Open Mind

Sea Ice, North and South, Then and Now

October 8, 2007 · 44 Comments

Having previously posted about northern hemisphere sea ice, in this post I’ll satisfy both a reader request, and my own curiosity, exploring the behavior of polar sea ice not only further back in time than the era of satellite data, but for both poles as well.

It’s undeniable that this summer we witnessed a truly dramatic decline in northern hemisphere sea ice. Yet most graphs and discussions have shown data only from satellites, which data don’t begin until late 1978. Some net searching turned up sea ice data for both hemispheres, covering the time period 1870 to the present: the HadISST data set from the Hadley Centre for Climate Change Research.

The data are monthly average ice concentration for each 1-deg. by 1-deg. latitude/longitude grid. Concentration is the fraction of the grid box which is ice-covered. Satellites provide the same data, but in addition NASA has computed “combined” data, namely estimates of total sea ice area, and sea ice extent. Ice area is simply the area covered by ice, so if a given grid box is 50% ice and 50% not, we count half its area toward the total. Ice extent is the area which has, on a large scale, more than 10% ice cover, so if a given grid box is 10% (or more) ice and 90% (or less) not, we add the box’s entire area toward the ice extent. Also, the satellite data use smaller grid boxes, giving greater resolution.

HadISST data come in separate (and rather large!) files containing blocks of 30 years, so I downloaded the data up to the 1961-1990 block; satellite data cover the remaining time. This gives a little more than 12 years overlap between the two data sets. I used the raw HadISST data to compute sea ice extent. Computing area is trickier, because before the advent of satellites, anything north of the “marginal ice zone” was assumed to be fully covered (100% ice). Hence these historical data provide a better estimate of ice extent than of ice area, and I shall limit my discussion to that parameter.

The HadISST data are far from ideal for studying the time variations of sea ice. They’re not designed for that purpose. Rather they’re inteded as input data to climate models, for which it’s necessary to have an uninterrupted “best guess” of sea ice concentration for every grid box over the entire globe. Hence the good folks at the Hadley Centre have filled in all the gaps using the best available information. For some rather long periods of time, the paucity of data makes it impossible to quantify the year-to-year variations; during those episodes, they’ve used the best information for the average behavior during that epoch. As a result, there are long time spans for which the data are the same year after year (”NH” is northern hemisphere, “SH” is southern hemisphere):



These graphs are rather crowded, making it hard to see the “fast” fluctuations, although it’s easy to see the upper and lower annual limits. For the northern hemisphere, the data before 1901 are a repetition of the same annual cycle, and the same is true from about 1940 to 1953. Nonetheless, since 1901 we have data suitable for studying changes over time. For the southern hemisphere the situation is much worse; until 1963 the data are representative of long-term averages rather than momentary values. Hence we may use these data to characterize sea ice extent changes for the northern hemisphere from 1901 onward, and for the southern hemisphere from 1963 onward. For anything after late 1978, I’ll use satellite data.

But first we should note that although the long-term southern data don’t give us annual resolution for very long, they do give us information about the overall long-term behavior. For the southern hemisphere we can reliably say that ice extent was, on average, higher from 1870 to 1940 than from 1950 to 1960, and that since 1963 it’s been lower still. Therefore over the long haul (since the late 19th century) there’s been an overall decline in southern hemisphere sea ice extent. The same cannot be said for the northern hemisphere; there are lots of fluctuations but the sizeable decline in ice extent doesn’t begin until about 1970.

We can also conclude that the recent record maximum in the satellite record of southern sea ice is just that: a maximum of the satellite record. It appears that prior to 1940, the long-term average of southern hemisphere sea ice was considerably greater than the satellite-record maximum observed this year. Again, the opposite is true of the northern hemisphere; the recent record minimum in NH sea ice is a record for all available data, far below any estimate in these data of NH sea ice from 1870 to the present.

