« 2011 Slush Puppie Animation | Main | Poll Results 2 »


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


The time the factors were warm enough to start the melting process it would never stop unless the factors of melting the ice reversed to factors of stopping the melt or reversing it.

However lets say those factors stopped getting stronger shortly after the ice started to melt

the ice would still melt, it might take 200 years but eventually it would melt out in the summer.

Co2 and climate variation like the PDO, NAO, AMO could all have equal partner ships.

but two things have started the rapid feedback.

While the initial issues helped potentiate the newer factors which has lead to rapid feedback.

1960s: C02 has started to go up at an exponentially increasing rate.

1970s: Co2 is starting to go to work on the Earth. At this time the overall Pattern in the NH was condusive for colder temps. The Co2 Feedback was far to weak attm to make a dramatic impact on it's own accord Wind Patterns were non effective to this ice pack for long term ice loss. Temps were not an issue.

19800s: Co2 with wind patterns started to take a toll on the arctic ice. With this shift warmer summers started to emerge up there. Ice slowly started to lose Volume.

1990s: The Beaufort Gyre broke down. Co2 is now increasing faster then ever. The Sun is revving up. The ice melt is feed-backing on itself now, albeit slowly. Warmer waters are filtering under the ice and ice free parts of the arctic are warming enough for albedo melt. The volume is slowly melted away each summer and not being replenished. by the late 90s the train leaves the tracks.

2000s: Co2 is now taking a toll on the entire globe. However Ice melt is steadily increasing. It is not until 2007 when the pattern overwhelms the weakened state of the ice and opens up 1/3rd of the arctic to water. The water is warmed dramatically and the ice thickness crippled. This now allows SST feedback to run amok.

2008-10 the winds are favorable for the ice to recover. But the SST feedback and the loss of volume are to much and the ice still losses volume during a calm period.

2011- Winds are still decently favorable but the Co2, 2m, and SSTs have caused overwhelming feedback that can't be stopped and the ice melts out faster then ever inside its own arctic ocean.



Unless there is dramatic climate shift to much colder weather.4-6C in the arctic alone, the ice will melt out. The feedback is out of control.

We will just need one summer of horrible winds and warm temps and the ice will be nothing but a 2-3km2 slab sitting next to Greenland.

that drop will cause the winter ice to never recover fully allowing much of the arctic away from that ice to be to warm and to much cold will be lost trying to cool that water.

This is clear to anyone with there eyes open.

I am sorry there is nothing we can do now.

We all think once ice free hits the Ecosystem will feedback on itself to raise global temps. This will change everything.

the clock is ticking...


This place will start filling with trolls the worse it gets in the arctic.

many of them have deep psychological ties to denying the reality of this.

That reality is being shattered, there will be repercussions.

Bob Wallace

Thanks for all that Paul. Helps me a lot.

I don't know how accurate the ARC drift projections tend to be, but they are predicting about three days of hard movement into the Fram. Since, as you point out, the chute is open we could see a lot of ice lost there in the next week.


Charles Monnett is back at Ak Region BOEMRE tomorrow:



Posted by: Neven | August 25, 2011 at 20:13
Anu, I believe the ice retreated quite far over there very early in the season, as air temps were very high and there was lots of insolation as well. I could be wrong, but this seems to be in large part albedo-induced warmth.

I suppose there was significant melt/spread as early as August 10th:

Would high air temps and cloudless conditions persist in one small location for two weeks? And this condition happened only in this one small region? Are you suggesting that once the melt started, the open waters absorbed more sunlight and it became a small regional self-sustaining melt ? I suppose the melt area could be the result of strictly surface features, I wasn't really paying attention to weather maps and cloud cover in this area in August.

It still seems significant to me that this most-northerly region of unusual melt sits right over the Lomonosov and Gakkel Ridges:
    The ... Lomonosov Ridge ... rises 3,300 to 3,700 m above the seabed. Slopes of the ridge are relatively steep, broken up by canyons, and covered with layers of silt.

I think the "ocean" part of the Arctic Ocean sea ice melt is sometimes overlooked.


Related to the fast ice in northeastern Greenland; cloud cover has prevented clear MODIS views for a number of days now.

These overlapping radar images from ASAR (http://envisat.esa.int/instruments/asar/)

show some fracturing of this residual fast ice.




Checking my post, I see the images are not loading.

I found the radar images of Northeast Greenland here; http://ocean.dmi.dk/arctic/northeastwater.uk.php


Okay, that link appears to work. To find ASAR images to view the ice below the clouds select a Greenland region of interest and a date to be viewed. Then find the ASAR images.


Getting to the bottom line, ASAR images (from a few days ago) do show some fracturing of the remaining fast ice in northeastern Greenland, but, so far, most of this ice is hanging on.


-38K for tonight's prelim.




Guys, please find out if you can asap when those maps are made?

how can that be?


Frivolous, that first link is to that Russian research station. That's definitely not where the Healy is.


I know that.

The map self explains where the Healy is.

And it is pretty far into the ice pack for showing so little ice the last few image updates.


The Healy has shown almost no ice concentration the last 6 hours except for one image.

and it has been right over 80-90% concentration during that time.


I see what you mean. There were similar situations last year.

Speculation: Right now it would be in a slightly darker patch, I would say 50-75% concentration. A grid cell (how big was it again, 6.25 km x 6.25 km?) with 25% ice gone could easily contain an area of 1 km x 1 km. How far can we see with that webcam?

And of course, it's misty now, unfortunately.

Artful Dodger

Neven said: "Let the ice do the talking."

Ho! respect, brah! (I support your position unconditionally).


Bremen Extent has taken another big drop. Moving further below 2007.

Expect Jaxa to drop below 5,000,000km tomorrow.


