Has Arctic sea ice returned to normal?
Posted on 24 April 2010 by John Cook
An argument gaining popularity in 2010 is that Arctic sea ice has recovered from its long-term decline. Winter sea-ice extent bounced back strongly in the cold Arctic winter of 2010. Anthony Watts tells us that Arctic sea ice is returning to normal. However, this argument only looks at one piece of the puzzle while neglecting the full picture of the Arctic sea ice. Consequently, I've added a new rebuttal to the argument 'Arctic sea ice has recovered' (the 108th skeptic argument). This is basically a simplified version of Peter Hogarth's comprehensive review of peer-reviewed research into Arctic sea ice (and also includes useful information from the comments).
When people talk about the state of Arctic sea ice, they most commonly refer to sea ice extent. This is the area of ocean where there is at least some sea ice (the most common cut-off is 15%). Sea ice extent shows a strong seasonal cycle as Arctic ice melts in the summer, reaching a minimum in September, then refreezes in the winter, peaking in March. Temperature is the main factor driving changes in sea ice extent, although other factors like wind patterns and cloudiness play a part. While sea ice extent has been steadily declining in recent decades, it fell to a record low in 2007 due to a combination of factors.
Figure 1: Sea ice extent with trend from 1953 onwards.
Sea ice extent gives us a reasonable indication of the amount of Arctic sea ice but does have its limitations. Extent tells us about the state of the sea ice at the ocean's surface, not what's happening below. A better metric for the total amount of sea ice is, well, the actual total amount of sea ice, measured by sea ice volume. Satellite radar altimetry (Giles 2008) and satellite laser altimetry (Kwok 2009) find that Arctic sea ice has been thinning, even in 2008 and 2009 when sea ice extent showed a slight recovery from the 2007 minimum. So while some claimed Arctic sea ice was recovering after 2007, the total volume of Arctic Sea ice through 2008 and 2009 were the lowest on record (Maslowski 2010, Tschudi 2010).
Figure 2: Continuously updated Arctic Sea Ice Volume Anomaly Polar Ice Center.
Currently, websites such as Watts Up With That are using sea ice extent in 2010 to claim Arctic sea ice has returned back to normal. A few days ago, Watts claimed that we had "more ice than any time on this date for the past 8 years". On the contrary, in March 2010, the total Arctic sea ice volume was 20,300 km3 - the lowest March value for total sea ice volume over the 1979-2009 period. Those who claim Arctic sea ice has returned to normal are focussing at the thin shell at the top and neglecting the steadily thinning sea ice below.
If ice volume is a better gauge of arctic sea ice and if the trend is more worrying why is this not regularly used as the standard for measuring arctic sea ice condition?
The way you word things it sort of suggests that WUWT are cherry picking the ice extent to tell a certain story. But most of the serious science website that provide day-to-day arctic sea ice coverage focus on ice extent as well (NSIDC, arctic roos, DMI, cryospere today etc).
Are there uncertainties with measuring ice volume?
I remember when he cheerfully tried to capitalize on the 2008 La Niña temperature drop, and now totally ignores the recent temperature rise.
Many people believe him, but that's just the human problem of sticking to anyone that says what one wants to listen. It does not have anything to do with the quality of data, or accuracy of analysis.
Because of the large short time variations, I don't think that one shoud be too impressed by either the sea ice minimum of 2007 or by the alleged recovery. These could be random events, and do not touch the long term trend.
The available data on ice thickness are strongly suggestive, but they only span about 5 years. Because we already know that we have to deal with strong short time variations, it is not clear to me that they do represent a long term trend. Zhang's thickness graph above goes back to 1980, but it seems to be based on models, so we only believe it if we believe in those models.
Still, the data on ice extent very clearly points to a trend towards less ice. The most obvious explanation would be global warming. But is it certain that this decrease is linked to that - yes, it does sound like a very foolish thing to say, but after all, we know that the situation in the Antarctic sea is very complicated. The sea around Antarctica is warming, but the extent of ice in that warming water is increasing. There are several interesting attempts to explain this, including a subtle model by Jinlun Zhang et al., but I don't think that we really understand what is going on there. If we don't understand the dynamics of ice around Antarctica, how can we be sure that we understand the dynamics of ice around the North Pole?
However, if you read the linked information you will see that Zhang's model has been confirmed by recently released US Naval measurements from it's under polar ice submarine patrols.
Scientific papers often use the expression "this is consistent with..." instead of categoric assertions like "this is proof that...".
There are lots of evidence pointing to AGW. Arctic ice loss is just one of them. Is it possible that other factors concur (as you rightly pointed out)? Sure. Do we have any evidence of that? Hardly.