We should also check that sea ice extent, as computed from HadISST data, is at least in rough agreement with satellite-derived estimates. Here is a comparison of the two data sets during the overlap period:



The agreement is pretty good, but not perfect. In particular, the southern hemisphere satellite data are slightly lower than the HadISST data. There are at least two reasons for disagreement: first, the satellite data have a finer grid size (a 25-km stereographic grid) than the HadISST data; second, the data are processed by different organizations using different algorithms. Hence we can expect good agreement at best, but we should not expect precise agreement between them.

The most obvious feature of sea ice is its annual variation: more in the winter, less in the summer. We can remove the annual cycle in many ways, including computing “anomaly” (the difference between a given month’s data and the average for that month). One disadvantage to this approach is that it removes the average annual pattern; if the annual pattern changes, then the anomaly will include the difference between this year’s annual pattern and its long-term average. Another way, which avoids this drawback and simultaneously estimates a smoothed value for the annual mean, it to apply a well-tuned wavelet transform. Doing so yields these results:



This emphasizes that the agreement between HadISST and satellite data during the period 1979 through the end of 1990 is better for the NH than the SH, but the SH disagreement is far smaller than the decline in SH extent since the early 20th century. The summer-2007 record low for the NH is, as far as the available data go, an all-time record, while the record high for the SH isn’t even close (even more true when we consider the pre-1963 data). You may wonder why the SH data doesn’t show an upward peak for the most recent data, in spite of the record maximum recorded this summer. That’s because this summer we saw a record high (satellite-era only) in SH sea ice area, but not in extent.

These facts really drive home the vast difference between the records set this year in the NH and SH. First, the SH record applies to the satellite era only; before that, SH sea ice was dramatically greater than it is today, so this summer we did not see a record on century-long time scales. Meanwhile, the NH broke the entire-data-span record. Second, the SH maximum applies to area but not extent, while the NH set a record in both. Third, the SH data broke the previous satellite-era record by a miniscule 0.9%, while the NH sea ice area broke the entire-data-span record by a whopping 27%. So although SH sea ice area set a record maximum this year, it’s really not an extraordinary event. The NH record-breaking year, however, is truly extraordinary.

Wavelet analysis supplies us with more than an estimate of the smoothed annual mean; it also supplies an estimate of the amplitude. This is the difference between the summer minimum and the winter maximum. Here is the semi-amplitude (which is simply half the amplitude) for each hemisphere:



It appears that during its recent sharp decline, NH ice extent has also shown an amplitude increase in its annual cycle. This is primarily because the maximum winter extent has declined less than the minimum summer extent, leading to a greater difference between the annual extremes. For the SH, the amplitude has fluctuated but doesn’t show signs of a secular change.

A further piece of information which emerges from the wavelet analysis is the phase. This is the moment (within the cycle) at which the smoothed curve reaches its minimum. I’ll note that this is not the actual minimum of the extent, it’s the minimum of a best-fit sinusoid for data near the moment in question. Nonetheless, it will give us an indication of whether or not the phase of the cycle is changing:



The NH shows an increase in phase from about 1960 to 1980; the summer minimum has tended to come later in the year. The SH likewise shows an increase in phase during the 1960s, with the summer minimum again tending to occur later in the year.

In summary, looking at longer-term sea ice data indicates these important points:

- This year’s SH record maximum is truly unremarkable; it applies to ice area but not ice extent, is valid only for the satellite era, and breaks the previous record by a mere 0.9%.
- The NH record minimum is truly remarkable; it applies to both area and extent, is valid for the entire time span covered by these data, and breaks the previous record by an impressive 27%.
- Over the long haul, NH ice extent has declined, almost all of the decline occuring since 1970.
- Over the long haul, SH ice extent has also declined, most of the decline occuring in the 1960s and 1970s.
- For both data sets, the latter part of the last century shows a progression toward minimum occuring later in the year.

I’ll close with the wavelet-estimated annual average ice extent for all the data, including those time periods which lack annual variations:



Perhaps you can see why I’m not impressed with the SH record maximum this year, but I’m very concerned about this year’s record NH minimum.