"The uncertainty makes it worse, not better. If you find uncertainty in this context reassuring, I'm happy for you.



I'm not sure where you got the idea that I think the uncertainty makes it better. It does make it worse. My point is just a suggestion that you might consider or not. In some measure you speak for the science I believe in. And since we both care about the planet I think reasoning together about the best choice of words is a small favor to ask given what's at stake. However, if you remain convinced and absolutely certain that the greater purpose is served by your take on things, then I'll keep those suggestions to myself. But I do find it odd that with so much at stake that people are so certain that their way of communicating is best. But that's understandable I suppose.

As I said, I think this is a nice place and I enjoy reading it. You've clarified what you mean by ice free. It wasn't that painful. I've explained why that's important, it wasn't that difficult. Not much more to say, just waiting for this year to beat 2007. Peace and love.

Rob Dekker

Peace and love to you as well Steve.
Thank you for your statement :
"But unless some positive feedback or strong natural variability kicks in, an ice-free summer around 2030 is virtually impossible."

Just so that we are as clear on what you mean as you are asking Neven to be :
What is your definition of 'ice-free' in this statement ?


Probably when they close the last ice cream shop in northern Greenland!


Hi all,

The Passages that run North South in the Canadian Archipelago no loner contain any sheet ice that would prevent movement. Over the last few days, ice has been gently eddying Northwards.

The Sea Surface Salinity map on the same US Navy page as the new PIPS displacement maps continues to intrigue me.

It has just introduced a change of chromatic scale which confuses the eye, but, running the 30 day gif animation, it does seem that there has been a large ingress of more saline water across the Arctic, flowing from East of Svarlsbard over towards the Bering Strait. This is very close to the area of unexpected high melt near 85N.

Neven - it may possibly be illegal for you to allow people to bet real money using this site as a forum, unless you have a gaming licence. I think you should ban it.

Rob Dekker

Idunno, Neven will not be allowed to BROKER a deal without a gaming license, but AFAIK there talking about bets is part of freedom of speach. Either way, it happens all the time on internet, so Neven would be in a very, very large company.

Seke Rob

Short: My fav Al Pacino movie after 'Scent of a Woman' is 'devils advocate'. Seeing SM referring to the 10K 'weasels', this gentlemen goes hooded as their advocate.

The Glaziers will be feeding from different outlets for millenniums to come, so even in heart of summer the will be ice, so to clone hairs, there'd need to be a definition of 'what ice source'.

Neven, let's talk the state of the Arctic, with a little lite metaphoric injection. We are smart enough to filter the material information.

My model has the 5 million now slip into the 27th, but only because it computes 1720 kmsq above that threshold for the 26th... it's statistics... but if taken over 30+ years, it's dire.

Today was confronted with the statement that between 2007 and 2011 we had 3 good years in between (only and solely recognizing the JAXA chart). He dons special eyeware, incapable to see what I presented him in charts, same as I've done here. We need sledgehammers to get through, but then 21-12-2012 was trotted out as it not mattering anyway... so that's a waste of time to even argue.

Whilst we hope Hurricane Irene slips off eastward, it's heading in amongst for Long Island. They're statistically overdue to have one per a video presentation by Prof.Scott Mandia.


Could someone please make a manipulated Modis-image and place the remains of the Ward Hunt ice shelf, I cant figure out where it is located, if it still exist? And link it to this blog, Thanks Espen


Hi Espen,

here you can see the location of the Ward Hunt ic shelf:


Chris K


At risk of being tangental to Daniel Bailey's tangent (only 24 hours ago, but already already a page back!) but...

Yooper, you probably shouldn't still use that version of that graph. It was my version 1.0 of a crude attempt to replicate the results of PIOMAS version 1.0. We now have improved versions of the actual output from PIOMAS 2.0.

Wipneus has been doing monthly updates here as PSC have released results for each new month. Much higher quality than my original Excel-o-gram. The last I can see is here: http://neven1.typepad.com/blog/2011/08/piomas-july-2011.html#more

There might have been once since, but I couldn't find it - "They're coming in too fast!"

Pete Dunkelberg

Did Prof Mandia say "statistically overdue?" I suggest that this phrase be reconsidered. Think about it, or look up "exponential waiting time."

Daniel Bailey

Sorry, FrankD. Dealing with sophistry and media manipulation by the Grima Wormtongues of the world pi**es me off unduly. In my "haste to lay waste" I grabbed the first link in my Arctic Sea Ice graphics folder. Thanks for the correction!

Let the general reader beware: The climate blogosphere is infested by those who seek to dissemble and confuse, as they follow agendas laid forth, not the science itself. Sweeping statements lacking proper context and unsupported with links to reputable, primary sources are to be considered untrustworthy until shown otherwise.

Caveat emptor, my friends.


idunno, I don't mind friendly bets that involve small amounts of money or quatloos. I bet myself with William Connolley.

Neven, let's talk the state of the Arctic,

Good idea! We have plenty to look at these last 2-3 weeks.

Seke Rob

Re: Pete Dunkelberg | August 26, 2011 at 14:30

"3 years overdue" was the description. Here the video I took that away from: http://profmandia.wordpress.com/2011/08/10/long-island-hurricanes-global-warming/


BTW, will start posting again tomorrow (need a day to recover from a week of deadlines and flash melting).

Ned Ward

I bet myself with William Connolley.

Now that's what I'd call a high-stakes bet!


Hi all,

Cryosphere Today has fallen to 2.98 million km squared, which is:

Lower than every minimum except 2007.

Lower than 2007 for this date.

2011 and 2007 the only years to break through the 3 million mark.


The remenants of Irene may hit Greenland with tropical storm strength winds.

