« Springer science summary | Main | Greenland “snow drought” spells trouble »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Ronal Larson

2.1 M sqkm

Based on extrapolation of plot of recent extent vs volume, and guessing volume at a bit less than 2 (in correct cubic km units) and necessarily going to zero together.(in 2015).


3.7 million square km. (+/- 1 million square km).

I'm guesstimating that it'll be about the same as last year based on current PIOMAS volume, NSIDC extent and Cryosphere area being about the same this year as last year, and based on general drivers not changing that much year-to-year. Put me down as an optimist, I suppose.

Kevin McKinney

Whew! For a moment there I thought I'd screwed up *again!*

Tor Bejnar

3.15 M km^2 NSIDC mean September 2013 extent. (M = million, in case my mother is reading)

I visualized a point on the graph Larry Hamilton posted based on the following:
Extent will go down as I expect volume and area to continue to decline (despite PIOMAS virtually duplicating recent years in the current season). A lot of rotten ice and slush will be counted as ice. September mean extent will decrease only a little because leads or polynyas north of 80 or 85 N will freeze over quickly once the sun goes down, and there will be more high latitude open water than ever before at minimum extent. (I’m not sure how wide a lead or polynyas needs to be before being counted as water, but I think some will be big enough.) This northern most area will only be ice-free for a brief time; the water will not have much time to absorb solar radiation (nor mix with deeper water) that would delay re-freezing, and the month average will be rather larger than the minimum. September cyclones, however, would speed melting and delay refreezing. The southern areas of the Arctic will continue to melt when the northern most areas start to refreeze, but there won’t be much ice in southern areas.

If PIOMAS volume is currently about the same as it was a year ago, and more area is covered by <2 meter thick ice, then >2 meter thick ice needs, on average, to be thicker than last year’s, and most of this (extent-wise) will stay around into another freezing season (whence some of it will be exported out the Fram).

I expect average September Arctic ice extent to go below 1 M km^2 about two years after minimum ice area and volume crash below 1 M km^2 and 0.75 M km^3, respectively. Increased methane will cause the Siberian Arctic to melt faster; Mt. Cleveland’s eruption may slow down Canadian Arctic melting that would otherwise melt faster with all the relatively thinly iced leads (re Cracks of Doom).

My mom always uses “M” for thousand (from Roman numerals). I hope Espen is using this nomenclature when he guesses “3,014.699 M km2” or, as a European, uses a comma for the decimal point.

Espen Olsen


Thanks Tor :

3,014,699 M kms

Ethan O'Connor

@Lodger: I think your discussion points were spot on but I'm cautious about #5:

"a Sep 2013 SIE < 2.9 M km^2 or > 4.9 km^2 rejects the linear trend model of SIE decline at the 95% confidence level."

I think the linear model over the 2003-2013 span is rejected such less strongly by such a value. We are evaluating the model with the new observation in mind, but it shouldn't have any more weight than any other year in evaluating the model.

This means the question is: "What is the probability of at least 1 measurement with a residual > 2stddev in a set of 11 measurements?"

This has a probability of about 0.4

Of course, you can look at 2007 and instead ask how likely it is to get 2 measurements with residuals of z >= 2; this is about 0.086.

In any case, however, I think that a viable linear-decline model would assume long tails on the residuals for the measurements -- possibly drawn from distributions with indeterminate std. dev. Thus the calculated probabilities of "2 sigma" events are misleading when sigma is estimated from 10 events.

Does this seem reasonable?

Ethan O'Connor

[Continued discussion of evaluation of linear trend 2003-2013 based on 2013 observation]

Another way to look at this is to perform distribution fit tests against the extent residuals. The residual z scores for 2003-2012 fit a normal distribution very well, and we can plot the goodness of fit versus a range of hypothetical 2013 zscores (using Kolmogorov-Smirnov P-Values as goodness of fit):


Frankd 1977

3.85 million km2 +- 0.89 ...here's how:

I took the September averages from 2000-2012 and plotted a trend line. Doing the monthly averages gave me a smaller SD then just listing the minimums.Then I derived the average deviation of the means from the minimums and subtracted it from the projected 2013 monthly mean.

Rob Dekker

Sorry for the OT, but need to correct something :
Erimassa said :

assuming average thickness of 2m for the 83000km2 daily melt. The heat of fusion for water is ~333550J/kg. The energy required to melt this much of 0 degree ice is 55,37 PJ which translates to 15,38 GWh. ..... which is only 0,026% of the daily production of electricity in the world ..... Correction : 26 %

Erimassa, you are still a factor 1000 off :o(

83000 km^2 x 2 m = 166 Gton ice, which requires 55 x 10^18 J (55 exajoule) to melt. That is 15,000 TWh. Per day.
That energy translates to 636 TW in 'cooling' power during the melting season.

Daily total electricity use is something like 57 TWh, so ice melt on a typical Arctic day requires about 263 times (26000 %) of world-wide electricity use.

That melting ice keeps the entire Northern Hemisphere cool in summer, which makes us all wonder what will happen once it's gone...

Mark Kosir

2,732,000 km^2. I'm visual and relying heavily on graph curves.


I meant minimum rather than mean. But I'll stick my neck out and leave it there.

Minimum has been extending further and further into September, regrowth has been slower. The mean for September must also be falling faster than expected.

Phil Scadden

3.9 - continued downward trend but guessing we wont get a weather extreme two years in a row.

Artful Dodger

Hi Larry,

I've thought about this for a full day now, I and think I've discovered why quite a few people has been confused about the question for this survey. I believe it's the Title:

"Crowd-Source Prediction of Minimum Arctic Sea Ice"

Often when people read "Minimum" unconsciously they append "Daily" to the thought. Instead, I suggest that word be changed to "September". Indeed, as the climate warms, serious people are considering if the daily minimum could occur in October or later!