Someone raised the issue of the winter maximum in NH sea ice. To facilitate consideration of that issue, here’s the greatest monthly NH sea ice extent for each year in the satellite data:


Categories: Global Warming · climate change

44 responses so far ↓

  • henry // October 8, 2007 at 6:23 am

    People are quick to jump on the NH sea ice MINIMUMS, but not so quick to mention the sea ice MAXIMUMS.

    Watch the following NASA animation, and discuss it.×240.mpg

    [Response: I don’t know why the URL didn’t seem to work, but after several attempts I succeeded in getting it fixed.]

  • fergusbrown // October 8, 2007 at 8:41 am

    Hi Tamino: you don’t mention what error is estimated for the pre-satellite data, or the possibility that the estimates pre -1960’s could be out by a significant proportion; can you enlighten us? Whilst this might effect some parts of the graphs, it wouldn’t, afaik, change the bigger picture overmuch.

    [Response: For me also, that’s an important and as yet unanswered question. I’m still digesting the published documentation of the data set (Rayner et al. 2003, Journal of Geophysical Research, 108, 4407); if I find relevant information, I’ll post it.]

  • William // October 8, 2007 at 8:52 am

    “For the southern hemisphere we can reliably say that ice extent was, on average, higher from 1870 to 1940 than from 1950 to 1960, and that since 1963 it’s been lower still.” is very dubious indeed. HadIIST is based on very very very sparse data from those earlier periods. It only really exists for the SH because people need globally complete datasets. You shouldn’t use it as you have. The only way to do this, if you must, is to go back to the original data HadISST is built from.

  • anon // October 8, 2007 at 9:15 am

    They are indeed two very interesting charts. Do you have any suggestions on what caused the SH decline from about 1945? And why it should have preceded the NH decline?

    Also, how credible really is that straight line to the left in both charts? It surely cannot have really been that even?

    ‘Just the facts’, as in this case, is so much more illuminating than more rhetorical approaches.

    [Response: The straight lines are surely *not* credible in the sense that we can be sure sea ice was *not* so even. For those time periods, HadCRU has substituted the estimated longer-term average values as though the data were constant. This is because time-series studies are not the purpose of this data set; it’s purpose is to provide an uninterrupted estimate for input to numerical models of other climate phenomena. But they are credible estimates of the long-term average for the time period they cover.]

  • Eric // October 8, 2007 at 11:48 am

    Amazing and very informative post. That certainly puts a couple of dozen very long nails in the denialist ‘oh-so-what-about-the-Arctic,-look-at-all-that-ice-down-south’ coffin.

    Are you putting this on RealClimate? Surely it deserves it.

  • Hank Roberts // October 8, 2007 at 3:26 pm

    Good clarification. I’d think ‘extent’ can change depending on which way the wind blows from week to week, rearranging the loose floating ice.

    “8 October 2007 (BBC)
    “On Monday, the Canadian Coast Guard is preparing to send one its research vessels, the Amundsen, through the Northwest Passage with about 40 scientists on board.
    ….BBC News will join the ship for its journey through the Northwest Passage. …”

    Interesting politics — Russia claims half the Arctic Ocean based on an underwater ridge, but Russia and the US both claim the Northwest Passage is international water.

    And Canada’s caught short on their plans to build a military base in the Arctic because the passage melted out and looks to be navigable a couple decades sooner than anticipated.

  • John L. McCormick // October 8, 2007 at 5:04 pm

    Tamino, a most valuable post and my thanks for the time you invested in the data analysis and presentation.

    Please excuse my snipping my RC post to this post but I feel the focus on meltback is ignoring the fr more important aspects of open ocean, i.e., what does this mean for western North American temp and precip?

    From the NYT:

    Arctic Melt Unnerves the Experts

    And well it should.

    However, the super-saturation of news articles, blogs, etc focused only on the discussion of melted ice ignore the paramount issues of what does this mean to NH climate; and particularly western NA temp and precip patterns.

    Can you or someone out there focus a bit on possible impacts of ice melt (including high and low pressure systems forming where they seldom do)?

    At least one meteorologist, Stu Ostro, Senior Meteorologist with the Weather Channel is trying to understand impacts of ice melt aside from reducing polar bear habitat.

