There’s been a great deal of static in the blogosphere lately about the supposed “recovery” of Northern Hemisphere sea ice. The stories usually go on and on about how much more ice there is this year than last year! Those who aren’t in denial know that single-year changes in noisy variables don’t mean a whole lot, unless they’re truly amazing (like last year’s record-smashing decline in summer sea ice). But just how large, really, is the “recovery” of NH sea ice? The National Snow and Ice Data Center (NSIDC) has just updated the monthly averages of sea ice extent and area as measured by satellites, to include October of 2008. Let’s take a look.
Here’s the time series of monthly average sea ice extent:
Here’s the data for sea ice area:
Both show signs of decline lately, expecially in late summer/early fall when ice is at its minimum. But the large difference between summer and winter masks the size of the changes. So, I’ll remove the annual signal in two ways. First, I’ll compute anomaly (removing the seasonal part of the signal directly), and second, I’ll compute the “mean” value with a wavelet transform. Here’s the result for extent:
And here’s the result for area:
Both extent and area show an unambiguous decline. Linear regression indicates that the average rate of decline since satellite observations begain is 51,000 km^2/yr for extent and 49,000 km^2/yr for area. And yes, the decline is statistically significant. In fact the decline is demonstrably nonlinear, and the rate of decline is faster now than it was earlier.
There’s certainly not much sign of “recovery.” Yes, this year’s ice extent and area are higher than last year’s, but this year is well in line with the prevailing trend and doesn’t show any indication of reversal, or even slowing, of that trend. Only by comparison to last year’s astoundingly low values can this be called a “recovery” — it’s by no means a reversal of the long-term trend. In fact, this year’s data represent a continuation of the long-term trend.
We can also take a look at the annual values for the months of September and October; here’s the extent data:
The October extent shows a decline, at about 50,000 km^2/yr, and we see an even larger 78,000 km^2/yr decline for September. The September 2008 extent is the 2nd-lowest on record, exceeded only by last year’s unbelieveable low, and October 2008 is the 3rd-lowest on record. For September, the three lowest years are 2007, 2008, and 2005, while for October they’re 2007, 2006, and 2008.
Looking at September/October data for ice area, we get:
Again, both months show a strong decline. For ice area, 2008 recorded the 2nd-lowest values for both September and October, with the incredible 2007 of course holding the record.
For those who are interested, the amplitude of the annual cycle is also interesting (here’s the semi-amplitude, which is just half the amplitude):
The amplitudes have been increasing, indicating that long-term, the decline is faster in summer than in winter so the difference between winter and summer has been getting larger. Both area and extent also show a dramatic increase in amplitude for the last two years, due to the truly breathtaking summertime declines.
Is the increase this year over last year a “recovery”? Suppose we watched adult males passing through a doorway, and we noticed a statistically significant increase in their average height as time passed. We might even be shocked at the magnitude of the trend when Shaquille O’Neil passed through the doorway! But if the next man through the door was only 6 feet 6 inches tall — well in line with the long-term trend but shorter than Shaq — should we conclude that the trend of increasing height has reversed? If the fellow next to you pronounced an astounding “recovery” of shortness, proving that the “increasing height” trend was only an illusion, should you believe him?
No.
The trend in sea ice has continued. Then why are so many bloggers calling it an amazing recovery, even proof of the incorrectness of global warming? Because that’s what they want to believe. But to support that mistaken belief, they have to focus on comparing this year to last year, ignoring all that came before it.
They do that a lot.
39 responses so far ↓
michel // November 4, 2008 at 7:21 am |
There is clearly a downward trend over the last 30 years, and equally clearly the last two years have been exceptional in decline. The thing that is much more problematic, surely, is what it means. We don’t seem to know how exceptional the last 30 years are in terms of the history of the last few hundred. The skeptical sites mention anecdotal evidence of large scale ice retreats in the early 1800s and 1920s, but they never have any evidence of how extensive or prolonged they really were.
It also seems, as you say, very difficult to assess the significance of the two year observations. We can see that they are most unusual in the history of the last 30 years, but whether they are so unusual over a few centuries?