Additionally, since there are three common metrics for sea ice, specifying the measure for this topic in the title would also help focus the mind. Thus, a revised subject could read:

"Crowd-Source Prediction of September Arctic Sea Ice Extent"

hth. ;^)


David Goldstein

Well, the author of the following Huffington Post article wants it to go to zero this year! Some kind of nut-job I guess :) (ahem, that would be myself, actually). I also mention and include the url of the Arctic Sea Ice Blog in my article "Burn, Baby, Burn: A Climate Activist's Wish for Superstorms, Wildfires and Drought." Please check it out here: http://www.huffingtonpost.com/davidgoldstein/burn-baby-burn-a-climate-_b_3224411.html?utm_hp_ref=climate-change


3.25 million square km

Watching the numbers, we are due for a dead cat bounce. Looking at the ice it could be very low. Over 4 would surprise me, just under 2 would not.

Clearly I am putting too much faith in the bounce of numbers.

Mathew Weatherwatcher

3.8 m km^2


Thanks Rob Dekker for getting the scale of the problem right...

Meanwhile the crowd-source estimate has again risen a bit (82 entries) and is now at

2,9031499 Mkm2 with +/-0,636875

and the the size of circular slab of 2m thick ice to melt per day has shrunk to 260 km (162 miles) of diameter. Ain't maths fun done correctly, one could almost think this situation is out of our hands?



I agree with Hans that 4 consecutive years of decline is unprecedented in the record; however, I suspect that the loss in ice volume is such that the buffers provided by multiyear ice in the past aren't available for the natural fluctuation that has occurred in prior years.


Nice poll, Larry,

My prognose for mean September extent as measured by NSIDC is 3,28 Mkm2, derived from my latest April prediction of 2,0 Mkm2 for SI Area minimum (3,13 for SIE and 2600 km3 SI Volume).

The remains almost confined within the ‘arbitrary’ limits of MODIS tiles r03c03 and r04c04, the rest of the Arctic Ocean will be only good for lurking at unusual weather and algae patterns.

I started off my prognosis in February, suggesting two modes (2,5 / 4,0 if ‘normal’, 1,7/2,6 if ‘dipole’).
In two steps (down) I got to my present prediction. I have three main arguments:

1) mainly FYI 2) not enough winter power 3) Ice quality.

Meanwhile, I expect the mediagenic consequences of accelerating changes in the ocean-atmosphere dynamics in parts of the mid-latitudes. It will still be difficult to get a grip on the whole picture, but this will become clearer…

Happy to join up again. I’ve reset my blogging routines because I concluded it was affecting the quality of my work (and I expect my boss to be less tolerant than my wife).
There’ll be less word-salad by Werther, but you can consider me watching…

Timothy Astin

3.7 M km^2

My method is eye-balling an exponential best fit extrapolation. While knowing that I should stop there, I then add some unscientific guessing that the pattern of deviation means some rebound from last year is quite likely. Then I add a gut feeling that the 12/13 arctic winter remained warmer than average because of exporting cold to mid-latitudes.

Hans Gunnstaddar

Regarding David Goldstein's Huffington article: I concur! Too bad it will take such harsh environmental responses to penetrate the density of self preoccupation, but at this stage whatever it takes.

Not only transition to renewable energy sources away from FF, but also build so much renewables we can build manuf. plants to draw in air and sequestor carbon to use for road base or some other products to bring down CO2 levels, not just for stabilizing the weather but also to reverse acidification of oceans. With healthier oceans more phytoplankton will grow and that will also seqestor carbon.

Running past the 3rd base coach's stop signal to surpass 400 ppm CO2 is no way to run a planet!


3.64 million square kilometres.

I think one has to be very cautious about deviating from the excellent predictions of Tamino, which are for 3.97 +/- 0.9 million square kilometres. However, I have chosen to do so.

My prediction is that there is nearly a 50% chance of a new record being set this year, with my 3.64 figure being slightly above last years 3.61. My reasons for this are:

1. Last year did not seem to be set up for a big melt. We thought the ice was thicker in the Beaufort Sea. There wasn't a consistent Arctic Dipole as in 2007. Yet the record was still smashed.

2. PIOMAS has been largely vindicated by the Cryosat observations. At some point the decline in thickness will accelerate the decline in extent - perhaps the cracking indicates that this year is that year?

I've not gone too far below Tamino's forecast though. It's worth remembering that the NSIDC mean September sea ice extent has not yet set a new record two years in a row. And tomorrow will be six weeks from midsummer.

We are already in the quarter of the year with most solar insulation into the Arctic, and there is still 13 million square kilometres of sea-ice reflecting that sunlight back to space. On Midsummer's day itself, it is likely that sea-ice extent will still be more than 10 million square kilometres of sea-ice.

The final interesting point I would make is that following the 2007 minimum, none of the subsequent years have had a minimum extent above the previous minimum (of 2005). All have had less sea-ice than in 2005. If 2013 follows this pattern then we would expect that it would also be below 2007, if not necessarily below 2012.

Tony Duncan

4.4. I just am conservative and imagine am somewhat hopeful that it won't be really bad. I also will enjoy it when denier's slap their backs on another resounding recovery


2.2 million square kilometres.

If the estimate required is for average sea ice in September it might be a bit higher but I'm going lower than most because

1. Piomas volume is dropping and I have suspect that Piomas is beginning to estimate slightly higher figures than it used to.

2. Much of the ice is in a position which will melt more easily. (NOSEDIVE! on DOSBAT)

3. Rotten ice. Reinforced by reports of algal blooms.

4. The cracks in the ice.

5. More warm water coming in from the south.

6. More mixing of lower and warmer salty layers with surface water.

7. Tamino's statistical predictions set a target before most of the above are considered.

8. Methane's lurking in my mind somewhere too.

Sources: Neven and the people on this blog, Chris Reynolds' DOSBAT, Tamino, David Barber's talk on rotten ice last year. Thank's to you all. Especially the amateurs.