    His blog at:

    offered the following observation:

    [And the Arctic is interconnected with the rest of the world.

    Some who are skeptical of the seriousness of global warming like to point out that there’s uncertainty in the forecasts made by global climate models. I’m certainly not one to argue with model uncertainty; in fact, a couple of my blogs on about recent tropical storms and hurricanes focused on errors in the weather models which entail much shorter-range forecasts than climate models. But what seems to be happening is that if the latter are erring in one direction, it’s that they have failed to predict the rapidity with which some of the climate-related changes are taking place. ]

    I say, lets get past the hand wringing about extent of ice melt and get on with understanding its global implications for world grain supplies and hydro power in the West.

    It is time the NSF got into the act and invest as much time and money as it did to write up the Abrupt Climate Change report. We are looking at an abrupt climate change in the Arctic ice melt back but all we can see is the open ocean. There is more to this story but nobody seems ready to write Page Two.

  • Hank Roberts // October 8, 2007 at 6:57 pm

    What’s living there now, just a sample from much published work coming out:

  • John Mashey // October 8, 2007 at 6:58 pm

    1) Good post: it’s certainly a reminder of how very different the Arctic & Antarctic are. The latter (at least East Antarctica) is certainly helped by having a lot of land at 8,000+ feet, and a weather pattern that gives it some isolation.

    Whereas in the NH, when it is warmer, and the arctic sea ice melts, there’s no replacement, in the SH, if it is warmer, presumably glaciers flow faster and there are more icebergs calved off?

    Put another way, in some sense, one really wants to have Antarctic ice mass + sea ice mass to compare with Arctic sea ice mass. I’ve looked for such a chart, but haven’t found it.

    2) So NH looks in freefall, with ice-albedo feedback going hard. SH looks like a two-state system, with a 40-year transition in between, although of course, the zones of incomplete data affect the visual appearance.

    3) It’s worth looking at the GISS 3-zone chart:

    4) Minor presentation nit: on the semiamplitude charts, I sort of wished that the scale was inverted, to make it have the same orientation as the Extent charts. I also wonder whether anything interesting appears if you compare Winter(n) with Summer(n), or Summer(n) with Winter(n+1).

  • Gareth // October 8, 2007 at 10:23 pm

    JLM - thanks for that link. Verry interesting. The pdf linked on that page ( includes much more detail. Ostro’s comment about the intensification of anticyclones echoes a comment made to me by a local sever weather forecaster…

    And another tantalising hint in the context of Rahmstorf’s DO discussion at RC. Arctic/Greenland warming in a DO event is accompanied by Antarctic cooling, perhaps due to changes in oceanic circulation. Do we have that sort of effect, driven by Arctic warming, overlaying on Ostro’s speculation.

    There really is an urgent need to look at this sort of thing… Time for the climate community to sit down with the weather people.

    (And T - thanks for the excellent post. I must say I share William’s reticence about the usefulness of the SH dataset - if not his views about the Arctic ;-) )

  • WhiteBeard // October 9, 2007 at 7:22 am


    Thanks much. Didn’t post an additional request, believing you were headed in this direction. Happy I was correct.


    Winter ice extent is of less importance, as the high arctic is a place where the solar input is insignificant during winter.

    The ice reflects the sun’s radiation very well during that portion of the year when it does shine. Loss of the ice means the Earth’s total annual “retained heat” is higher.

    [Response: I’ve updated the post to include a graph of the maximum monthly sea ice extent for the NH, for each year in the satellite record.]

  • anon // October 9, 2007 at 11:14 am

    tamino, you replied to the question about the straight lines to the left. But not to the question about “what caused the SH decline from about 1945? And why it should have preceded the NH decline?”

    The trouble with the straight lines is surely what we do not know. We don’t know whether on a scale of hundreds of years the present Arctic events are as unusual as they appear in the light of the last 30. I agree that they are very striking and require careful investigation, but its hard on the facts presented so far to form a view on how alarming they are.

    Probably the refreeze and the thaw next summer will be indicative. A couple more years of the current shrinkage, and there would be less doubt that something of a very unusual nature is going on. Though even then, I find it hard to assess without knowing how much natural variability there is over centuries.

