

    Thank you for establishing this site. It will no doubt serve as a valuable resource. I will bookmark it immediately.

  2. 2
    SwimJim says:

    Why didn’t the ACIA report have a reference section? This (missing) aspect of the report has been highlighted as a detriment in several skeptical responses to the report. There are some really interesting trends and phenomena described in the report, so it would have been a good resource to follow up with citations to the original research papers!

  3. 3
    Joe Jackson says:

    Excellent site. I’m a Civil Engineer in the water industry and so “climate dependent” to a great extent.
    BTW, I’ve heard the 1930’s artic warming linked to the dust bowl storms of the same era.

  4. 4
    Mark Bahner says:

    “…unprecedented melting of the Greenland ice sheet,…”

    More than the melting that allowed the Vikings to inhabit Greenland (and call it “Greenland”)?

    More than the melting during the last interglacial? My understanding is that Greenland’s ice cover during the last interglacial period was a lot less than at present. Is my understanding wrong?

    Greenland Ice Sheet during Eemian (previous interglacial)

    Quoting from the above abstract, “Our results suggest that the Greenland ice sheet was considerably smaller and steeper during the Eemian,…”

    [Response: The unprecedented melting discussed in the ACIA report refers to the period over which observations of melting exist from satellite, i.e., since 1979 – see the ACIA highlights report (p. 6) for details.

    We do not know much about melt rates from before that time. However, the question regarding the time of the Vikings and the last interglacial seems to refer to the size of the ice sheet at that time, rather than to melting rates.

    Of course even if the Greenland ice sheet vanished altogether this would not be unprecedented on geological time scales, since during much of Earth’s history the planet was entirely free of ice sheets and greenhouse gas levels were much higher than today (see Royer et al. 2004 for more).

    When Erik the Red started a colony in Greenland in the year 982, climate and ice sheet size were probably not very different from today. It is often forgotten that this region in south-west Greenland (near K’agssiarssuk) is lush and green also today and supports agriculture and sheep farming. The oxygen isotope record from the Dye3 ice core, the closest core to the Viking settlement, suggests the warmest temperatures there of the past several thousand years were reached in the 20th Century. This does not prove anything about global warming (it is just a single point, the story at Greenland’s summit is already different), but it is interesting with respect to the Viking settlement. The Dye3 record shows how conditions deteriorated after the Vikings arrived and that they abandoned their colony after temperatures hit an almost 1,000-year low.

    The ice sheet size during the last interglacial (the Eemian, ~120,000 years ago) has attracted much interest recently, but data are still uncertain and controversial. Several studies (e.g. the one quoted by the correspondent) suggest smaller ice sheets and correspondingly a sea level several meters higher than the present. This is of some concern (hence the recent interest in this topic) since estimates for the global mean temperature during the Eemian suggest that it was only 1-2 ºC warmer than the present. If such a small Eemian warming did indeed lead to such a large sea level rise, prospects for the future of ice sheets and sea level would be more pessimistic than those given in the last IPCC report. – stefan]

  5. 5
    James McCarthy says:

    Regarding comment #2, only the Overview document of the ACIA has been released. The full report of about 1200 pages, with comprehensive referencing, is in final phase of editing and will be printed early in 2005.

  6. 6
    Erik D. Hilsinger says:

    To supplement Comment 4, Cambridge Press will be publishing the complete report in short order and it can be pre-ordered on Amazon, Barnes and Noble, or at Cambridge Press’ web site. The summary was prepared for Congress, I believe, however my senators (one of whom, Ted Stevens, I believe helped fund it) have pooh-poohed it as has my congressman, Don Young. The summary has already contributed to current policy documents and like the NAS report on global climate change, will have significant effects on policy documents if not actual policy and dedisionmaking.

  7. 7

    I carefully read several times the Key findings #9 ACIA article on ozone thinning and UVB increase. I was horrified on that rehassle of UVB scares (melanomoas, eye cataract, etc..) for the Inuit: the existing snow cover increases UVB today by about 50% through reflection; if it disappears, human UVB-dose will go down, even if the ozone layer will thin out by 20%; so at least for the human aspect the danger is nil. The scares and warnings on increased UVB are well known since many years: everybody agrees that too much is bad, but I never found and ophtalmogist who had a patient suffering from normal level UVB induced eye cataract (I do not speak of the dangers of industrial over-exposure). I would have appreciated a much more scientific and less populist-alarmistic chapter on this!

  8. 8
    Mike Koefman says:

    A beautiful, necessary and long-overdue site. Well done! Having witnessed superb US climate science presentations at successive unfccc conferences, especially in Milan December 2003, and experiencing at the same time the silent embarrassment of these same scientists (from the “land of the free”!) when challenged on their readiness to disown the misuse of their science by well funded “skeptics”, I can only say “hooray”. I do hope that we at the Campaign for a Hydrogen Economy will be allowed to quote you without too much hassle when we reiterate the obligation which this generation has to switch from fossil fuels, in their entirety, to the only feasible alternative, renewable electricity and renewably generated hydrogen.