Hans Verbeek

Neven, if the relation between summer Arctic sea-ice-extent and AO is real, we can use the minimum-sea-ice-extent of 2013 to predict the value of the AO in the winter of 2014.
Less sea-ice could mean a more negative Arctic Oscillation, couldn't it?

Account Deleted

About 3.0 M km2 if weather favors the survival of Beaufort Sea MYI; 2.0 M km2 or less otherwise.

I lean toward the 2.0 M km2 value, for the poll.

Less sea-ice could mean a more negative Arctic Oscillation, couldn't it?

Everything's possible, Hans. The good, the bad and the ugly.

Susan Anderson


Real amateur (not like you skilled amateurs), but following along with interest and passion. So many things. Volume loss pre-eminent. Early wildfires. Weird circulatory alterations - cooler ocean to the south. Dead cat bounce trumped by above.

also, plus or minus a whole M km^2

Paul Beckwith


R. Gates

Hans Verbeek,

A negative AO may have some association with lower sea ice, but correlation not causation. Planetary wave activity, SSW's, and the Brewer-Dobson circulation also are also associated with the AO index with the most negative index we see occurring in days, weeks, and months following a large SSW event.

Matt Arkell

2.75 mean area, 2.5 minimum.

I'm expecting volume to be lower again this year, which will be reflected in extent dropping again.

Patrice Monroe Pustavrh

Mauna Loa average daily concentration of CO2 is 400.03 ppm for May 9th 2013. Damn!!

L. Hamilton

When I first wrote this post, I wondered whether it would get 8 answers or 18. The flow seems to be tapering off now but my rough count has 95 so far. Thanks, folks! The polls are still open of course, but I will definitely follow up with new versions for the SEARCH July and August cycles. Then comes the fun part of analysis, comparing the crowd-sourced results with estimates by scientists and the arbiter for us all, the real sea ice.

To provide some quick feedback here, the mean answer so far is about 2.9, with a median of 3. The distribution has a mild negative skew so I'm inclined to favor medians (or other outlier-resistant statistics) as summaries. Here's the whole distribution:

Stem-and-leaf plot for extent (June SIO estimate by comment in ASI blog)

extent rounded to nearest multiple of .1
plot in units of .1

0* | 000
0. |
1* | 013
1. | 578
2* | 00000012222444
2. | 555555778888888888999999
3* | 0000000111112222333334
3. | 5567778888888999
4* | 000011334
4. |
5* | 0

L. Hamilton

Lodger makes a good point above that my title might have confused some people about the monthly vs. 1-day minimum. I'll be more obvious next time.

The SEARCH Sea Ice Outlook focuses on September monthly mean, so that should be our target for this crowd-source estimate too. I stated that somewhere in the original post but it not prominently enough. There's a learning curve for us all.

L. Hamilton

Hmmm, an earlier post might have been spam-filtered? In it I mentioned that we have (by my rough count) 95 estimates so far, with a median of 3.0. Here's the distribution:

Stem-and-leaf plot for extent (June SIO estimate by comment in ASI blog)

extent rounded to nearest multiple of .1
plot in units of .1

0* | 000
0. |
1* | 013
1. | 578
2* | 00000012222444
2. | 555555778888888888999999
3* | 0000000111112222333334
3. | 5567778888888999
4* | 000011334
4. |
5* | 0


3.1 M Km2

And Mauna Loa CO2 at 400.03 ppm yesterday.

Just a number I suppose.

Hmmm, an earlier post might have been spam-filtered?

Yes, sorry about that, Larry. Apparently not even a guest blogger can comment on his own guest blog. :-|

I'm still in contact with TypePad, but apparently there's nothing they can do. I've mentioned the word whitelist, but apparently it's complex.

A median of 3 million km2 for NSIDC smacks of cojones, but why not? I've been way too conservative so far. That reticence creeps in as soon as you feel some responsibility, I guess. I'm afraid of screwing up and then getting used by fake skeptics as proof that AGW is a hoax.

Espen Olsen


"That reticence creeps in as soon as you feel some responsibility, I guess."

I am getting worried!?

Hans Gunnstaddar


Climate Milestone: Earth’s CO2 Level Passes 400 ppm

From the article: An instrument near the summit of Mauna Loa in Hawaii recorded a long-awaited climate milestone today: the amount of carbon dioxide in the atmosphere there has exceeded 400 parts per million (ppm) for the first time in 55 years of measurement—and probably more than 3 million years of Earth history.

The last time the concentration of Earth's main greenhouse gas reached this mark, horses and camels lived in the high Arctic. Seas were at least 30 feet higher—at a level that today would inundate major cities around the world.

The planet was about 2 to 3 degrees Celsius (3.6 to 5.4 degrees Fahrenheit) warmer. But the Earth then was in the final stage of a prolonged greenhouse epoch, and CO2 concentrations were on their way down. This time, 400 ppm is a milepost on a far more rapid uphill climb toward an uncertain climate future.

If the seas were 30 feet higher when CO2 levels were previously this high, means the heat hasn’t had sufficient time in the atmosphere to penetrate the oceans. So that is how much sea level rise is already baked in, presuming CO2 is not sequestered enough to avert those consequences. Camels in the Arctic? Anyone else on board to transition to renewables?


2013 should turn out to be a very interesting year and it should provide the answer if the arctic ice is deceasing ala the exponential trend or the Gompertz trend.

My estimate 2.98 Million Km^2.

Since I am inclined to believe the exponential decline over the gompertz although Wipneus's gridded data does show Gompertz behavior I take the prediction for 2.1 Km^3 from PIOMAS and I divide it by the average September thickness these last 3 years of 1.1 m and I multiply it by 1.56 to get extent instead of area. This provides my estimate of just under 3MKm^2.
My reason for going with the exponential decline is that when old ice melts the density of the melt is very low and forms a barrier undeneath the ice keeping the warmer water away from the ice reducing bottom melt. When the melt is first year ice this low density layer is significantly reduced and as a result the ice melts faster. With the ever smaller amount of old ice the ice bottom melt rate wil be enhanced and PIOMAS will continue to show an exponential decline. Since the North Pole Web Cams clearly show that melt ponds are totally frozen over by the beginning of August this bottom melt controls the September minimum.