Not sure of the consaquencies but it may have enough residual heat to mean a lot of rain rather than snow there?


CT Area: Now 11 days ahead of 2007 and one more day like the last reported drop (78k)would give a record low (68k needed).

2007 had its minimum in 13 further data days time. Last four years have had area minimum consistently in 13 to 15 further data days, but 2006 didn't have minimum for a further 28 data days.

Wayne Kernochan

Re CT area: It is now only 60K from 2007. I continue to predict that it will wind up between 2.8 and 2.5 mkm2, and that with a 61 or 62% CAPIE that would mean extent would also most likely go below 4.4.

Wayne Kernochan

@dorlomin: The latest forecast I have seen gives only 30% chance of tropical storm winds to Maine. Since after that it will need to push across some land, and the waters after that aren't as warm, I think it's unlikely that it will reach Greenland with winds that strong.

However, it's more likely that it will run up between Canada and Greenland with warmer air and higher winds than are there right now. I am wondering if this will push on into the Arctic along that line and affect melt -- e.g., add a couple of extra days of above zero temps at 80 degrees N to extend ice surface heating.

Seke Rob

Wonder if L.Hamilton could get the minimum dates to print vertically in the 2 bar charts, or print them on the X Axis in place of the years? Could even be done with a secondary Y axis, and print them as day # of year line graph on top of the bars for an instant feel of earlier or later.

William Crump


Wow your site has really taken off.

Thanks for the response. I saw the .9 meters measurement for the North Pole from an item you posted and was wondering how it compared to the past.

Are these measurements for a single location or are they an average of thickness samples taken over a wide portion of the Arctic Basin?

I like the comparison you did, but I would like it to be made with more data points.

I would guess there are areas in the Arctic Basin that had ice last year that have no ice this year and areas which have ice this year which did not have ice last year. This makes extrapolation for a single point (or a relatively small area) unreliable.

Does the data you are citing include broadly dispersed sampling points or is it just North Pole data?

Anyone want to guess when the Central Arctic Basin on MAISE or the Arctic Basin Region on Cryosphere Today will start increasing on a regular basis?

I guess that the Cryosphere Today Arctic Basin chart for ice area will top 3 million km2 around September 30, and the ice extent chart for the Central Arctic Basin on MAISE will not go below 3 million in 2011 even though it is below 2007.

I do not see that there is data base for ice volume, thickness, area or extent for the Arctic Basin Region which indicates it will be "ice free" by 2013 or 2016.


Hi William,

The thickness measurement of 0.9 metres was made by the Alfred Wegener expedition which has just been through the North Pole. It was the most commonly found measurement, approaching the Pole from a sector that has noticeably thin ice. It is not an average, or mean, measurement, it appears, but a median.

I must say that I find your whole take on the recent fall of the ice in the Arctic Basin to a record new low is, while in keeping with your pet theory, somewhat drifting away from the data.

I would expect that Central Basin ice will start increasing in area in a few weeks time, after it has finished setting a record, slightly later than usual, as the Sea Surface temperature measurements all around it are very high, and there seems to be anomalously high Sea Surface Salinity building up underneath the ice.


Hi William, that's all the info I got out of that AWI press release. There's an article on ScienceDaily, I believe, with a better translation (but comes down to the same thing).

Wow, Cryosphere Today SIA! Like I thought, the flash melting has come with a small lag. I'm doing a post on that tonight.


Now that's what I'd call a high-stakes bet!

That's because you don't know me, Ned! :-P

Lord Soth

I not sure if 2011 will break the 5M barrier tonight, but it should definitely do it by the 27th.

There is tremendous potential for compression, so once the weather patterns become favorable, we should see some nice extent reductions.

The real action will be in 2012. With the transpolar drift, we should see first year ice over the pole for the start of the 2012 melt season, just like it did in 2008.

If weather conditions are favorable, 2012 will give us the possibility of an ice free pole.

Wayne Kernochan

On another note, if I'm reading the data correctly, the record negative area anomaly for any time of the year was 2007 (early in August) at -2.076, and 2011 is only about 0.02 away. We've already passed last year's -2.03.


Not exactly, we are well short of -2.635 anomaly:

2007.8000 -2.6349814 4.4081578 7.0431390

That is much later in the year though.


Wasn't that around the time when the global sea ice area anomaly also hit. This year was close to breaking that one a couple of times, but it's also highly dependent on the Antarctic SIA.


Thanks Chris

For the Ward Hunt image, I also suspected that little piece of ice to be the one, but I thought it was too small to be given a name!

Regards Espen

Wayne Kernochan

Egg on my face. Thanks, crandles. Could we maybe say "record for any time of the year before Sept. 28"?

Re global sea ice anomaly: I think we're past the time of year when we could break that one. It's unlikely either Arctic or Antarctic area anomaly will drop 450K at this point, although I suppose 2011 Arctic could pull a 2007 in October ...

Seke Rob

Hyperactive, Thomas Dolby... go look it up on Youtube :D

MASIE dropped 51K from the 24th to the 25th. The spaghetti chart shows the Central Arctic pushing that 3M barrier, but it's shied away a few times now.

MASIE BTW is closing in again on JAXA... 161KKmSq over latter.

The actual average per CT through Aug.24
236 days 2007: 10,022,447 KmSq
236 days 2011: 9,949,049 KmSq
Variance: -73,398 KmSq.

Is 2007 still the year to beat through the point where the annual minimum is reached? Marginal statistics to me, but if those can round out, there will be little left to argue. The debate has already shifted to the ACE value for the Atlantic it the WUWT wurreld, though just reading at Ryan Maue's site that the NH ACE as at August 17 was 4 days ahead for the 30 year average at 200. Irene will add a bunch.