It is probably a matter on which more caution is indicated than is usually shown. Certainly, as you often say, little significance should be attached to one or two years observations. Though these particular ones do seem rather exceptional.
[Response: For a look at historical data preceding the satellite era, see this.]
SimonP // November 4, 2008 at 12:13 pm |
Thanks for an interesting blog. I’m not a scientist or mathematician, but I feel I owe it to myself to read as much as I can from both sides of the argument on global warming so that I can come to a balanced conclusion. Right now, I still don’t know whether AGW it’s a real phenomena or not. Recently, I have read much about the recovery of sea ice on ‘denier’ sites and blogs, so it’s interesting to see detail to the contrary. Maybe it’s a little unfair of me to ask you this question, but I’m going to anyway; in your opinion then, does the continuing trend in the reduction of sea-ice prove the ‘correctness’ of global warming?
[Response: By itself? Of course not. It's pretty powerful evidence that the arctic is getting warmer rapidly -- but we really new that already. It's also strong evidence that warming of the climate can bring about significant change on a massive scale. But strong arctic warming alone doesn't tell us much about the *cause* of the warming.
And the crowing of the denialosphere about sea ice "recovery" tells us quite a bit about...]
Georg Hoffmann // November 4, 2008 at 12:48 pm |
Tamino
of course you were right concerning the way how the “recovery” is sold and “how this year is the first step of an long attended cooling trend”. But, as I mentioned in the discussion on the UAH vs RSS data, I think in fact something new started last year with seaice. When the seaice minimum falls below a value of about 20% of the climatological longterm mean for September then we start to see a clear seasonal cycle in the anomaly series.
So I would make the specific prevision that 1) a seasonal cycle will show up in sea ice data and 2) due to 1) the october/november increase and the may/june decrease of seaice will become bigger and faster. I might check that with CMIP model data but I am pretty sure that this is what the model show and what we start to see in your anomaly plot or here in Bill Chapmans plot:http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/sea.ice.anomaly.timeseries.jpg
Why not computing a wavelet transform of the anomalies and see if a seasonal cycle slowly starts to appear in the last years?
Sekerob // November 4, 2008 at 2:54 pm |
If the Ratio of Area to Extent is an indicator of the State of the Cryosphere, than this is a plot showing the %. August 2008 was the worst ever found. Now we watch how November is going, and it’s not.
http://i137.photobucket.com/albums/q210/Sekerob/ArcticSIA-SIERatio2.png
October ‘08 is based on 31 days average of daily data published by IARC-JAXA and CT. Will replace soon as NOAA/NSIDC releases their monthly figure.
John C // November 4, 2008 at 2:57 pm |
One observation is that the decline in sea ice trend is steepest over the last 6 or 7 years. Which coincides with basically globally flat temperatures. Why is that ? And if the Arctic is warming anomalously (compared to the global average) doesn’t it mean that the rest of the world must have been cooling slightly (to keep the average flat) ? I’m not trying to confirm or deny anything, I just want to understand this arctic warming in the context of a global picture.
[Response: One thing to remember is that some global temperature indices (HadCRUT3, satellite temps) omit the arctic altogether. As far as I know, only GISS includes it (estimating by interpolation), and GISS shows the strongest warming globally recently.]
Sekerob // November 4, 2008 at 3:30 pm |
Consider that sites like these: http://arctic-roos.org/observations/satellite-data/sea-ice/ice-area-and-extent-in-arctic are the seeds upon which the ‘uninformed’ draw their notions from.
Richard C // November 4, 2008 at 5:41 pm |
If you examine volume, I believe this year was the lowest. Click link/name to a graph I uploaded. (It got a bit mangled during ftp). This summer the ice extent was lower than the mean for over two months.
Philippe Chantreau // November 4, 2008 at 6:52 pm |
The arguments used by “skeptics” push ever so far the limits of the baffling. Back in August, someone who got his info from Watts’ blog was arguing to me that the SUMMER sea ice was showing a recovery in 08 because it was higher than 07.