R. Gates

Hans said: "If the seas were 30 feet higher when CO2 levels were previously this high, means the heat hasn’t had sufficient time in the atmosphere to penetrate the oceans."

Hans, I am not picking on you, but there are a few misperceptions here in your comment that should be cleared up. First, the direction of energy flow on a net basis across the planet is always from ocean to atmosphere. Literally the oceans keep the atmosphere warmer than it would be without the oceans. The vast majority of the energy in the oceans comes directly from SW solar penetrating down into the ocean. Second, the majority of the extra energy retained by the Earth system due to the rapidly increasing GH gas concentrations has already "gone into the oceans". I put this in quotes because it is the common way even scientists speak about the energy imbalance caused by adding GH gases, but it is thermodynamically inaccurate. GH gases in the atmosphere are not adding heat to the oceans, but rather they ultimately slow the rate of exchange of energy from ocean to atmosphere. This is no different than you putting a jacket on to keep warm in the winter. The jacket (i.e. greenhouse gases) does not add energy to your body, but slows the rate of flow from your body to the cold air outside the jacket (i.e. outer space). Thus, adding GH gases to the atmosphere slow the rate of energy flow from ocean to atmosphere to space. This is the exact reason that ocean heat content has been rising so dramatically over the past 40+ years. Finally, and most importantly, the energy that the ocean is retaining is creating huge effects on the Earth system already, and will continue to do so now for centuries (and with increasing GH gas concentrations, the effects will even get more severe). Melting sea ice is just one of the effects. And in the coming decades, once we are well passed the first ice-free summer Arctic milestone, and well on our way to a Pliocene-like climate, the oceans will be many meters higher than they are today, but not because the atmosphere is heating the oceans, but because the warmer atmosphere and warmer oceans combined will be reducing the amount of ice in Greenland and Antarctica so dramatically.

Hans Gunnstaddar

Thanks R. Gates for pointing that out. Putting on a jacket in winter sealed the concept. Always open to learning more.

Chief Hydrologist

5.7 million km2

‘With respect to climate signals, the highest degree of seasonal predictability occurs during La Niñas that are concurrent with the negative mode of the PDO. This is a result of the increase in frequency of negative PNA pattern (zonal) during negative modes of the PDO (Bond and Harrison, 2000).’ http://onlinelibrary.wiley.com/doi/10.1002/joc.686/pdf

This year started off fairly warm but is cooling rapidly as La Niña evolves in the eastern Pacific. It is an early and dramatic cooling of the eastern Pacific -it has me wondering just how substantial this event could be.


In particular Arctic temps are nowhere near as high as last year and will remain very much cooler.


Longer term – decadally – it is difficult not to anticipate a cooling influence for the Arctic from an intensifying Interdecadal Pacific Oscillation.



So we have a year that is nowhere near the top 10 in temps – and some longer term cooling possible if not probable.

I would use 1999 as an indicator of September ice extent – 5.7 million km2


4.5 million /+-0.3 million

It is impossible to judge by the state of the ice so early in the season. But given that we've had three decreases in a row, statistically we can expect a bounce back. However as the trend is down, I don't expect that will bounce back to 2009 level.

R. Gates

Respectfully, I sincerely believe Chief Hydrologist's guess as to sea ice extent needs to thrown out if the mix. Not only is it absurdly high with a flimsy basis, I happen to think he has intentionally posted this high estimate in an attempt to raise the overall average. I've followed his posts for many years and quite frankly-- I think he just enjoys being a fly in the ointment-- especially with those who happen to think humans are altering Earth's climate in significant ways.


3.4 M km^2 +/-
I dont have any v logical scientific data to base this on:
I was initially hopeful of a 'dead cat bounce' (goodness what a descriptive turn of phrase that is!), well would really prefer a 'live cat' one better, taking things back to pre1970 levels.
But when I think about the positive & negative forces potentially operating this season my instinct is that any +ve 'bounce' will get swamped leaving us a bit below last years minimum.
The error margins are obviously large; more cloudy days? strong storms again this year? etc etc

Peter Ellis

No need to throw out data, it's irrelevant given how many estimates are already in. Let's say he's 4 million above the consensus (which he isn't). With ~100 guesses already in, this one wild guess will only skew the average up by 0.04 M km^2, which is completely irrelevant.

If you look at the graph, it's pretty symmetrical: the tail at ~5M balances the tail at zero. Both are equally unrealistic.

Peter Ellis

To make the same point another way: the mean and the median of the crowd-sourced estimate are both currently around 2.9, which means that Hydrologist is less of an outlier than the folks predicting zero.

Lawrence Coleman

between 2-2.5mil sq kms.
I now believe that the annual variation in ice albedo is less than the 5 yearly averaged downward trend. Thus I sadly expect to see from now on every consecutive year posting record lows. I'll put a rough date on an ice free summer arctic at 2018. Based on the decimation of the hard perennial arctic ice over the last 5 years and anecdotal reports of the state and quality of the rest of the ice sheet by research vessels and submarines etc.
I am not a climate scientist but I have been researching climate change now for just over 12 years.

L. Hamilton

To R. Gates and others who might wonder about outliers --

Internet polls are, you might say, a dime a dozen. I don't know if this one will be any righter than others, but I can promise it will be thoughtfully analyzed. (And not just by me, it's crowd-sourced, after all.)

Some statistical summaries such as mean, standard deviation, correlation, OLS regression (the classical family of moment or least-squares statistics) are notoriously not resistant to outliers, meaning that one or a few wild values can move them any distance.

Other summaries, however, such as median or robust regression, have higher resistance to outliers. These are good choices if you don't want your statistics to be controlled by a few wild values, but also don't want to arbitrarily exclude certain values from the data. A typo could shift the highest guess in our poll from 6 million to 600 million, and the median would not move at all.