Seke Rob

Re: Wayne Kernochan | August 26, 2011 at 19:51

CT Global SIA in a picture.

Yes, that absolute record chance has passed, but the cumulations, see my post above, are well ahead. There's a 365 days rolling average left to beat that 2007 had in November. We're 33KKmSq away from that number. If you look at 2010, you see that there could be a second wind.


It's unlikely either Arctic or Antarctic area anomaly will drop 450K at this point, although I suppose 2011 Arctic could pull a 2007 in October ...

Wayne, there's no need for SIA to drop more for the anomaly to be greater.

In fact, chances are best later in the season because during the 1979-2008 baseline period the seas would quickly freeze over at the end of the melting season, but now it takes longer due to warmer waters, which means the anomaly can increase very fast. If you catch my counterintuitive drift.

Speaking of sea ice area, there's a new post up: Area graphs: 3 million km2 mark passed

Wayne Kernochan

@Neven: I think we are actually in massive agreement, with the caveat that last year around this point I think Antarctic area starting going significantly positive. Thank you for summarizing far better than I could.

And before I forget yet again, thank you again for creating and maintaining this amazing site, from which I have learned so much. Please know that you are welcome to shut me up any time I interfere, and that I really hope you'll keep on keeping on. - w


I think we are actually in massive agreement

Ah yeah, on closer inspection I see what you mean. I guess I was talking to myself. :-)

Don't thank me yet. We still have 2-3 weeks to go.

PS Fram Strait animation has been updated.

Paul Klemencic

I don't have a lot of time today, but thought I would quickly make two comments. First, I want to address confusion between the MASIE reported regional results, contrasted to the Bremen map, then compare with the IJIS reported SIE. If this all been discussed before, I apologize.

I have been confused why the MASIE reported E. Siberian ice extent remained above 600k, while the Bremen map clearly shows significant extent loss, and the ice shown on the Bremen map below 80N clearly doesn't have an extent of 600k sq km.

It turns out, that the E. Siberian region, seems to extend up almost to 82N in places. Unfortunately, the MASIE site map for the E. Siberian doesn't label the latitude parallels shown, but they appear to be the 71N, 76N and 81N parallels. So much of the main ice pack is included in the E. Siberian region report.

Secondly, the ice extent is shown in white on the linked page, and again the ice extent is much larger than shown on the Bremen map. Clearly the MASIE system shows ice extent over 15% in a significant amount of area not shown as that concentration on the Bremen map. The possibility exists, that the MASIE report has a delay, or that the MASIE system is more sensitive, but in the E. Siberian region alone, the difference appears to be over 150k sq km. You can get an idea by following the 75N parallel, which bisects the easternmost island in the New Siberian Islands, or the 76N parallel; across both the Bremen map and across the MASIE regional map. There is a huge difference in ice extents shown.

Next, the total extent from the MASIE regional maps is about 5.20 million, about 165k higher than the IJIS reported number, but still relatively close. This means the MASIE maps are likely closer to the extents reported by IJIS, than either is to the Bremen map that I follow.

No wonder we can't seem to estimate daily extent losses by watching the Bremen map! In most of the critical regions, the Bremen ice pack edge varies significantly from the MASIE ice pack edge (with the MASIE edge much further out), and the IJIS reported extent numbers seem closer to MASIE than the Bremen map.

Clearly the current Bremen map doesn't show over 5.00 million sq. km. Now what is surprising to me; the Bremen map also seems to overstate ice extent; this conclusion reached from examining the Healy photos from seas that should be filled with 70-90+% ice extent the last several days (yet show very little ice).

I am increasingly skeptical of the accuracy of the IJIS and MASIE reports. Either there is a serious delay, or they clearly overstate extent. Even the Bremen map seems to overstate extent.
If someone can explain all this, please help out.

Wayne Kernochan

"I was talking to myself" -- to quote Tolkien: "It's a habit of the old. They like to talk to the wisest person in the room."

Patrice Monroe Pustavrh

Paul, I have to slightly disagree with you about complaining about extent measurements. Yes, there must be much of open water in some cell (6.25x6.25 km2 or 12.5 x 12.5 km2 or even 25 x 25 km2). But, things are measured consistently over the years and similar errors in measurement occurred in past years. And extent is quite a variable measure (you can have same amount of ice and you can measure it in one or six cells, so extent with given area can theoretically vary for more than factor 6 given same area covered just by distribution). So my point is to: let us accept the extent measure as is. It will have errors, but those errors will distribute across the years and show us general trend. And yes, we shall account for some errors in measurement. Otherwise, we are no better than deniers, which do try to overthrow the very basic measurements and we end up in really non scientific way. If any issues are raised against measurements, they should be raised on much more sound observations than just couple of photos from icebreaker. P.S. I do trust scientists that Arctic ice is in bad condition and that we ought to do something to cut our CO2 emissions. But, we should be very very conservative, when we ask the scientific observations and should take them as reliable unless we are not knowing better.

Paul Klemencic

Or look at the MASIE map for the Chukchi where the ice edge isn't anywhere close to the Bremen map. In fact, the MASIE map seems to be about 4-5 days old; thats were the ice edge was at that time.

We generally seem to see good correlation between the Bremen maps and the MODIS images (before the Healy photos); so I tend to believe the Bremen map. The MASIE (and likely the IJIS) seem to be have a time delay of about five days; i.e. they show the ice where it was five days ago.



this conclusion reached from examining the Healy photos from seas that should be filled with 70-90+% ice extent the last several days (yet show very little ice).

Paul, I discussed this briefly with Frivolousz21 last night (and also touched upon it last year): "Right now it would be in a slightly darker patch, I would say 50-75% concentration. A grid cell (how big was it again, 6.25 km x 6.25 km?) with 25% ice gone could easily contain an area of 1 km x 1 km [with hardly any ice]. How far can we see with that webcam?"