It proved fruitless to point out that it was still the 2nd lowest on record, far below the 3rd lowest and even farther below the 79-00 average. The same guy seemed to believe that Antarctic sea ice showed a significant increase over time, in spite of me pointing him to the sea ice graphs on NSIDC that clearly show no trend or a trend below margin of error. Some people just won’t be rational.
[Response: A linear regression to SH ice extent *does* give a statistically significant result -- but just barely. The rate of increase according to that analysis is small, about 12,000 km^2/yr, compared to the overwhelmingly significant and much larger decline in NH ice extent of about 51,000 km^2/yr.]
Gareth // November 4, 2008 at 7:45 pm |
Good post Tamino. I’ve updated my recent thoughts on the subject with a link to your analysis.
Phil Scadden // November 4, 2008 at 8:55 pm |
As always, curious about the maths. The wavelet transform you used for smoothing – is that a “calculate cooff ‘denoise’ for seasonal frequency and then reconstruct?” Looks like a useful approach.
[Response: I'm not exactly sure what you mean. This particular wavelet transform fits, for a given frequency (in this case 1 cycle/yr), a function of the form x = m + A cos (wt + z), where "m" is the mean value, "A" is the semi-amplitude, "w" is the frequency (in this case fixed at 1), and z is the phase. The parameters m, A, and z can then be taken as estimates of the mean value, semi-amplitude, and phase; in this case it's "m" that is used as a smoothed estimate, (the seasonal component goes into the other terms), and "A" is used as an estimate of semi-amplitude.]
Phil Scadden // November 5, 2008 at 12:47 am |
“This particular wavelet transform fits, for a given frequency (in this case 1 cycle/yr), a function of the form x = m + A cos (wt + z), where “m” is the mean value, “A” is the semi-amplitude, “w” is the frequency (in this case fixed at 1), and z is the phase.”
Okay, very different. Does this particular wavelet transform have a name? Looks like a useful approach for looking at data with diurnal cycle (eg river temperature).
[Response: It's called the "Weighted Wavelet Z-transform," or "WWZ." You can read about it here.]
Phil Scadden // November 5, 2008 at 2:05 am |
WWZ – thanks very much. Even found some software. I’ll look more closely at it when I get some time.
TheWonderer // November 5, 2008 at 3:38 am |
It seems reasonable to me that the amplitude of the annual cycle wrt area and extent will continue to increase until winter temperatures become warm enough to inhibit widespread freezing. If this is true, we can look forward to dramatic claims from denialists in 30 years regarding “unprecedented recovery” as the open Arctic ocean of the summer freezes over in winter.
Richard Steckis // November 5, 2008 at 4:09 am |
Is the satellite data for sea ice free of the sort of annual anomaly that you observed for UAH? Have you tested for it?
fergusbrown // November 5, 2008 at 1:24 pm |
Thanks for the moral support.
:)
Check out Walt Meier’s comments on WUWT?; clear enough; but subsequently completely ignored by W.
As has frequently been said, you can lead a horse to Walter, but you can’t make him think.
mauri pelto // November 5, 2008 at 7:49 pm |
I recently reviewed a paper that is in press. It noted the emergence of asurface based arctic atmospheric temperature amplification signal. The signal was most pronounced in the later summer and early fall after considerable open water was present. The additional heat in the system went to melting the sea ice earlier in the season. And the rapid spread of sea ice at the start of the winter ended the amplification. This would increase the amplitude of the annual change from the minimum extent to maximum extent. http://www.the-cryosphere-discuss.net/2/601/2008/tcd-2-601-2008.pdf
chriscolose // November 5, 2008 at 11:17 pm |
There are a number of other recent papers on attribution of polar changes as well…
http://www.agu.org/pubs/crossref/2008/2008GL035725.shtml
http://www.nature.com/ngeo/journal/v1/n11/abs/ngeo338.html
William Connolley // November 5, 2008 at 11:17 pm |
You’re right, arctic ice is on a long-term decline. No-one serious is arguing about that.
However, a lot of otherwise serious folks apparently took last years extreme anomaly as the start of a new much steeper trend. I believe that people are quite right to object to that interpretation, and this years ice is evidence against it.