I think of it like a crime investigation. The accused was seen entering the victim's house before the estimated time of the murder, and a car like his was seen leaving the neighborhood an hour later. The crime weapon was his property, and his hand had the chemical marks of firing the gun. He had the gun when the cops searched him afterwards. The victim had an affair with the accused's wife (there was a motive).
Does it prove beyond refutation the accused is guilty? Not quite. We can imagine some unlikely story in which he practiced shooting in his backyard and went to a friendly visit to the victim, and a third party stole his gun for an hour just to make him look guilty. But there's no evidence of this made-up story.
On the other hand, do we have enough information to take concrete measures about it (eg convict the accused)? Most likely.
Your question is a good one, and maybe some other commenter here knows a cause attribution study about this ice loss. I just feel it's important to keep the issues in perspective.
Alexandre> My question was not about AGW or not, but about the connection between the warming of the Arctic and the loss of ice. I just wanted to make the point that this connection might be more complex than what one would immediately assume.
Understood. Answering this good question of yours would be the role of a cause attribution study. I think I already saw one about this issue somewhere.
So there is nothing wrong with reporting extent or discussing it. You just have to keep in mind what extent is, and what it is not. It is sea ice cover, and not sea ice volume. If you have ever watched the ice break up on a frozen lake, you will know that cover and volume can have very little to do with each other.
If Watts is trumpeting a claim that sea ice has returned to normal because extent has returned to normal, while at the same time volume data show the lowest volume yet recorded, then cherry-picking is about the kindest assessment you can make of that.
This is an example of the interannual varibility that says nothing about the trend. And this is why it makes no sense to say that from 2007 arctic sea ice has recovered.
Unless one means that the this year number is higher, which no one doubts, a recovery or reversal of the trend cannot be assessed nor we have (at the moment, at the very least) reasons to believe it's going to happen.
I think it argues against looking at any year or short span of years.
While volume does seem the superior measure, the 5 years of consistent data available are insufficient to make a solid claim (although even at 5 years the data is a notable counter to "ice extent is recovering").
If the volcano in Iceland goes on for a long time it might affect sea ice this summer. We will have to wait until September to see what happens this summer. WUWT will always spout off when there is a small shift toward recovery, even if it is just noise in the data. They will quiet down when the cycle goes back to the long term trend.
http://www.jamstec.go.jp/frsgc/research/d2/masayo.ogi/2009GL042356-pip.pdf
Abstract
Based on a statistical analysis incorporating 925-hPa wind fields from the
NCEP/NCAR Reanalyses, it is shown that the combined effect of winter and summer wind forcing accounts for 50% of the variance of the change in September Arctic sea ice extent from one year to the next ( Δ SIE) and it also explains roughly 1/3 of the downward linear trend of SIE over the past 31 years. In both seasons meridional wind
anomalies to the north and east of Greenland are correlated with September SIE,
presumably because they modulate the export of ice through Fram Strait. Anticyclonic
wind anomalies over the Beaufort Sea during summer favor low September SIE and have contributed to the record-low values in recent summers, perhaps by enhancing the
flux of ice toward Fram Strait in the trans-polar drift.
Normal sea ice extent is probably better taken as the extent in 1970 or earlier where the records show a long period of steady not declining cover. After that the extent declines so any average of recent data is below the starting point and not useful as a definition of normality.
Sometimes average can be used to mean normality and arguments are made where the two terms are interchangeable. This is not the case here. The skeptics deliberately make a false argument by exploiting the fact that these different concepts may be treated as equivalent by the unwary reader.
Honestly the way you talk you would think Watts invented sea ice extent. All the international and national bodies that concern themselves with Arctic sea ice generally use ice extent as their primary measure. For example NSIDCto make their monthly assessement and daily measurements in terms of ice extent, and I'm not cherry picking NSIDC to make a point either.
It's just a ridiculous claim to say Watts cherry picks ice extent. I checked his blog he's been reporting this metric back to 2007 when it was showing the all time low.
In fact you could actually accuse SkepticalScience of cherry picking by changing from ice extent, which no longer shows what it would like to report, to ice volume which is more convinient. I actually don't believe this I understand John's wish to show the full story but I equally think you can't really blame Watts for using the 'standard' metric.
I still wonder why all the international and national bodies don't use ice volume as their standard metric?
The issue is the conclusions you draw from the sea ice extent data. It's inappropriate and misleading to say sea ice has returned to normal based on extent data because this is decidedly not the case - the total amount of sea ice is at record low levels.