  9. 9
    Ferdinand Engelbeen says:

    The Greenland photos are at without the “).”

    In addition to my comment at #9:

    The warming between 1905-1945 was global, as good as the warming in the second period 1975-1998. Between 1945 and 1975, there was a global cooling trend. That can be seen on the GISS graphs at

    But the regional climate of Greenland (and of most of the Arctic) remained longer cooler, compared to the global trend in the second warming period. I have looked at all data of circumpolar stations. Some 30% had a peak in the 1930-1940s, followed by a cooling until 1975 and a warming thereafter, exceeding the warming of the 30-40s. That were mainly stations in Alaska and Eastern Siberia. The remaining polar stations had similar trends, but a longer cooler period until around 1990 but don’t surpass the 1930-1940s in recent temperatures.

    This is at odds with climate models, which predict higher increases in temperature at higher latitudes than near the equator, if GHGs are the main cause of the increase in temperature…

    [Response: On this and your previous comment, I can only again point to the analysis of the warming pattern published in Johannessen et al., Tellus 2004. Their Figure 1a clearly shows the difference between the 1930s and the recent warming and concludes recent warming is likely to be anthropogenic. If you disagree with the findings in this peer-reviewed paper I propose you contact its authors. If you think their analysis is flawed in any way, submit a comment to the journal, as is the custom in such cases. – stefan

  10. 10
    Ferdinand Engelbeen says:

    In addition to message #4, I have the impression that current Greenland climate is not very different of that in the 1930-1940s. See the temperature trends of all Greenland stations from the GISS database at:

    As far as can be seen in the long trends, the temperatures today are about as high as in the period 1930-1940. Thus local warming now is not worse than in that period. And the main increase in temperature was before 1930, when the human-made extra GHG level was much lower than today. Thus mainly natural.

    Btw, Greenland, even in the south, has little or no agriculture (see: They have some sheep and reindeer farming, but most income is from fishing (shrimps and halibut), some halve of the GDP is from subsidies by Denmark… When we were there, they were investigating the growth rate of different wood species to see if that could work.

  11. 11
    Ferdinand Engelbeen says:

    The difference between Johannessen and my comment is that I have only looked at land temperature trends, while Johannessen e.a. mainly looked at sea surface temperatures and ice cover. But nevertheless, their graphs are interesting.

    I suppose that you are a little confused with the height of the temperatures. The 1930-1940’s globally were colder than today, but the polar temperatures were near equal (land a little lower, sea a little higher today). But as the graph (1a) shows, even the lower latitudes were higher in temperature in that period, thus the temperature increase was globally. On land, this is even more pronounced, see the temperature trends of the contiguous USA at

    Moreover, the other graphs of modelled GHGs influence (1b) and GHGs+aerosols (1c) clearly show that the models don’t reproduce reality. for GHGs alone the latest decades are much too warm and the 1930-1940 periode too cold. For GHGs+aerosols, the low-to-mid latitudes in recent decades are too cold and the 1930-1940 period is even cooling further… If a model isn’t able to reproduce reality (i.e. not validated), it is inapropriate to make any conclusions from the results… And the researchers clearly underplayed the solar cycles in these matters.

    As (recently) retired process automation engineer, but interested outsider in climate research, I don’t think my remarks will pass any peer review (appropriate or not). I have some experience with modelling, be it for chemical processes. I know how difficult it is to tune a model, even if all physico/chemical parameters are (or should be) known. Therefore I am a little skeptic about climate models, where a lot of physics and feedbacks are not even known to any accuracy…

  12. 12

    I am in possesssion of several reports which show that UV radiation in the Northern and in the Southern atmosphere is decreasing (and even if it had moderately increased in the Arctic this would be much less the radiation for inhabitants in the Alps or at the Aequator.) I can share these data with those who want to receive them.
    There are also reports which claim that the icecap at the Antarctic and in Greenland are increasing. This would be logical as a consequence of more evaporation of the oceans and consequently more precipitations in these cold areas.

  13. 13
    Eli Rabett says:

    F. Engelbeen’s claim that he only looked at land temperature trends, while Johannessen e.a. mainly looked at sea surface temperatures and ice cover is a Tech Central Station like. Johanneseen, et al’s surface temperature data set was taken from:

    “… a unique century-long SAT dataset focused on the high latitudes of the Northern Hemisphere. The dataset is provided through the Arctic and Antarctic Research Institute (AARI), St. Petersburg, Russia ….(Alexseev et al., 1999). The input data are daily temperature from 1,486 meteorological stations in the Northern Hemisphere, including land- and drifting-stations from the Arctic. A gridded dataset (5° lat. x 10° long.) based on these data has been developed from several sources.”

    I rather doubt that the majority, or even a large fraction of those 1486 meteorological stations were drifting, so the claim that Johanneseen, et al mostly looked at sea surface temperatures is misleading. Their data set, does have the advantage that covers both land and sea.