In my (inadvertently) dual posts above I cited the median for this reason, although currently the median and mean are pretty close so it doesn't make much difference. Also, I showed the whole distribution as a stem-and-leaf display, a method invented by the very clever John Tukey as part of his Exploratory Data Analysis (EDA) approach -- which emphasizes graphical displays that make no assumptions about the data, and can highlight unexpected features such as outliers. Noticing outliers turns them from statistical pitfall into information, which could be interesting in its own right.

Anyway, at this point I'm not worried about outliers, apart from correcting obvious mistakes (mine or yours) if such occur. The analysis presented will include a fair central value (such as median), along with a graphical look at how estimates varied around that -- out to the extremes.



Are you looking at all at the comments associated with the numerical answers?

I know when I have had to analyze survey results, the comments provided as much information clues as did the distribution of the results.

james cobban

2.8Mkm^2 - based on the weakness of this year's ice, the trend in volume reduction, decreased albedo, increased methane from melting permafrost and clathrates, increased CO2, increased particulate matter from burning forests and tundra, and all the other bits of info I've gleaned from this great site.

L. Hamilton

Our raw data consist of exactly what you see here, the OP plus all the comments. Analysis starts with the numerical estimates. I don't know how that will turn out, but the comments will be a first resort to interpret.

For me this whole undertaking has an exploratory, EDA-like flavor. Take a robust/visual approach and see what we learn.


It is just a guess as it remains very difficult to make any reliable prediction at this stage, so many uncertain factors may change the overall picture like down welling short wave radiation, long wave radiation, atmospheric circulation (dipole anomaly “DA” and how it can direct the transpolar drift sea current), Canada and Eurasia snow extent anomaly (which may influence the DA http://onlinelibrary.wiley.com/doi/10.1029/2012GL053268/abstract?deniedAccessCustomisedMessage=&userIsAuthenticated=false, http://dosbat.blogspot.co.uk/2013/03/eurasian-snow-cover-and-atmospheric.html therefore transpolar drift), Arctic cyclone, water pools, warm sea current from the Pacific /Bering strait…

I would like to believe that there should be a very slight extent upsurge after the Sept 2012 record, as it happened after 2005 and 2007, however it seems difficult, although many factors may change the overall picture until the end of the melting season. My guess is based on a few hard sticks:

-PIOMAS volume (largely backed up by Cryosat-2 and the fantastic work from late Seymour Laxon and his team) in April 2013 nearly at the same level as in April 2011 and 2012, while 2008 and 2009 September rebounds were preceded by significantly higher PIOMAS volume in April compared to 2007.

-After a record NH snow extent in April 2013 since 1996, the extent anomaly became negative only recently. At week 18, snow cover remains higher than in 2007, 2011, 2012, particularly for North America, and significantly lower than in 2009, 2010 (http://climate.rutgers.edu/snowcover/table_area.php?ui_set=0). At day 131, May 11, snow melting is well ahead of 2012 in Siberia (excluding Central Asia) while lagging behind in Canada/Alaska (eye balling), however still probably lagging behind 2011. At this stage, it is probably too early to conclude anything in relation to a potential impact on DA in June, July and August. However, the quicker snow melt in Siberia and the recent appearance of numerous polynias on the Arctic Siberian side, if not impeded by external drift of ice, may lead to a rapid melt on the Siberian side, and will it favor a DA?

- Chris Reynold’s Dosbat interesting comment on a sharp decrease of area of ice thicker than 2m, and correlative increase of area less than 2m thick, although the 2010 drop in area of >2m thickness only materialized in SIE in 2011.

-No more multi year ice barrier in Beaufort / Chukchi sea in front of Bering strait, which disintegrated last year, this will leave an open door for the warm pacific water, which is known to be rather shallow and plays an active part in melting this part of Arctic sea ice (however, the intensity of this warm current is highly variable from one year to another).

-I do not see any potential impact from Cleveland volcano at the moment as it was emitting (on the 04/05) “A small, low-altitude ash cloud along with high surface temperatures at the summit were observed in satellite images starting at 0717”, and on the 05/05 a “continuous low-level emissions of gas, steam, and minor amounts of ash producing a faint plume that drifted East below 4.6 km (15,000 ft) a.s.l.”. Besides similar low level ash clouds happened in June, July and August 2012 (source: http://www.volcano.si.edu/world/volcano.cfm?vnum=1101-24-&volpage=weekly). Only a Pinatubo 1991 eruption style might give a real bounce back (or respite?) to the deadly cat above a potential small bounce back due to year to year usual climate variability. (Dorn et al, 2008, http://www.benthamscience.com/open/toascj/articles/V002/91TOASCJ.pdf)

My overall unscientific guess will go at around a minimum of 3.7m km2 SIE (considering that a small amount of 2012 melting was due to the super cyclone although the 2007 record would have been broken without, and it is statistically unlikely that would happen two years in a row) and 3.9m km2 of monthly average extent, while a drastic drop may happen in 2014 or 15, and a deadly stroke a few years after (the 2007 perfect weather storm happening on average every 10 years).

Shared Humanity

I realize, to be entirely on topic, I am suppose to vote on minimum extent but I do not feel informed enough to do this. I do thoroughly enjoy visiting each day to read comments and grow my understanding.

R. Gates.....I would like to thank you on your explanation of heat transfer between the oceans and atmosphere. It cleared up my own misconception as well. One more bit of personal knowledge.

On this topic and the impact of 400 CO2 levels and sea rise....

Hans said: "If the seas were 30 feet higher when CO2 levels were previously this high, means the heat hasn’t had sufficient time in the atmosphere to penetrate the oceans."

I believe the most relevant point of this 30 foot sea level rise and the 400 CO2 level is that it did occur on the down slope of CO2 levels. Since ocean levels rise slowly in relation to warming (thousands of years if you are an optimist; hundreds if you are a pessimist) the 30 foot higher levels is more reflective of peak CO2 levels preceding the 400 CO2 level during the decline. What were the peak CO2 levels from this period?