Last night's images were very misty, but they are much clearer now and showing much more ice around.

What I'm trying to say: Healy webcam images are a lovely, but very limited indication.

Bob Wallace

The Healy has moved from 80N to 82N since yesterday. What's showing now is not the same location as last night.

Paul Klemencic

I am not questioning the scientists, I am questioning my own understanding of how to interpret these reports. The Bremen map conflicts with the MASIE maps, and not just a little, but by a large amount. The Bremen map seems to agree with MODIS. The MASIE extent measurement seems to agree with IJIS. Yet the two sets (MODIS and Bremen map) vs. (MASIE regional extent map and IJIS reports) disagree by several hundred km on where the edge of the ice pack is located.

I just want to understand this.

Patrice Monroe Pustavrh

Paul, it may not be so confusing, but different filtering alghoritms, 5 (NSIDC) or 2 (IJIS) smoothing do make a difference on short time scales. But nevertheless, I am asking myself a question: Given the ice is much more scattered at larger areas than in previous years (and there is big if with this statement, so treat it as a question), could it be, that in the following years the measurement of area, despite known issues would be more accurate indicator of state of ice than extent ? I am very sure, it was not so in the past, but what about seeing large areas of clearly not 100% areas of ice as viewed from MODIS ? For the trends in long term it is clear that any measure we take, it will reflect the state of ice, but can it happen that variability in ice distribution will cause greater error than our inability to separate melt water ponds and wet ice from open water ? If someone from the field informs us on this issue, the help would be very well appreciated.

Paul Klemencic

Neven, I am not trying to bring the Healy photos into this right now. The Healy is now above 81deg 40"N and into the very solid pack. Lets simply compare the MASIE map ice edge in the Chukchi with the Bremen map.

The MASIE map shows mostly high concentration ice pack extending down to the 75N parallel, with some ice extending down to 73N. Yesterday's Bremen map doesn't show any significant ice extent below 76N and a lot of open water extending up over 77N. I went back five days, and the Bremen map back then still didn't show the ice as low as the MASIE map shows.

The difference in extent is quite large. At least 100k sq km out of a total of 287k shown for the Chukchi region by MASIE. Thats a pretty big error, and hard to miss.



Bremen and IJIS use the same data.



Take a closer look.

They are nearly identical. The only divergence comes when the pack has a lot of open water that a 6.25km vs 12.5km grid resolution would pick up.

Jaxa uses a 2 day running average where Bremen is day to day.

Jaxa plays catch up to create more smooth data. While UB will have a big drop then an even day then another drop. Jaxa would show two slower drop days but it evens out.

I would expect Jaxa to have a little bigger drop day the next 3-4 days. those arms of ice in the Siberian Sea and in the Kara/Lapteve play havoc on higher grid resolutions.



is even further behind because open water like between the two arm bands in the ESB cause those large grids to have to much ice and there will be to many grids with 20-30% concentration adding there entire grid to the data.

As you can image Jaxa would have this issue over UB. Today the UB map showed a huge opening in that hole up there, it also showed a big recede between 135E and 105E.

Compaction over that area that data is sparse until the final map as well.

this will give Jaxa more ice loss then Norsex and UB will will have the most. The faster those arms compact/melt out the fast Jaxa catches up


From my studies of the polar sea, I find the Modis Images and the Bremen Map the most reliable and the ASAR Images should I add, they all give a good base for understanding what is going on, much other material I find doubtful.

Regards Espen

Paul Klemencic

OK, I have looked at the MASIE charts and calculated ice extents for the E. Siberian, the Chukchi, and the Beaufort. In every case, the charts show where the ice was over a week ago (on the Bremen map). Each regional ice extent shown is consistent and does agree with the ice pack edge shown in the charts. The extents shown do NOT agree, not even close, with reasonable estimates of regional extents using where the ice boundary is in yesterday's Bremen map.

I can conclude that the extent data shown on the MASIE regional maps is over a week old! The total of those extents, show the total MASIE of over 5.2 million sq km. That seems an accurate estimate of the sea ice extent for LAST WEEK from the MASIE reports.

There is no way to reconcile the MASIE data with the Bremen map information. There almost certainly is a problem with the MASIE reports.

Kevin O'Neill

As has been noted, the Healy's webcam offers a limited field of vision; open water may be very local - and the captain probably alters his course to take advantage of any openings between floes, leads, etc. That said, some of those images definitely raise questions about the ice extent algorithms.

But, as Patrice pointed out, the same algorithms were used last year, and the year before, and the .... you get the point. Yes, these measurements have errors and uncertainties, but in annual comparisons we need to compare apples to apples.

What perhaps is frustrating is that we 'see' how vulnerable the ice is, but the numbers we've always relied upon seem insufficient to quantify the demise of the ice pack. Because of the way MYI is calculated it's possible we'll see numbers that say the ice is still 'recovering' - when any astute observer can tell you that just ain't so.

This is why area and extent are poor proxies for the health of the ice pack. PIOMAS has been telling us the ice is getting thinner, concentrations lower, resulting in lower volumes even during periods of 'recovery'. We're seeing a perfect example of that.

Paul Klemencic

More confirmation of week old MASIE extent maps used to calculate regional extent data:

The E. Siberian regional chart (for August 25th) shows a pod of ice about to separate from the pack near the New Siberian islands.

That pod separated a week ago, and now only the shell of that pod remains in the current ice map. The MASIE charts, and the calculated ice extents shown, are a week old.

Paul Klemencic

Wow, what an edge this would give my betting game. I can predict the future! What day was the flash melt? August 22nd, correct?