So the major point here is that citing sea ice extent data needs to be interpreted in proper context, being aware that the thicker ice below the surface is thinning and that total sea ice is at record low levels at the moment.
I'm not sure that you understand what cherry picking is.
It's to select carefully just the fraction of data that supports what you want to be true. Like picking Germany and Austria in Europe and say "my sampling shows that virtually all Europe speaks German".
Watts does not "cherry picks sea ice extent" (as a parameter). He just picks a moment of near-average sea ice extent and shouts "it's back to normal!", as if the problem, if once existed, would be gone now. He does not analyse the whole dataset - he cherry-picks the convenient moments. If he did analyse the whole thing, he would find this near-average blip is as insignificant as the Jan 2008 cool spell as far as the trend is concerned.
Marcus @ 1 points out that new ice is more vulnerable to melting - well, of course. But we don't know how much will melt and how much will eventually form part of old ice until we see the impact of the coming summers and the following winters.
Steve L @ 6 notes the possible relevance of albedo with increased ice extent. Steve seems to believe this will ultimately be cold comfort (dreadful pun, I know) to the sceptical camp. John D @ 15 speaks of the contribution of wind forcings.
The reality is that we're dealing with highly complex systems. Marcel Bökstedt @ 5 highlights this well. Extrapolations from one winter's data mean little irrespective of where you stand in the warmist - sceptical spectrum.
I'm sorry, HR, but this is a really silly statement. If you want to talk about the AMOUNT of sea ice, you really should be talking about the VOLUME of sea ice (or the MASS, if you like). Extent is useful, but it is not volume. The point is that volume has always been more important than extent, but the continuous time series of extent estimates goes back a lot longer. So there is value in measurements of both extent and volume, but they need to be interpreted based on what they are.
In any case, during the cold spell in North America, the area that was much warmer than average was more like the Hudson's Bay area -- well south of the Arctic Ocean and not really "Arctic". So I'm not sure why it seems "perhaps counterintuitive" that when a cold air mass moves south, it is replaced by other air that in this case was warmer (as opposed to leaving a vacuum in the north?), and in any case the "warm" temperatures in that area were still below freezing, and not in the Arctic Ocean. I don't remember the situation north of Europe -- that may have been more relevant for Arctic sea ice formation, but I just don't recall.
Compare this with the earliest winter on record, 1980. You can see much more pale purple regions, suggesting areas with as little as 80% coverage.
2007 didn't look great but then neither did 1986, most years seem to look like 1980. The images are generated using a tool on the Cryosphere Today website where you can check as much of the intervening years as you wish. 2010 looks so good you can't help thinking it's a data error.
Since we're after the full picture I thought it was worth throwing this into the mix.
I agree. Therefore the quality of the ice volume data has to be highlighted because this is the drawback of that particular metric. I don't see any of this in the article and given we're after the full story then maybe it should.
For example from what I can tell Figure 2 is MODELLED data based on ice concentration.
and much more ...
warmer than average weather in the Arctic is still plenty cold enough to result in the formation of more sea ice, because the surface air temp is well below 0C.
Fair point. However, most of the Arctic was warmer last winter - not just Hudson Bay. As I recall, the argument on this site relating to Northern Hemisphere snowstorms suggested they arose in the context of increased precipitation due to warmer weather. Indeed, we have just had the warmest March since 1979 as per the satellite record.
Even if much of the Arctic has temperatures below 0 degrees, I would expect the overall area below 0 degrees to be smaller and for 0 degree conditions to persist for shorter periods hence leading to less by way of new ice formation. This is what I consider counterintuitive - hence, my wondering (naively perhaps) just how direct is the dependence of sea ice on temperature.
Given all this, Humanity Rules @ 25 sums it up well saying, 'There is something a little bit extraordinary about this years Arctic ice.'
First, there are the submarines. They did measure ice thickness from below, using sonar. Kwok and Rothrock report that there were two periods of submarine voyages under the ice. The first one was 58-76, the second 93-97. The reported difference in thickness is 1-2 meters, which is a lot. The problem with those data is that there were not that many voyages, it seems to have been about 10 in all in either period.
Then there are early direct observations 1955-2002, and buoys. Rigor and Wallace report on those. They show a decline in ice thickness. R&W; state that this happend very fast in 1989-1990, and blame it on the Arctic Oscillation which was in an extreme "positive" state at this point. A positive AO index means low pressure at the pole. This would be compatible with the submarine observations. They predict (in 2004) that the trend will soon reverse, and the ice thickness will build up again. This prediction has not come true, but on the other hand, the AO index has stayed on the positive side.