If this is correct, it gives me some hope that immediate action on CO2 emissions would allow us to avoid such an increase in sea level.

R. Gates

Put me down for 2.95 million sq. km. for the Sept. mean extent.

I do have to say that the nature, timing, and speed of the descent of the current high pressure anomaly in the Stratosphere to the troposphere could have an impact on the melt at a key point in the summer melt season. This high pressure anomaly can be seen here:


This high pressure could remain in the stratosphere, though it has indications it will not and is descending. Should it hit the lower troposphere with strength over the Arctic at about the same time as peak insolation, we'll get vigorous melt and lots of ponding.

R. Gates

Here's a more contextual image of the current high pressure anomaly in the stratosphere over the pole;


Will it work its way down to the surface? If so, when, and with what force and how long will it linger? The timing during peak insolation over the Arctic could have a major influence on early summer melt.


4.1 M km^2

Three years in a row going down, so it's time for a little rebound. My intuition.

Hans Gunnstaddar

I'm right there with you on that one Surveyork. My prediction is 4.35 (from earlier in these postings), both in the same camp as 'Rebounders' after 3 straight years of declines.

There are 3 camps, the Rebounders (greater than 2012), Decliners (lower than 2012) and Zero's in which most if not all the ice melts or is transported by currents out of the Arctic.

Nature is not linear, i.e. even in an arcing trend-line there are fluctuating years. So far in the limited years shown on the graph up top, there hasn't been more than 3 years running of declines and there has not been two years running of new record minimums set. It took from 2007 to 2012 to set a new record, so the chances of a new record this year, in spite of volume losses, is in my opinion very unlikely.

This non-linear, non-predictability is what gives the denial crowd fuel as they are able to keep up their recovery mantra. By the time a new record is set, the previous recovery mantra will have been forgotten.

Possibly homo sapiens can be separated into two camps; those that can understand an uneven trend occurring over time, and those that can only understand a trend if it is purely linear. We may even find out there are gene markers that predispose people into one camp or the other. Or we may find it is a factor of education, but in any case there seem to be those two levels of thinking.


currently there are 104 entries by my count, I may have missed one or two if they're not clearly presented (embedded in long posts with lots of numbers). Average is now pretty close to 2,97Mkm2 with +/- 0,6 (this is the average of the entries that have the error bar included, f.e. 2-2,5 is interpreted as 2,25 +/- 0,25). For this to be correct, the daily melt required now (may 11th to september 11th, which is pretty close to the september average) is now at 84200km2. The melt should of course start speeding up shortly after snow has left N.Siberia and CAA.

Espen Olsen

But the fact is it has been growing for the past 2 days IJIS: 12,388,281 km2 (May 12, 2013)


whoopsie, now that I've actually done the maths, I must inform it's the september MINIMUM in CT area which is pretty close to September 11th. The September AVERAGE is pretty close to August 31st or September 21st so the area calculation is also wrong. Maybe the organizers will calculate this from the correct data (NSIDC extent) too.

(correcting....) 92550km2/day would be the amount of melt to August 31st, thank you for your patience.


3.57 M km^2
At the moment, 2013 looks similar to 2012, so just guessing that the mean will be similar. Waiting to see if the MYI around CAA gets blown apart. And there is no "science" for my estimate, just the trusty eyeball :-)


I will go with 5.5.

Weather patterns are noticably changing in this northern land and this time. AO is turning positive. Polar vortex strengthening. Arctic temperatures below average.

Based on experience with similar years, I predict, with the aid of my crystal ball, that these conditions will persevere throughout this summer, significantly reducing the melting of the sea ice compared to the previous few years.

Rob Dekker

3.8 million km^2 (+/- 600k)

This is still a bit of a wild guess, but based on the following data :
- 2012 NSIDC Sept ave min was 3.61.
- Volume by PIOMAS at this point is pretty close to 2012.
- Extent and area at this point are pretty close to 2012.
- I address 200 k melt in 2012 to the Great Arctic Cyclone of 2012, which is pretty close to the estimate by the PIOMAS team :
who estimate that the GAC caused 150 k km^2 of extra extent loss.
- Regression between May snow cover and Sept ice minimum suggests that 1 million km^2 snow cover loss correlates with about 500 k km^2 Sept sea ice extent loss.
However, data from NH snow cover is not yet giving a clear difference with 2012 :
This chart shows that snow cover is at this point WAY below average, but not much different from last year's (record) snow cover anomaly.
So I can't subtract or add a delta wrt 2012 based on snow cover just yet.

That makes by prediction at this point 2012 minimum + 200 k for GAC is 3.8 million km^2.
So I assume (perhaps optimistic) that there will be no GAC this year.

The error margin is based on standard deviation of regression analysis of May snow cover versus Sept ave sea ice minimum, data over the past 20 years.

For the end-of-June prediction, I'll attempt to put in extent prediction based on volume reduction and snow cover and extent priors. That should get us more accurate numbers and reduced uncertainty margins.


If we go by the NSIDC data, the year 2013 SIE could be 3.9 M Sq. km. (with an UL of 5 and LL of 3.0 M Sq. km.). However, if we go by an independent study [based on the 1999-2009 QuikSCAT sigma-0 data], we may probably give the extent value at around 4.8+/- 1.2 M Sq. km.

Patrice Monroe Pustavrh

I'd go for 4.1 mio km2.


4.15 million km^2.

I expect a rebound from 2012's record. 2012 had much more open water at this date, soaking up solar energy throughout June. Snow cover near the Arctic Ocean was also lower. That will be hard to catch up with.

The continents are warm now, but the Arctic itself is lying in a cold cocoon. This could be a pattern.


The SIE IJIS gauge is now (June 02) close to the 1990s average. On that basis, it is not impossible (although I must admit not very likely) to see a minimum close to average minimum for that period i.e 6.6 million.
I am not saying we should read anything significant into this, just that climate in particular arctic climate is so impredictable that we cannot rule out the possibility of a high rebound for a few years even if in the medium/ long term the downward trends proves correct. I believe a paper published last year predicted just that: i.e the possibility that we might see SIE at higher levels for 5 or 10 years before resuming on its downward trend.