  • henry // October 9, 2007 at 11:40 am

    ” The NH record minimum is truly remarkable; it applies to both area and extent, is valid for the entire time span covered by these data, and breaks the previous record by an impressive 27%.”

    Glad you included the last chart, does the NH max and min have the same % of drop?

    “Winter ice extent is of less importance, as the high arctic is a place where the solar input is insignificant during winter.

    The ice reflects the sun’s radiation very well during that portion of the year when it does shine. Loss of the ice means the Earth’s total annual “retained heat” is higher.”

    It sounds like you are saying “sun shines, ice melts; sun doesn’t shine, ice freezes”.

    How much of the loss of winter growth/summer melt can be attributed DIRECTLY to sun exposure (albedo)?

    Increased CO2 is year round, but sunlight is seasonal, correct?

    And I’m still trying to research to see if the winter “bite” occurs in the same place year-to year (appears to be above SE Asia, where the largest coal-fired plants in the world are.) Could that increased soot be causing massive ice melt?

  • Dano // October 9, 2007 at 12:21 pm

    Well done sir, again.



  • John L. McCormick // October 9, 2007 at 12:57 pm


    Yes, it is way past time for the National Academy of Science or its equal to take on the study of the impact of Artic ice meltback on the rest of the planet. And, if anyone doubts the potential for a collapse of the Amazon rainforest, they are not keeping up with the literature. Smack in the middle of the Amazon and the Arctic is the world’s grain basket.

    I have tried repeatedly and without success to convince the American Meteorological Society to appeal to the NSF to take up that investigation and research….not any interest, thusfar.

    Maybe we are too fixated on the polar bears and the open ocean to ask such complex questions because time and money are tight and maybe we need another twenty years of observation to be certain we are not imagining all of this.

    You said: [There really is an urgent need to look at this sort of thing… Time for the climate community to sit down with the weather people.]

    Yes, tell your friends!

  • John Cross // October 9, 2007 at 1:00 pm

    I didn’t see it in the thread above so sorry if it was already posted. This is an interesting paper that presents the idea that the rapid ice loss over the last two years is due to wind circulation as much as anything else. Note for skeptics - no, it does not say that global warming is not happening and if you read the paper you will see that they even mention enhanced IR radiation and other thermal effects as instigators. What they seem to be saying is that a climate shift has been seen over the last few years that tend to cause wind to move the ice in ways not seen before (sorry, I am posting from work but left the paper home so I am posting from memory - a bad thing for me to do). More later if anyone is interested.


  • J // October 9, 2007 at 1:52 pm

    Thanks, Tamino.

    I am curious about the predictability of changes in seasonal ice extent. Just for fun, I took the satellite data and used linear regression to model March NH extent (ice maximum) as a function of the previous year’s Sept extent (minimum). The model is highly significant but with a kind of iffy r2 (0.53).

    Then I plugged in this Sept. extent (4.28 m km2) and calculated the expected value & confidence intervals for next March. Best estimate is 14.3 m km2, with a standard deviation of 0.3 m km2. Of course, this is based on a value outside the existing range of data, since this Sept. was so much lower than previous years … so take this all with a big chunk of rock salt.

  • henry // October 9, 2007 at 2:25 pm

    “John Cross // Oct 9th 2007 at 1:00 pm

    I didn’t see it in the thread above so sorry if it was already posted. This is an interesting paper that presents the idea that the rapid ice loss over the last two years is due to wind circulation as much as anything else.”

    Also absent in that article was ANY reference to AGW. The wind patterns changed at the beginning of the century (I’m assuming they mean 2000).

    When did the large coal-fired power plants in China go into operation?

    Must do more research…

  • san quintin // October 9, 2007 at 7:05 pm

    The data from New Zealand re absence of Younger Dryas cooling is certainly interesting…..but there is still a lot of evidence to show that a cooling event occurred during the YD Chronozone in the Southern Hemisphere. For instance in Patagonia we dated a 30km glacier advance which ended around this date, and there’s other stuff too. Problem is….it also coincides with the Antarctic Cold Reversal.