Well, I predict that by August 31, we will see a one day drop of over 120k sq km in the total of the Chukchi, and Beaufort regional ice extents reported by MASIE. I think are likely to see this drop by August 29, but I am giving myself a cushion.

If only I had this info when Lucia was around; I could have won some quatloos!

Rob Dekker

I think Lucia is a bit busy debugging her analysis of the Connolley-Dekker 2016 sea ice bet using her models :


Steve Bloom

Paul, it should go without saying that people overly focused on quatloos care more about the Rules of Acquisition than they do science. :)

(Yes, I'm well aware of the thoroughly mixed Star Trek metaphors.)

Paul Klemencic

Golly guys, I am going to look like an idiot if I am wrong on this, but here goes:

Frivolousz21, thanks for this information:

Bremen and IJIS use the same data.


Take a closer look.
They are nearly identical. The only divergence comes when the pack has a lot of open water that a 6.25km vs 12.5km grid resolution would pick up.

Jaxa uses a 2 day running average where Bremen is day to day.

Jaxa plays catch up to create more smooth data. While UB will have a big drop then an even day then another drop. Jaxa would show two slower drop days but it evens out.

The lag appears to be at least 4-5 days, and possibly with two day averaging, up to 6-7 days.... HMMM, a similar lag that is in the MASIE data.

If MASIE showed over 5.2 million a week ago, I estimate we have lost at about 420k since then (with the flash melt etc.), then the current up-to-date ice extent would be around 4.8 million. And close examination of the Bremen chart of ice extent shows about 4.7-4.8 million, but hard to read.

Can we get an accurate numerical value for the Bremen extent?
Could you point me where to find it?

It seems that the Bremen leads the JAXA IJIS by 5-7 days. To me, the Bremen report predicts what the MASIE report will show in seven days, and leads MASIE as well.

We should be looking at the leading indicator, and that is the Bremen ice extent report. Interestingly the NSIDC ice extent seems to track the Bremen ice extent report. The JAXA report is the outlier, and the laggard among the three.

It seems that the real extent that matches the ice extent maps right NOW, is only 4.8 million sq km. And when NSIDC reports for August 26th, their report will be close to that number. And in about a week, MASIE will report that number (4.8 million), down from the 5.2 million they report now, when their lagged information comes in.

We need an up to date metric, to compare with real world observations. No wonder we are having trouble matching ice pack observations with the reported JAXA numbers and the MASIE regional ice extents. They both have big lags.

Yvan Dutil

Paul, there is an easy way to check this out. All we have to do is calculate the cross-correlation between each data set. Time lag could be easily determined. My general impression, is that over a year,number will average out and the residual lag will be very small.

If someone can point me to the right data source, I could do the analysis very rapidly.


@idunno: "... somewhat drifting away from the data."

Indeed. Cryosphere Today's graph for the Arctic Basin is ~190,000 km^2 down on 2010, a 7.5% delta. The data does not support the claim that there is "no significant difference" between the two years.


Wayne K, crandles etc (not trying to teach you to suck eggs, just expanding the point): The maximum anomaly is driven more by timing than anything else. When ice extent is increasing by up to 200K per day (in Oct-Nov), a small delay translates to a big change in anomaly. So we see the maximum local anomalies where the peak rate of melt / freeze is a little earlier or later than usual. A million K anomaly in June or November might translate to 5 or 10 days change. But because an extra 5 days at the rate of change we see at mx or min does to translate in to a large number in absolute terms, the final effect is smaller.

We usually see the anomaly reduce at maximum and minimum extent. This effect is seen most clearly with the anomaly in Hudson Bay (for obvious reasons).

Which is not to say a big anomaly isn't important, its just that what it means depends on whether we are talking anomaly at max or min (which is mostly the amount of ice) or at time of greatest change (which is mostly about timing). 2007's record anomaly was certainly helped by the low baseline at minimum (~ -1.7M km^2), but was mostly because the freeze that year was also very late.

For the maximum anomaly, it wont matter whether 2011 beats 2007's minimum. It's close enough that it will or will not set an anomaly record based mainly on whether the freeze-up is earlier or later than 2007. That depends on how fast the Arctic Ocean dumps its excess heat, and on the vagaries of the weather.

Bob Wallace

Prelim 5,009,375

Down 46,406

That's a good gain on 2007, cuts the 2007 lead to 190,937. Promises a break below 5 million tomorrow.

Paul Klemencic

Yvan Dutil, thanks for the offer to check the lag in the reports, but unfortunately I can't provide the file of numeric data for the Bremen ice extent report. According to Hamilton, who has access and comments on this site , Bremen won't release the actual numbers yet, and only publishes the graph.

Finding the numerical result for JAXA is easy (one of the reasons this site is popular), and the MASIE site allows you to download the spreadsheet. But without Bremen data, we can't do the analysis.

I have read from the guys here that NSDIC uses a five day average, so a comparison of rolling five day averages might show the lags. The NSIDC report seems to compare with Bremen well; its JAXA now that is the clear outlier.

L. Hamilton

"It seems that the real extent that matches the ice extent maps right NOW, is only 4.8 million sq km."

Paul, that corresponds roughly to the Bremen extent for 8/24.

Paul Klemencic

For me, finding the lag in the regional extent numbers is an eye opener. I couldn't understand why the regional extent reports held up as ice disappeared dramatically in the Beaufort, Chukchi, and E. Siberian regions. I can roughly estimate the amount of area covered by the ice on the Bremen map, but the regional extent reports showed extents much too high.

Now that I realize there is a significant delay in the MASIE regional reports, then I can roughly estimate the current "real" ice extent, and its less than 4.80 million sq km already, and could be as low as 4.70 million. This compares well with the Bremen graph of ice extent, and is consistent with the NSIDC graph.