There are additional data from satellites and observations from locals, that all point to a decrease in ice thickness since about 1955.
The conclusion seems to be that we have had a decrease in both extent and thickness, giving an even bigger decrease in ice volume.
Humanity Rules> Yes, I noticed that darkening of the purple too, but I don't know what it means. Maybe we should mail the guys at cryosphere today and ask if it is real or an artifact?
Marcel Böksted #29 - It is very reasonable to assume that the AO index has some influence in arctic sea ice. On the other hand, there is also some debate whether this oscillation is in some way influenced by the anthropogenic sea temperature rise.
The math (ie the models) predicted Arctic Sea Ice would melt under AGW. It is actually melting faster than that.
I still want to find that paper about sea ice loss cause attribution. I was hoping some other commenter could help us here...
I've just done a quick research myself.
This Francis 2005 looks like something I meant before. They basically say direct longwave radiation (ie GHE) accounts for the largest part of sea ice loss (40%). Solar variation in the region is negative. Positive feedback due to albedo variation is also important, and there's some discussion (a bit outside the scope of the paper) about other factors, such as increased precipitable water.
About winds: "Anomalies in meridional winds also explain significant variance in most areas except for the Barents and Beaufort Seas, but the influence is much weaker than expected."
(I did not read it all, just a quick look at the abstract and conclusions)
Alexandre> I find the paper by Francis et al. confusing. It is mostly about statistical correlations, and it is not so easy to figure out what is cause and what is effect. As you say, they don't believe that winds are important - the polynyas don't matter. But they must be there, after all that word is Russian.
michael sweet> That's good to know. But it still does not explain why the maps from this year look different from the maps from previous years.
What really impressed upon me was that most of the ice in question was in the Bering Sea, not noted as a site for Perennial Ice! This really does illustrate the confusion that persists in many people's opinions between long-term and weather-related trends. It has been a spectacularly aberrant Winter this past year, according to many regional expectations for normal weather. It is important to argue and defend a more circumspect perspective e.g. in this case that cold air masses get pushed around by the warmer, especially in a Season characterized by El Nino and a strong NAO, and that persistent cold temperatures WILL yield more Seasonal Ice. In contrast to the dearth of Ice Pack in Eastern Arctic/Canada of course.
Anyway I'm new posting here, mostly here to be better informed given that I am not a Climate Scientist, so my contributions to the discussion are pretty much rhetorical. I do want to help combat the ignorant, pasted blogtrotting that goes on in commentary in Mass Media however. So I'll be checking in more or less irregularly!
This is underlined by a recent report from the Pew Environment Group concerning the “climate services” provided by arctic snow and ice and the potential economic costs of their disappearance.
From a summary of the study: “The report calculates that this year alone, Arctic melting [of sea-ice, snow-cover, and permafrost] may warm the Earth an amount equivalent to pumping three billion metric tons of CO2 into the atmosphere. 'That’s equal to forty percent of all U.S. industrial emissions this year or bringing on line more than 500 large coal-burning power plants,' said Dr. Eugenie Euskirchen, co-author of the report and a scientist from the University of Alaska at Fairbanks’ Institute of Arctic Biology.”
The average winter temperature in the Arctic Ocean is not just below 0C, it is far below 0C. In Barrow, Alaska (Arctic coast), the winter average temp is around -25C (Dec through March), that's the daily mean not the low. I don't know how much this varies across the entire Arctic Ocean basin, but the bottom line is that air temps are far below freezing and this will favor the production of sea ice everywhere in the Arctic Ocean all winter.
I would bet that the answer is known to your question about how direct is the dependence on temperature -- but I don't know it. I don't think the question is naive. Simple physics says that everywhere there is open water and air at -20C or -25C will mean there is significant heat loss from the water and this will favor sea ice formation, but the water doesn't just sit there in place so wind and waves causing water to mix will also have an impact, and the formation of some ice will also change the rate of heat flow out of the water+ice to the air.
And the upshot of what Alexandre explains in #19 is that amateur skeptics, Republican politicians and right wing media take what people like Watts say, and turn it into another global warming skeptic/denier meme, which then becomes more proof to them that AGW is hogwash. It then is a new "proof" in their littany of arguments, which get repeated endlessly in the skeptic echo chamber.
An example of this are the claims in Febuary that global warming is over because a big part of the U.S. got a lot of snow. Actually one of the snowstorms that hit Washington DC happened on the warmest Febuary 6th on record, globally that is. But this nonsense was repeated in every right wing media outlet, the blogosphere and by Republican polticians. Donald Trump even said Al Gore's Nobel Prize should be taken away because this was proof against AGW. It was a record warm winter.