Dano // November 6, 2008 at 2:58 am |
I just got an e-mail notice about polar ice:
Best,
D
Aaron Lewis // November 6, 2008 at 4:18 am |
Not all ice is equal. Seals need ice over the continental shelf that is thick enough for a den to pup in. Polar bears need seals and ice that is within 60 miles of land if the piece they are on should melt. We could still have ice over deep water and lose our seals and bears due to lack of suitable ice.
Richard Steckis // November 6, 2008 at 4:30 am |
Dano,
Just a few problems with that study.
1. Almost all the stations considered were on the Antarctic Peninsula. It is well know that the peninsula is warming out sync with the rest of Antarctica.
2. The Scott-Amundsen station at the South Pole shows a long term decline in temperature.
3. The east Antarctic stations show an insignificant slight warming.
4. None of the weather station data show significantly abnormal long term increases in temperature.
Generally the weather station data show about a +0.1C change per decade.
[Response: Did you really think I wasn't going to look at the data? The Amundsen-Scott station shows a trend of -0.006 +/- 0.014 deg.C/yr -- the uncertainty level is more than twice the estimated rate, we can't really tell whether it's up, down, or zero. Yet you're unabashed in referring to it as a "long term decline in temperature" while referring to east Antarctic warming as "insignificant slight"? The most generous appraisal is that you're confused.]
coby // November 6, 2008 at 4:52 am |
Thanks for that very comprehensive discussion, Tamino. I have a simpler take on this talking point here:
http://scienceblogs.com/illconsidered/2008/09/arctic-sea-ice-recovered-in-2008.php
Gareth // November 6, 2008 at 7:07 am |
Mauri, thanks for that pointer. Credit given!
Joan // November 6, 2008 at 12:56 pm |
I am not a scientist, and don’t claim to even begin to understand all of the ins and outs of global warming and the receding of the Northern ice. I do however understand people and I can recognize wishful thinking when I see it. People would often rather not look reality in the eye and will go to amazing lengths expending huge amounts of energy doing so.
Eli Rabett // November 6, 2008 at 12:59 pm |
I think William does not understand what I was saying. The 2007-8 summer ice minima are not the harbringers of a steeper linear trend but step down. The trend may be the same, but it will be displaced downward.
TrueSceptic // November 6, 2008 at 3:31 pm |
Eli,
How can we tell which it will be? Surely we need some more years’ data?
Dano // November 6, 2008 at 5:05 pm |
What Joan said (November 6, 2008 at 12:56 pm).
Best,
D
Eli Rabett // November 6, 2008 at 9:35 pm |
True, yes and no. The ice pack is characterized by a hysteresis effect. New ice is thinner and melts more easily than old ice, therefore if you wipe out a lot of old ice in one year 2007, it forces a step in the ice coverage measure.
Andy // November 10, 2008 at 4:05 pm |
It’s intuitive looking at the seasonal amplitude graph that at some point NH winter sea ice will stop recovering completely. This amplitude stretch can not go on foreever. Perhaps a future where winter ice is restricted to sheltered bays, etc. similar to what is observed in the southern great lakes?
Sekerob // November 11, 2008 at 10:43 am |
There’s an update report http://nsidc.org/arcticseaicenews/index.html
to include a comment that the total melt March-September exceeded 2007. Air warmth observed as sourced from the ocean and release from the ice freezing.
What continues to obscure me is how nothing is actually said about the real ice [AREA]. It was per their numbers an average 5.72 million km^2 square for October, the second lowest on their record. The 31 days of Cryosphere Today (daily reports apparently lagging 5 days)) averaged on my sheet about 5.633 mkm^2, close enough.
Also no word or update on Ice volume, which in several press discussion as for instance ESA was reduced substantially.
Hank Roberts // November 11, 2008 at 4:23 pm |
Sekerob writes:
> What continues to obscure me is how nothing
> is actually said about the real ice [AREA].
> It was per their numbers ….
Followed by numbers.
Could you make another attempt at that sentence? I can’t figure out what you’re saying.