Given that this is the real extent of ice left, and that JAXA is simply catching up to the current reality, with their averaged and delayed report; then I can say we have a much better chance of wiping out the 2007 minimum, probably over a 70% chance. There is too much weak ice pack in the current extent of less than 4.80 million not to lose at least 550k of extent from this point. The weak ice in the E. Siberian, Chukchi, and Beaufort regions, along with the hole near Severnaya Zemlya will do the trick. Any further ice loss on the east side, or flush from the central Arctic Basin simply pushes it down further. And there is 3 1/2 weeks to do it. Eventually JAXA will stop reporting ice where "there ain't any".

The Bremen and NSIDC extent graphs show that 2011 has caught 2007 already.

The only thing that could prevent a record new minimum low would be very unfavorable weather. And since the ice pack in 2011 is in much worse shape than 2007 at this point, the weather will need to turn really cold, with a big low on the Canadian side to prevent the new extent low.

Paul Klemencic

Thanks Hamilton, for the confirmation. I was fairly certain we are below 4.80 million at this point.



If you compare the UB map for Aug 26. with the map for the same day in 2007, the extent for this year looks slighly greater, say by about 200 k matching Jaxa extent gap.
Also, if we want to compare SIE between the two years, we should be using the same gauge. The record extent measured by Jaxa in 2007 was 4.2 million. The minimum measured by UB in 2007 might have been even smaller. Until we have another source for SIE mumbers, we have no choice but to rely on Jaxa.



If you compare the UB map for Aug 26. with the map for the same day in 2007, the extent for this year looks slighly greater, say by about 200 k matching Jaxa extent gap.
Also, if we want to compare SIE between the two years, we should be using the same gauge. The record extent measured by Jaxa in 2007 was 4.2 million. The minimum measured by UB in 2007 might have been even smaller. Until we have another source for SIE mumbers, we have no choice but to rely on Jaxa.



If you compare the UB map for Aug 26. with the map for the same day in 2007, the extent for this year looks slighly greater, say by about 200 k matching Jaxa extent gap.
Also, if we want to compare SIE between the two years, we should be using the same gauge. The record extent measured by Jaxa in 2007 was 4.2 million. The minimum measured by UB in 2007 might have been even smaller. Until we have another source for SIE mumbers, we have no choice but to rely on Jaxa.



If you compare the UB map for Aug 26. with the map for the same day in 2007, the extent for this year looks slighly greater, say by about 200 k matching Jaxa extent gap.
Also, if we want to compare SIE between the two years, we should be using the same gauge. The record extent measured by Jaxa in 2007 was 4.2 million. The minimum measured by UB in 2007 might have been even smaller. Until we have another source for SIE mumbers, we have no choice but to rely on Jaxa.



If you compare the UB map for Aug 26. with the map for the same day in 2007, the extent for this year looks slighly greater, say by about 200 k matching Jaxa extent gap.
Also, if we want to compare SIE between the two years, we should be using the same gauge. The record extent measured by Jaxa in 2007 was 4.2 million. The minimum measured by UB in 2007 might have been even smaller. Until we have another source for SIE mumbers, we have no choice but to rely on Jaxa.



If you compare the UB map for Aug 26. with the map for the same day in 2007, the extent for this year looks slighly greater, say by about 200 k matching Jaxa extent gap.
Also, if we want to compare SIE between the two years, we should be using the same gauge. The record extent measured by Jaxa in 2007 was 4.2 million. The minimum measured by UB in 2007 might have been even smaller. Until we have another source for SIE mumbers, we have no choice but to rely on Jaxa.

Paul Klemencic

Phil263, you don't need to compare the two maps to see if there is a difference in Bremen extents between 2007 and 2011... just look at the Bremen graph of ice extents (on Neven's daily graph page), and you can see 2011 is lower now than 2007. So the estimate of 200k difference based on your eyeball review of the maps is not accurate.

Besides, I arrived at the 4.8 million number several ways. I used the MASIE estimate of 5.2 million for about a week ago, and broke that into the regional extents, then adjusted each regional extent by estimating the ice lost in the last week in that region. The total of the extent losses totaled about 400k in the last week, so removing that 400k from 5.20 million gives the same approximate 4.8 million as the Bremen graph (confirmed by Hamilton) for August 24. Also the NSIDC shows 2011 tracking 2007, so the difference is negligible.

JAXA hasn't caught up with the flash melt last week, or may be picking up a bunch of dispersed ice, but I really don't care. The main reason I accepted the JAXA result in the first place, is because at 5.1 and change, it fell between the MASIE total of 5.2 million and the Bremen and NSIDC result. But now we know the MASIE correct number is 4.8 or so; that leaves JAXA as the odd man out.

I also don't really care about the reporting differences. If the ice extent really hits a new low this season, all of the systems will likely show 2011 extent as the new low. The main reason I needed an accurate current ice extent number (the 4.8 million reading), is because I was trying to estimate the actual final minimum, by estimating the ice extent melt and loss in each region. It makes a huge difference on August 24 if I subtract 600k from almost 5.2 million (result = 4.6), versus subtracting 600k from 4.8 million, resulting in a reasonable estimate of 4.2 million for the minimum.

Its game over, now that we know Bremen agrees with the estimate of 4.8 million. Only a near miracle event of cold weather and a low in Canada could save the day, and keep 2011 from becoming the new record low ice extent.

Rob Dekker

Paul, I have not followed your entire analysis, but when you state It seems that the Bremen leads the JAXA IJIS by 5-7 days. and JAXA hasn't caught up with the flash melt last week do mean the SIE or the SIA ?

Since temps are dropping rather quickly (DMI 80N+ chart) right now, I expect that melting ponds will freeze over and get covered by snow in the next couple of days, which should show up as increased ice concentration, and thus increased SIA.