    Second, Johanneseen et al, are presenting a complete reports and discuss both temperature and ice coverage, and try to integrate observations of both. The discussion of ice cover does not detract from their discussion of surface temperatures, but on the contrary compliments it

    Third, although as Engelbeen notes from Fig 1a of Johanneseen, global surface temperatures were higher at all latitudes in the 1930s than in, say 1910 or 1950 they are warmer today at all latitudes than in 1930, and much more so at most latitudes.

    Fourth, as to the skill of climate models I note that Johannessen, et al reference a single 2000 such study which is certainly not the last word, but they appear impressed by it

    “This simulation without anthropogenic forcing is able to produce an anomaly similar to the observed highlatitude warming in the 1920s-1930s. Therefore, we strongly support Delworth and Knutsonâ��s (2000) contention that this high-latitude warming event represents primarily natural variability within the climate system, rather than being caused primarily by external forcings, whether solar forcing alone (Thejll and Lassen, 2000) or a combination of increasing solar irradiance, increasing anthropogenic trace gases, and decreasing volcanic aerosols, as suggested from an analysis of 400 years of temperature proxy data from the Arctic (Overpeck et al., 1997).”

    Another example of the perfect being made the enemy of the useful.

  14. 14
    Eli Rabett says:

    I would be curious to know which data sets that P. Lutgen refers to. AFAIK, the best summary of the subject can be found in Chapter 5 of the 2002 WMO ozone report which indicates an overall trend upwards in UV.

    Other useful sources of data can be found at the World Ozone and UV Data Center

    Now there very well may be sites at which UV has decreased due to aerosols, but on balance UV dose at the surface has increased as the ozone layer has thinned.

  15. 15
    Ferdinand Engelbeen says:

    As already said, I have looked at all land based circumpolar stations, in my case from the GISS database. The result is that the land based part of the Arctic in the last decades is a little less warm than in the 1930-1940’s. That Tech Central Station has found the same trends (see: ) from the GHCN network and other sources, simply confirms my findings. The AARI data include drifting stations and ice information, although not the majority (my fault to see that as “main”), that means that the difference between only land based and total is in warmer sea surface temperatures. Which is highly probable, as some pole ward currents are stronger during a positive NAO. See: last paragraph.

    An important point is that the temperature difference between lower latitudes and the Arctic (at least for land based) is smaller now than in the 1930-1940’s. Which is at odds with climate models, which predict that GHG driven temperatures should increase faster at higher latitudes.
    This also means that the increase in Arctic temperatures probably is tropical in origin, as the tropical sea surface temperature increased 0.85 ºC/decade in the past decades, see: . This leads to more heat transport to the poles, as well as via air (increased Hadley cell circulation) as via sea currents.
    The authors of the latter report haven’t decided if the warming is natural or GHG driven, but point to a higher probability for natural:
    “The possibility that lapse rates were decreasing instead before 1980 (23) suggests that the observed intensification of the Hadley-Walker cell may be due to natural variability on decadal or longer time scales rather than to a forced climate change; however, the length of the satellite data record is too short to distinguish between these two driving mechanisms.”

    The results of the 1999 state-of-the-art climate model are not very impressive (to say the least). If a model only accurately describes some 10% of the time-latitude area, then such a model has not the slightest value to make any conclusion of it. That the model can reproduce some (internal) natural variation is more an artefact of the model (the temperature jump may occur anywhere, even in recent decades, see previous paragraph) than of reality. Current models anyway underestimate solar influences, as they only incorporate direct insolation and ignore secondary effects on changes in jet stream position, cloud cover,…

  16. 16

    I think the arctic sea ice colapse in summer will come faster than the IPCC models say. Even perhaps before 2020.
    And the climate change in Arctic will be very fast after that. Ocean warming (the albedo of the water is lower than the ice on and Gulf Stream warm courant will be cooled no more in summer, without floating ice.
    And, with ocean warming, what will hapen with the oceanic clathrats (methane hydrates)?

    Alain Coustou (from France). I’m the main author of “L’effet Venus” published by EONS (

  17. 17

    The geologist John Atcheson has published (december 15, 2004) an interesting article in the Baltimore Sun “Tickling Time Bomb”. That is a complete confirmation of my analysis in “L’effet Venus” (november 2004, ). He says: “The arctic council’s recent report on the effect of global warming (…) ignored a tickling time bomb buried in the Arctic (…). There are enormous quantities of naturally occuring GHG trapped in ice-like structures in the cold northern muds and at the bottom of the seas (…) called clathrates (…).”

  18. 18
    Ferdinand Engelbeen says:

    I have made an error in comment #15, the surface temperature in the tropics increased with 0.085 K/decade, not 0.85 K/decade in recent decades…

  19. 19
    Jarl Ahlbeck says:

    The Johannessen report that link recent arctic warming to GHG seems to be based on the assumtion that the positive phase of the NAO in the 90:s is caused by anthropogenic warming on the southern hemisphere. As NAO is to blame for warming, GHG have caused it.This is a weak point. The more common El Ninos have caused recent warming of the oceans, and the connection between El Nino and GHG is not proved. I think that the explanation of Polyakov: The recent arctic warming may be a part of a natural cycle has to be taken seriously.