Sekerob // November 11, 2008 at 5:48 pm |
Re: Hank Roberts // November 11, 2008 at 4:23 pm
My point is that “extent” is getting all the highlighting in that news release whilst the real ice [AREA & Volume] is hardly getting a word in. None in fact.
As confirmed in the OP charts, October was third lowest in Extent, but Area was second, and that’s significant, for that and it’s thickness is the stuff that determines how far or not the melt will go next year (ignoring any other influence for now)
As for the denialist blogosphere, if you see the last few days at WUWT and CA, one can only wonder. They’re not interested at all in the consequences of GW, Anthropogenic or not, only in the “look, I found 0.001C difference between 2 pages on the same site and therefore if this is wrong everything else must be too, and that fits my believes”.
jyyh // November 23, 2008 at 10:15 am |
Sorry I use an alias… in order to not use your time unwarrantly i go straight to the point… looking at the yearly ice area graphs provided by NOAA, there seems to be a reduction in the ice growth speed at something like 8.6 M km3, this looks (to me) like it happens yearly. Is there a simple physical reason for this or is it forced by the geography of the continents surrounding the Arctic ocean? There seems to be no correlation with time so I doubt the reason could be astronomical. If this is real, what would that mean for the surrounding areas in terms of snowfall or rain?
jyyh // November 23, 2008 at 10:43 am |
And that was supposed to be the NSIDC
thanks in advance
C Higley // December 8, 2008 at 8:59 pm |
Unfortunately, I must agree that 2 years of records do not a real trend make. But, then , neither die s 30 years. The NW Passage was one for a number of years just after 1900 (ships went through under sail) and several years at least in the 1940’s. I cringe at all of the mentions of so much warming in the last 50 ears, when they ignore that almost all of the 50 years before that was warmer than the last 50 years. 1998 barely reached the temperatures of 1953, just before we lost temperature until the 1970’s. [This temperature pattern is obvious, if you honestly stay away from urban stations, which the IPCC and followers refuse to do.] Sure we have had a 30 year warming trend, but we are still colder than the 1920-1940’s. So, what’s the problem?
As CO2 was actually above 440 ppm in the 1940’s and temperatures then plunged, I have trouble seeing CO2 as having a warming effect. And CO2 then declined 5 years later – a lag as one would if the oceans were involved. [ See Beck’s paper on 180 years of CO2 data. The IPCC CO2 curve is a patent lie as they purposely ignore ALL real direct CO2 data (100's of papers) and claim that indirect ice core data is better (known now to be 30-50% low in their measurements). That way they can totally ignore the 440+ ppm of the 1940's and pretend it did not happen.]
Thermodynamically, at best CO2 can warm the planet on the order of 0.01-0.10 deg C, not enough to do anything about. Even AL Gore admits that decreasing CO2 by Draconian measures will avid maybe 0.1 deg of warming. But, changing a constant of thermodynamics to suit themselves, the IPCC has inflated CO2’s ability by 10-fold and then factored water vapor as huge positive factor, which it is not. In fact, the water cycle makes it a huge negative factor as part of the global heat engine. CO2 cannot, will not, and has never driven our climate. The “greenhouse effect” from such a small constituent of the atmosphere is meaningless and is a never proven hypothesis by Arrhenius back in the mid-1800’s. Never proven because it is not a valid idea.
Whether we warm or not is not in our hands. The solar cycle drives that. The only thing in true lock step with our global temperature is the length f the cycles’ periodicity. Recent work has clarified this relationship very well and, also, shown the mechanism by which this can severely influence our temperature. The IPCC discounts the Sun entirely and stands by the opinion that the Sun’s influence is negligible. The solar cycles directly correlate with our major climate shifts. THe IPCC simply assumes that all natural climate factors are entirely swamped out by CO2. How convenient to have everything gone, so that they can blame EVERYTHING on us.