  • EliRabett // October 10, 2007 at 3:42 am

    I’ve been thinking about this issue. IEHO the third dimension, the thickness of the ice at maximum coverage is the predictor of the summer melt. However this is hard to measure and there is little data. What data there is suggests the trend will continue, and next year may very well be a new minimum. It is also interesting to note that winter 2007 was VERY warm in the Arctic which limited the refreeze depth.

  • henry // October 11, 2007 at 2:20 pm

    “EliRabett // Oct 10th 2007 at 3:42 am

    I’ve been thinking about this issue. IEHO the third dimension, the thickness of the ice at maximum coverage is the predictor of the summer melt. However this is hard to measure and there is little data. What data there is suggests the trend will continue, and next year may very well be a new minimum. It is also interesting to note that winter 2007 was VERY warm in the Arctic which limited the refreeze depth.”

    I see your point. It would be nice to get this thickness data, to see if there is, for lack of a better term, “core” ice.

    Since, as you said on your blog, the Arctic is a closed ocean and pretty much freezes over every winter (max ice), is there a way to overlay the summer data to see if one area in the Arctic sees little melting?

  • EliRabett // October 12, 2007 at 1:16 am

    The National Sea Ice Data Center has month by month images of ice concentration and coverage going back essentially to the first availability of satellite images. Great site. Just follow the link. The best ice thickness data is on Rothrock’s blog (links in the comments AMS) most of it comes from nuclear sub cruises. There are rumors of other (Russian/Brit) data sets out there and the US one is not complete due to secrecy issues.

    Still, the point is that once an area clears during a summer, that means that the multiyear ice is gone, and melting in the next summer is highly probably. Given a warm winter and the ice coverage goes to zilch.

    [Response: Eli has an excellent post on the topic himself; recommended reading. Also, Fergus Brown has pointed out that although we’re already into the re-freeze season, the sea ice anomaly has continued to decline.]

  • fergusbrown // October 12, 2007 at 12:46 pm

    Julienne Stroeve is probably the best person to ask about ice thickness at the moment, and Possibly Haas. The UK sub-based paper was published out of Southampton last year; they may have the dataset for HMS Tireless. They measured the W. Arctic and concluded that thickness changes were comparable to those in the E. Arctic from the R. et. al. work.

  • cce // October 12, 2007 at 3:02 pm

    Sorry for this off topic post.

    I recently gave a presentation to a friend’s Audubon group titled “The Global Warming Debate.” I’ve added a narration and converted it to a video. My goal was to create something that is both easy to understand and covers the topics that people might hear on TV , on the Internet, or in skeptical books. It should be a good companion to any (good) documentary on Global Warming. The popular press and blogosphere obsesses over these details, so it’s important for the average person to be aware of the actual facts and evidence.

    I’m hoping to get feedback so that I can correct any significant problems or inaccuracies. I don’t want to be accused of spreading misinformation.

    It is broken into 5 parts, but if you watch it from start to finish it is about 3 and 1/2 hours. So, it is not insignificant in content. While recording the narration, I went from not sick to sick to not sick so my voice can be pretty hoarse in some places. I also make no claims to my skill as a narrator.

    I cover these topics:

    Volcano Case Study/Lyndon LaRouche
    History of Global Warming
    IPCC 4th Assessment Report WGI
    The Scientific Consensus
    Attacks on the Consensus
    Greedy Scientists
    A New Ice Age
    Attributing CO2
    It’s Cold Outside
    Cosmic Rays
    Climate Models
    The Hockey Stick
    Why Now?
    Energy Issues

    I spent some time at the library and purchased a number of “historical” books. For example, I spent a lot of time documenting the “Volcanoes cause more ____ than all of mankind” skeptical claim, which many might find amusing. I retrieved some additional “global cooling” articles from the ’70s. I have Hansen’s entire 1987 Senate testimony (his ‘88 testimony was, sadly, missing). I have the original graph from Lamb first published in 1965 and covering the temperature of Central England. This eventually became the “global temperature of the last 1000 years” chart from the first IPCC report. I have a lot of other graphics and charts taken from journal articles, and swiped from the internet. Some should be familiar, and I hope that no one minds that I used them (I cited the sources). I have more than what is in the presentation, and can provide these documents to those who request them. However, my first emphasis is on creating a reasonably final version of this presentation.