James Lovejoy

My WAG (I won't dignify it as a prediction) is 3.75 +- 0.9 for the average, and 3.40 =- 1.0 for the minimum.

Others are talking about "recovery" or reversion to mean. I disagree. I think we are just going to have a very unfavorable year for melt. Unfortunately, the usual suspects are going to use that as an excuse to ignore that volume is still going down.

Yep, I expect Piomass to be lower this year than last, but not by much. See above the unfavorable melt year.


I agree with you James, we will just have a very unfavourable melt year.
Unfortunately this will be seized upon by deniers of all stripes to justify inaction on carbon emissions. We can already hear things like " the Antarctic sea ice is growing" as a result of the very high SIE minimum down there in February and we can expect the same sort of argument if the arctic SIE bounces back in September.
Unfortunately there is more to this than just science.It is the societal perception of reality rather than scientific facts that will affect global human behaviour and this will ultimately affect reality.


I completely agree with you James, it will be just an unfavourable melt year.
Unfortunately this will be seized upon by deniers of all stripes to justify inaction on carbon emissions. We can already hear things like "the Antractic ice is actually growing" following the high minimum down there in February. We can expect similar arguments from the deniers camp if the Arctic sea ice minimum bounces back to pre 2000 levels in September.
I believe this issue is not just about science. Scientific facts about the climate are clouded in misunderstanding and therefore believed to be controversial. Perception of reality (particularly perception spruiked by the Murdoch media) will determine collective human behaviour rather than the examination of facts and this is very scary!

L. Hamilton

Kamaljit_raj, should I put you down for 3.9, 4.8 or neither? To include in the stats I need a point estimate, although discussion can range far from that here.

The median so far (111 by my count) is 3.0, with late-arriving estimates tending to be higher than those from early May.

The SEARCH SIO deadline for June estimates is this Friday, June 7. Since this is our first iteration of this crowd-source experiment I had not planned to send the results directly to SEARCH. A few bugs need working out -- confusion about minimum day vs. September mean, and the problem of drawing most of the estimates a month before the official deadline. I'll try to improve both points for the next iteration/July deadline. Part of the interest will be in watching how estimates evolve with the season.

If you haven't sent yours in yet and would like to, there are still a couple more days.


For this month I'm going to go with 2.9, up just a nickel from last month, due to the slow start. The reason for staying with another record breaking year is the condition of the ice, not the current extent. And as the discussions and animations on the other threads have indicated, the current condition of the is is deplorable.


Put me down for the 1990s average of 6.6 million.

1. the whipflash effect. After 2007 area rebounded for 2 years with 2009 being pretty high. If the climate/weather is getting more erratic than we could see a faster/larger/shorter duration bounce back from 2012 record melt.

2. Dispite all of the evidence of how weak and thin the ice is, it has shown a remarkable resiliance in hanging in there so far this spring.

3. There has been enough talk of colapse (not all here), that I'm predicting there will be a colapse of the colapse theories, before the colapse actually comes.

I suggest it might be human nature that if something is going to be bad. It has to be really bad, ultimate bad, at ever faster rates. To which nature can decide to change her mind, throw us a curve ball, and confound us all.


I suggest it might be human nature that if something is going to be bad. It has to be really bad, ultimate bad, at ever faster rates. To which nature can decide to change her mind, throw us a curve ball, and confound us all.

Couldn't agree more opensheart! As David Attenborough once said : "In the end , the planet will always look after itself". The problem is that humans (at least human plans) may not be part of Earth's plans.

Dr Tskoul



Anthony Watts has put up a poll for the SEARCH SIO, but most votes so far have gone to ' Greater than 6.0 Million km2'. Are fellow alarmists skewing the vote there, or do they really think the slow start to the melting season will cause the mother of all recoveries? With ' 5.0 to 5.1 Million km2' in second spot (also quite unreasonable IMHO) it doesn't look like pranksters at work. Besides, there are much less CAGW-folks visiting WUWT than before.

Either way, looks like Watts won't be sending in the bin with the most votes (like he did last year), but some sort of average or median.

If it stays like this I might turn it into a blog post...

L. Hamilton

Just to be clear, I'm not planning to send our results to SIO this month, viewing it as a trial period -- unless you folks tell me otherwise.

Next month's post will be more clear about Sep mean in the title, and I think we should send those results to SIO. I'll also soon post an analysis of the June results, which have about 113 individual estimates (median still 3.0).

L. Hamilton

BTW, one well-known sea ice researcher suggested to me that if we averaged our Sea Ice blog value with that from WUWT, it would be a mean between pessimists and optimists, and closer to most of the science teams.


I said the same thing last year, also because the WUWT forecast wasn't completely out there. But if 'Greater than 6.0 Million km2' becomes top of the vote, and Watts sends that in...

Jai Mitchell

Interesting, WUWT has been coaching his voters because of previous over estimations. http://wattsupwiththat.com/2013/06/06/sea-ice-news-volume-4-2-the-2013-sea-ice-forecast-contest/

"Bear in mind that traditionally, forecasts in June have been too high. Last year’s minimum was 3.41 million square kilometers (1.32 million square miles) at its lowest point on 16 September, and in June, WUWT readers forecast 4.9 million sq kilometers."

L. Hamilton

I'm making a point of keeping my own thumb off the scales here! But I will publish some analysis soon to give y'all some feedback. That will include comparison with the distribution of predictions by SIO science teams, and a preview of the analytical strategy.

michael sweet

L. Hamilton:

I think you should send in the numbers this month. Until it's in writing it's not a prediction. We want to document our choices versus the deniers.

Just my$.02.