Whether we have food or not is in our control. CO2 is plant food. You read about the flowers blooming 2 weeks earlier than previously and the caribou missing the Spring growth in the calving grounds. ‘Not global warming necessarily, considering that the Arctic Rim stations all report cooling, but one fact that most do not know is that plants grow faster in higher CO2. CO2 is their limiting growth factor. There is no need to invoke global warming when the CO2 alone can have this effect. Is it a bad effect? Not necessarily, the caribou will adapt as they always have. It is ignorant and simplistic for us to assume such a small thing would endanger them. Just plain dumb, got-to-be-us, thinking. How arrogant. Greenhouses are standardly kept at 1000 ppm to accelerate plant growth – they love it.
Why do people defend a hypothesis that is not scientifically founded? Because, if it is not our fault we cannot fix it. AND, if it is not our fault, but we convince you that it is, in your ignorance they can tax you for living in ways no one ever did before. There’s big money out there to be had and you can do nothing about it. You have to live and they WILL get you. Sound paranoid. Of course, but it is not paranoia if it is true. Gore has made over $150 million in the last 5 years with his carbon trading company. He’s filthy rich and we pay for the carbon credits as companies pass the expenses of the credits onto us.
It is also a massive power grab as we can also strong arm third world countries into not developing their carbon resources and competing with us on the world market. This also plays perfectly for the environmentalist extremists, who believe that the way to save the world is to stop the industrialized world = bring it down. If this means a world-wide depression, they do not care. Instead, they should be looking to stimulate technology and development, so that we can move beyond the need for carbon fuels more quickly. You see, in the long run, we will need to conserve our carbon resources for making plastics and pharmaceuticals.
Global warming, as the steady, unchanging, and mild warming we have has since the Little Ice Age (yes, it did occur despite the IPCC trying to disappear it) is happening. We have our ups and downs along the way. The recent warming is still cooler than the early 1950’s and with the latest solar cycle 24 taking very long to start, we can look forward to cooling for quite some time, even something similar to the Dalton MINIMUM or worse.
I am not a global warming denier. I am a scientist who takes the narrow point of view that scientific observations, conclusions and decisions should be made on the facts and not opinions of science designed by politicians and “scientists” who think it is acceptable to cherry-pick their data, lie about the historical record by selectively discounting or accepting or altering the data, and putting disproportionate faith in computer models which are designed to provide the results that they desire.
The IPCC is a political arm of the UN. They have in no way proven Anthropogenic Global Warming (AGW). Their mission is to describe its effects with the unfounded assumption that it exists. And for that matter, if it was not happening, they have absolutely no reason to tell anybody as disproving global warming is entirely against their mission.
The IPCC is a business with a huge budget and massive grants attached. There are thousands of jobs related to proving or studying an effect that does not exist or cannot be described except by not looking at any given phenomenon too closely and then invoking global warming as the cause. Without exception I have found a claim of AGW to be either the result of human mismanagement or some other natural cause. That’s not denial, because it is backed up by science. That’s the facts, sir.
AGW is claptrap and should be relegated to the trash heap while we get on with the many other things we should be focusing on: disease, poverty, medical services, pollution (NOT CO2) of water, air, and land, recycling of all kinds, alternative energy sources, etc.
[Response: The only claptrap I see is you comment. If you believe that "CO2 was actually above 440 ppm in the 1940’s" the you're not a skeptic, you're a denialist: so deep in denial that you've lost all connection with rationality.
Go away and let the adults continue our conversation.]
Ian Forrester // December 9, 2008 at 12:26 am |
Wow what a load of absolute rubbish. I checked out this denier’s website :
http://realscienceworld.blogspot.com/
and it seems he is a “scientist” and is actually involved in teaching. What a disservice to his students.
That type of denier should have his teaching credentials removed since he does not have a clue about the science he is rubbishing.
dhogaza // December 9, 2008 at 7:08 am |
Dude, *try* to keep your story straight.
Bob North // December 9, 2008 at 4:32 pm |
C Higley – Puh-Leeze. As a true skeptic, I am shocked at some of the stuff you spout. Anyone with a decent sceince background should be able to see the fatal flaws in Becks interpretation of historic CO2 measurements.
Lori // February 3, 2009 at 8:06 pm |
I just want to ask you a question? Do food crops and forrest grow in warm tropical weather, or bitter cold? How so they grow? I thought carbon is needed to create growth.
Just curious :)