    I’ve uploaded the parts to one of those free web hosting sites. It is here:

    This site may not be adequate, but for now it will have to do. I picked it mainly for its high monthly bandwidth limits, but it seems pretty sluggish sometimes.
    (I looked at Google video but the videos are too small to provide the text based detail necessary.)

    The video uses the H.264 MPEG4 codec and requires a recent version of Quicktime. It is 640X480 which is large enough for the detail I needed. I went to great lengths to make sure it will play in real time over a dialup modem. I might make a “broadband” version that will be 800X600.

    Because it is still a draft, I ask everyone not to publicize or distribute it, beyond the blog sites that I am posting this to.

    Please send me any feedback that you wish (email address on site). I have worked on this for 5 months, so I am not eager to turn it inside out. I am more interested in correcting significant errors.


  • Hank Roberts // October 12, 2007 at 3:03 pm

    Hey Tamino — looking back at Dr. Bitz’s piece at RC on Arctic sea ice modeling

    in her models,

    ” … we examined the September Arctic
    sea ice cover in the 20th and 21st centuries in climate models, and found occasional decades of very rapid retreat. The most extreme case was a decrease from 6 to 2 million square kilometers in a decade (see Fig 1). ”

    Can you, looking at what she published, see a way to compare the actual single-year change observed last month to anything in the models?

  • Hank Roberts // October 18, 2007 at 3:39 am

    Looks like they may have to add more white space to the bottom of this chart this week, if the anomaly continues to increase. Yes, this is today’s news.

    I wish RC hadn’t closed down their relevant threads so quickly.

    [Response: And there’s this too.]

  • Hank Roberts // October 18, 2007 at 5:05 am

    Hmmm, Tamino, how hard would it be to plot the Mauna Loa CO2 data against the Arctic ice data? Both have trends recently, and both have always had nice annual up and down patterns. Just wondered.

  • Andrew Dodds // October 18, 2007 at 1:32 pm

    Hank -

    It looks like that computer model has a stable state around 6 million km2 - just from eyeballing, it is stable-ish from up- to around the year 2000, where it jumps down to 6 million km2, which is stable till 2025-ish.

    There are a few possibilities here..

    (a) The model is good, but failed to take circulation changes into account, which have brought forward the 2025 collapse.

    (b) The current melt is a freak event that won’t be repeated for a couple of decades.

    (c) The first runs on the model kept predicting an almost immediate collapse of the ice sheet, and the model was then adjusted to remove this ‘Obviously wrong’ result.

  • Hank Roberts // October 18, 2007 at 2:28 pm

    I’d _guess_ the model gives output that looks like ’steps’ because Arctic sea ice has ’steps’ (one year, two year, three year and thicker ice, in annual increments. The ice gets pushed up into big broken piles of slabs by the wind, making a thick area, then refreezes and the new ice grows from below lifting the older ice.

    Just guessing, of course.

  • Hank Roberts // October 18, 2007 at 2:37 pm

    And the Southern sea ice is back to just about average today, the small positive anomaly lasted less than a month.

    Watch for people repeating the claims that it’s higher than ever, after today they’ll be wrong.

  • John L. McCormick // October 18, 2007 at 4:09 pm

    Hank, as I view the 1979-2000 mean anomaly going south it is more affirmation of how slowly is the refreeze occuring. So much open ocean giving off enourmous amounts of heat to reach the 29 .8 F temp that begins the sea water phase change to solid…which also gives up more heat in the latent heat of fusion.

    The October 16, 2007 data tells me the refreeze will not reach the mid-Oct 2006 level until mid November. That is when the beginning of the refreeze that will have to occur to materialize into something observed in early January of this year. Then, four months later the melt begins with less new ice area to shift to liquid.