Kevin O'Neill

I would agree with Michael. The numbers before and after the reminders (on May 8th) that we're looking for monthly average - not minimum extent - might be interesting to note in and of themselves. As would the predictions vs time - i.e., did people with more information (voting later in the month) have significantly different views than those that voted early?

I think it valuable that the numbers be published in SEARCH just so that others can see them.


If the deniers' forecast is ridicilously high, they look silly. It's a "publicity win" for those who care about the climate. If our forecasts are too dramatic and we see a bounce back in the minimum, we look like we have egg all over our faces and they (the deniers) brag "we were right all along".

Hence, my belief that we should be cautious in our predictions. The ice will be what it will be, nothing we say ain't gonna change it, but public perception is important because it ulmitalely influences politicians and climate action. The main argument I hear from deniers around me is that we are "alarmists" therefore all we are saying is a load of crock, and no doubt we hear some predictions that are needlessly alarmist. IMHO the best response to this sort of argument is to show that we are rigorous, do not get carried away and that we are merely stating scientific facts.

Doug Lofland

1.8 +/- 0.2 mkm2 based on a huge spreadsheet that reverse engineered the melt based on energy that will be in excess from the state change. Also consulted with a psyhic on Long island


right now there is a virtually perfect correlation between estimated extent and bin popularity (the higher the estimate, the more people like it) - watt is up with that, indeed

Dan Ellis-Jones

Yes, I agree with you on being cautious.

I've been an avid reader of this blog for a few months now, and I'm a little concerned that there is an element of (totally unintended) 'Doom Porn'. I think the effect of looking at the melt season in the minutiae that this blog does can seem like we/you are almost urging the Arctic to melt out. I absolutely believe that this isn't the case for anyone who reads or posts here.

I think it might be worth mentioning from time to time that we WANT to see the Arctic ice to bounce back. The reason this blog looks at each and every development is that there is an underlying reality that IF we are seeing a melt out of Arctic ice, the consequences are dire. The better and earlier we are informed, the greater our chance to adapt to a new climatic world.

Frankly, I want the deniers to be right, but I can't ignore the science and the observations that point to dangerous changes to our climate, and thus our natural environment which is the basis of human existence (including the economy) on earth. Sadly it looks like the deniers are wrong.

Dan Ellis-Jones

As for my prediction... after reading comments, looking at weather patterns, witnessing the more average start of the melt season but taking into account the 'slushie' conditions of the ice...

Sept average will be 4.4m km2
September low will be 4.1m km2

But my instinct is that it will counfound us all, but I'm not sure if that's by much less or much more ice loss than most expect.

L. Hamilton

Some might imagine a battle of the blogs here, but that's not where I'd like to take it. As described in the OP we've started an experiment, with most of the data yet to come. And the analysis, and the writing.

Some might imagine a battle of the blogs here, but that's not where I'd like to take it.

Me neither, but I found this year's poll on WUWT fascinating. 'Greater than 6.0 Million km2' is going to win the vote, and now I'm wondering if Watts is going to send that in...


My original was 1.8 but that was daily. I didn't bother revising it but I'll say 2.2 monthly.

Very sloppy guess. But I don't this 'slow start' as meaningful, especially given the rationales that were given previously.



Thanks to Bfraser, Lodger , Neven & FrankD, I have come to the conclusion, that I’ll need to revise my previous estimate of 1.1 mio. sq. km. First of all, I was basing my estimate on minimum daily volume (+6 %), then I was calculating area and not extent (+15 %), and finally, I was estimating a new minimum to occur around 1 Oct (+4 %). So to be fair, I would like to raise my initial number from 1.1 mio. sq. km to 1.3 mio. sq. km. Hope you will be able to accommodate this revised estimate just before the deadline.

All the best


PS The early snow melt in May, the two Arctic cyclones and the ECMWF and NCEP forecasts all point in the same direction.


I'm wondering if Watts is going to send that in...

I guess not. He has just closed the poll with the following announcement:

Thanks to everyone who voted, and especially thanks to all those trolls (and we both know who you are) who tried to skew the poll by voting for “6.0 or greater”, which happens every time. Got you covered though, since I anticipate such actions.

If true, I think this is the first time it happened that trolls tried to skew the poll. Like I said, I don't think that there are that many warmistas going to WUWT; I only go there to see what Watts has to say about Arctic sea ice, and what he doesn't say, of course. I'm not sure, but I think I only voted on his first poll, and voted truthfully, I believe. They're crazy and ignorant enough as it is, why would I want to make it look worse than it is? Arctic sea ice is a huge problem for fake skeptics.

But it might be true that warmistas went and voted 'Greater than 6.0 Million km2'. Maybe Lew and John Cook are behind it. ;-)

I wonder what Watts will send in now. Not that I'll lose any sleep over it.



please don't go there again. It's a waste of your precious time. Get some sleep and get your vegetables growing instead.

Cheers P

L. Hamilton

OK, the SEARCH deadline has come and this June poll is now officially closed. Thanks to everyone who contributed; by my count we've got 116 predictions. I hope there will be equal enthusiasm for the July and August cycles, still to come.

Our June median prediction is 3.0, possibly biased downward by my unclear title which caused some people to predict 1-day min instead of monthly. The July post will be better in that respect. In the meantime, I'll have another post soon to summarize the June results, and explain more about where this experiment might be heading.

Stem-and-leaf plot for asi_blog
plot in units of .1

0* | 000
0. |
1* | 033
1. | 5788
2* | 000000122223444
2. | 555555677888888888899999999
3* | 000000000011111122223333344
3. | 5566777888888888999
4* | 0000111123344
4. | 5
5* | 0
5. | 57
6* |
6. | 6

L. Hamilton

Neven, I think the spam filter just gobbled my wrap-up post.


Released now, Larry. Thanks for this first crowd-source prediction post.

L. Hamilton

As might be expected, the ASI median estimate came in substantially lower than the SEARCH SIO median. I'll write a post comparing the two distributions over the next couple of days, setting up an analysis that should continue through the season.

The comments to this entry are closed.