Another example is the fallacious claim that IPCC scientist Phil Jones said there has been no warming since 1995. He never said that. But it is now another proof for skeptics,- another phony argument in their littany.
And Watts feeds and encourages this nonsense.
And so it goes.
how I can get the month-by-month Arctic Sea Ice Volume Anomaly and Total Volume data, so that I can plot it in EXCEL?
If climate alarmists would denounce Al Gore's "proof" being melting ice caps (poor lonely polar bears) and freezing cold winters, then skeptics wouldn't use his own examples against the climate alarmists when those examples reverse.
The Arctic ice *is* melting though, and it has been declining for decades. Polar bear habitat *is* threatened. The situation has not reversed at all. It gets tedious as well that every denier thinks that attacking Al Gore is the same as attacking climate science. He's irrelevant to the science.
This may be off topic slightly, but perhaps if we are to find a basis for establishing "normal" conditions on our planet, and that includes Arctic ice, we need to establish a bench mark of some relevant indicator that all else can be measured against.
As humans we should know what a normal diet is, too little food and we waste away, too much and we become obese, a "normal" diet should be one that allows our bodies to grow and perform close to their full potential.
In the natural world plants also have their diet that allows them to grow and perform close to their full potential, and part of that diet is that basic building block of all life, carbon. It is known that many plants need nearly 3 times the present concentration of CO2 in order to reach their optimum growth, and this knowledge is widely used by enriching the immediate atmosphere of the plants with CO2.
Therefore should the establishment of a "normal" bench mark for all things related to climate be based on what conditions would be, including Arctic ice, with an atmosphere containing 3 times the present levels of CO2?
It seems rather absurd to look for "normal" in time frames so ridiculously short as what are being used.
There's lots of data in the nsidc.org website, like this for instance.
I requires some work to dig out the information you need, though.
Ditto only more so for "Ice Volume".
'Simple physics says that everywhere there is open water and air at -20C or -25C will mean there is significant heat loss from the water and this will favor sea ice formation...'
I'm relieved that my question wasn't totally naive :).
The other variables such as wind and waves are obviously important. However, heat lost from water has to go somewhere - presumably into the troposphere and detectable as a rise in winter temperature. Thus, so long as winter temperatures within the Arctic do not transgress critical boundaries, ice will form regardless even if the overall temperature in the Arctic is recorded as increased.
Hence, consideration of summer temperatures may be more relevant. The complicating factor as far as I can tell is precipitation, which is predominantly snow. This is very difficult to measure (see the Wikipedia article on ‘Climate of the Arctic’). Increased temperatures, as we know, lead to increased winter precipitation (snowfall), which would predispose to better preservation of winter ice in summer months due to increased albedo. Does anyone know whether there was increased Arctic precipitation over the last winter?
It may be that arguments over ice cover in the Arctic may be a red herring in that increases or decreases may tell us little about world climate at least in the short term. From my very limited understanding, one could plausibly argue for warming even in the presence of increases in Arctic ice cover. The converse - global cooling despite decreased ice cover - may therefore also apply though clearly we have a significant raft of evidence from other sources that the world is not cooling.
I suspect that the iconic status of the Arctic and Antarctic ice caps may distract us from less glamorous but more robust climate signals. However, the world at large gets excited by headlines - hence, the need to address the complexities.
Both ice caps make it difficult to determine when things are in balance simply because of basic logistics.
Energy takes the long road in and the short road out. Anything that forces a change on one side of the ledger may not be balanced by an equal and opposite force on the other side, at least within the region.
It may well not tell us much about the global climate because the balance desired is to be found elsewhere in a different form.
Personally, I'm not interested in any climate "icon" if considered outside the global changes, we always need to put things in the proper context. So I'm happy to leave these icons to the media. But I have to admit that climate changes are more evident in some variable or process than in others; the Arctic sea ice shrinking is one of these.
There are certainly uncertainties around the model values. Ice extent has been the de facto standard for decades simply because it was the only data available. For ice volume we've only got five years of reliable data (and even that has some uncertainties) which cuts off since ICESat broke down. The recently launched CryoSat II and forthcoming ICESat II should give us much more precise data for a longer period going forward.
All that said, the claim that Arctic sea ice is recovering or has recovered based on just two years of extent data is plainly absurd. Both because it ignores basic statistics and because extent is only one of three variables needed to compute volume (the others being concentration and thickness)... which makes it a poor proxy for determining total ice amount.