    All that open water for longer periods of time has to have an impact on the NH climate and meteorology. How, how much, and when are the questions that have no answers today. But, Atlanta might give us some evidence of possible impacts.

  • Gareth // October 18, 2007 at 10:51 pm

    CT’s current ice extent graph (as of Oct 19, NZ time ) - apart from being off the bottom of the anomaly scale, is also still 1m km^2 below this time last year.

    That there Arctic ice has got a lot of catching up to do if William’s going to win his bet

  • Hank Roberts // October 25, 2007 at 12:29 am

    Looks about to go off the bottom of the recently extended chart AGAIN today.

  • Hank Roberts // October 25, 2007 at 12:41 am

  • fred // December 12, 2007 at 8:07 am

    Yes, read that earlier thread. It seems to be that between 1920 and 1975 we know little, and before 1920 just about nothing except anecdotal stuff. After 1975 we know the detail.

    Very hard to compare. Doesn’t change that what happened last year was very big in 30 year terms, though. Personal jury is still out on how big it was in century or millenia terms.

  • Wolfgang Flamme // January 11, 2008 at 4:57 pm

    I still remember the AFAIK still unexplained corrections that Cryosphere Today applied to their sea ice datasets:

    So for the sattelite aera, feel free to add or substract some 25% and for the pre-sattelite aera, fell free to ad or subtract whatever you want.

    [Response: At the link you give I see a graph with no explanation, and a link which doesn’t work. Unless you can give more details, I’ll doubt that you’re correct.]

  • Wolfgang Flamme // January 11, 2008 at 4:59 pm

    sry, ’satellite’ meant :)

  • Wolfgang Flamme // January 12, 2008 at 11:34 am

    Over at CA, in 04/2007, John Lang noticed changes had been applied:

    March 2007:

    June 2007:

    So I used Engauge Dititizer and did the math:

    With respect to remote sensing, this requires explanation. It looks like some strange surface property has misled them to over-estimate the ice area in the past. The surface with that property appearently waxes and wanes with the seasons, but does not expose a clear trend signal in NH winter ice maximum.
    It did not expose a clear trend in NH summer minimum either, but since summer 2000 suddenly its area increased considerably and this summer conditions remained as long as there is comparable data.
    Now you please tell me what kind of NH surface or surface property has changed that way since 2000 - I myself have no idea at all.

    A more trivial explanation might be this: Although it looks like they have applied their corrected analysis methods to all historical data, they have in fact not or only partly done so. Thus the time series might expose not changes in ice but rather changes in methods estimating ice … or both.

    One way or the other: Can we be sure that finally they’ve got it right this time? Their sudden changes and missing explanations have raised some doubt. Is there confidence then anyone got it right for the pre-satellite aera or even for the remote past?

  • Hank Roberts // January 12, 2008 at 4:52 pm

    Wolfgang Flamme wrote:

    > this requires explanation. …
    > Now you please tell me …

    Did you, or whoever wrote about this at CA, take notice on the Cryosphere main page of the invitation to ask?

    > Questions and/or comments to
    > William Chapman

    Why are you asking for explanations here instead?

  • Hank Roberts // January 12, 2008 at 5:00 pm

    Oh, I’ll bet you’re referring to this:

  • Hank Roberts // January 12, 2008 at 5:15 pm

    Cryosphere Today is a one-man operation, using data from NCDC/NOAA as it says at the bottom of the main page.

    Here’s a link to the abstract of the first paper (of a series, with links to the rest) that may help:

  • Hank Roberts // January 12, 2008 at 5:18 pm

    This also; note the recent posting re corrections being made to past data, still in progress:

  • Wolfgang Flamme // January 13, 2008 at 1:47 am

    @Hank Roberts

    I love you, too.

  • Increasing Amounts Of Ice Mass Have Been Lost From West Antarctica « Fermi Paradox // January 14, 2008 at 6:27 am

    […] year spreads among denialists, and the evil conpiracists refuse to talk about it. tamino, has an extended post about the sea ice in October : We can also conclude that the recent record maximum in the satellite record of […]

Leave a Comment