So if you meant SIA with a 7 day delay, then the next couple of days would be a good test for your (and mine) hypotheses.


Haven’t posted some days, tired of mixing ‘blowtorches’and conservative opinions. I fled into CAD and had my own look at the state of affairs.
Outcome: 2011 will set a new record on volume loss. On area it’s close; normally 15 days of average melt weather conditions will take 2011 to a record. Maybe Rob is right, refreeze might be early this year over 80 degrees north.
On extent: new record isn’t likely. Only if there’s enough melt ‘still in the pipeline’, when the surrounding SST’s are indeed anomalously high and the weather would facilitate. Last year showed a steep decline in the last three weeks before the minimum. If that corresponds not only to facilitating weather then, but to the thin spread (2010 finally had 1300MK), 2011 has an advantage (2011 has only 1743MK high concentration!).
I’m not going to mix in debates on the quality of the diverse data-sets. I’m using my own eyes, CAD and CT maps as a basis to get a hold on the differences. I check MODIS for confirmation. I hold PIOMAS as reasonably reliable. I like weather forecasts, but rather focus on the actual situations.
When it all boils down to the minimum I think we may all agree that the decline of arctic sea ice has continued(a lot...).
Last remark. When the CT map shows 794K less +30% concentration (25082011 versus 25082010), what would you think that may tell us about volume?

Ned Ward

When I first tried my statistical model for predicting IJIS-JAXA minimum extent (back in early July), it gave the following:

4.5 million km2 [3.5 to 5.4]

I updated that on 18 August, by which point the expected minimum had risen slightly, and the uncertainty had narrowed a lot:

4.7 million km2 [4.2 to 5.15]

By 23 August it had narrowed a little more:

4.6 million km2 [4.1 to 5.0]

Here's today's prediction:

4.6 million km2 [4.25 to 4.9]

The predicted minimum has been remarkably stable, while the uncertainty keeps narrowing. The 2007 minimum is hovering right on the -2 sigma threshold, so exceeding it is still possible but increasingly unlikely. (Note that the SEARCH outlook predictions are assessed based on the monthly NSIDC minimum, not the daily IJIS minimum; the two are likely to be fairly similar).

This is a conservative model, and doesn't incorporate a possible trend towards more-rapid loss of extent in the final month of the melt season ... a trend for which there is some, but not a lot, of evidence. On the other hand, the fact that the predicted minimum has barely changed at all over the past three weeks (oscillating between 4.5 to 4.7 million km2) is a bit reassuring.

Lucia (The Blackboard)

This morning, my weighted model's best estimate is 4.51.

The method picks from a bunch of candiates models and weights using AICc criterion, the qualifying candidates with non-zero weights predicted: extent=(4.56, 4.44,4.57,4.64). weights for that model (5%,51%,35%,9%) The four regressions were : extent v extent, extent v( extent and area), remaining loss v (extent-area) and mean remaining loss loss.

So these are all near yours, but I end up a little lower.

Ned Ward

Lucia, I'm sure your approach is better. Mine is just something simple that I set up earlier in the season, and I've kept it running as-is for the sake of consistency.

My guess is that you are capturing some of the trend towards increased extent loss in the last few weeks of the season, which can probably be predicted better by including area. I've looked at other ways of accounting for that, and it led to a predicted minimum extent of 4.4 or so.

I have learned quite a bit from off-and-on lurking on your site, by the way. You seem to have found a useful niche in the climate blogosphere.

Lord Soth

Tonight will be the night for breaking the 5M barrier. We only had a 500km correction this morning, so we should be good for a decent ice loss tonight. The bar is pretty low anyways with only 10K of lost needed.

We are now 190K above 2007, and we are entering the horserace for the finish.

Unless 2011 stalls, this 190K should be erased by the five day 2007 stall that occured at the end of August.

We may be biting our nails for most of September as 2011 and 2007 trades first place, right up to the photo finish.

Paul Klemencic

Alfred Hitchcock put a great scene in "North By Northwest". Cary Grant goes to an assigned meeting place in the middle of nowhere, with empty fields and cornfields stretching off in all directions. An old farmer comes along, and at first Grant thinks the man is his clandestine contact. But then it turns out the man is just waiting for a bus. Grant tries to talk to the farmer, but the old man is taciturn and short of words, and just keeps studying a cropduster plane on the horizon.

But then, just before the farmer gets on the bus, he tells Grant: "Funny thing about that cropduster... He's dusting crops where there ain't any crops." Grant shakes his head, missing the point entirely... and because of that, he almost gets killed.

Well, Ned and Lucia, you need to know that the IJIS-JAXA report is measuring ice "where there ain't no ice". Over 200k sq km of the JAXA "ice" is already gone; only the memory of this ice lives on in the JAXA reported extent.

How do we know this?

For one thing, I know it because of the regional breakdown. The central Arctic ice extent has been stable for a long time, the Laptev, Kara, and all most all other regions have also been stable and predictable. But then I looked for the ice that must be there in the E. Siberian, Chukchi, and Beaufort. it is easy to see looking at the maps that the ice extents in those regions are over 200k short of the extent needed to get to the 5.01 million reported by JAXA.

Several independent measurements, the Bremen ice extent graph, and the NSIDC trend, show the same thing as an examination of the ice extents on the maps. These reports are consistent with a total ice extent below 4.80 at this time.

Somehow, the JAXA has some kind of lag and averaging system that causes their report to continue reporting ice extent that has already melted out. An accurate assessment of the ice extent, right NOW, is less than 4.8 million sq km. If the JAXA system is still picking up really weak ice areas, or has the lag I suggested, their reported extent will begin to close on Bremen's extent level as the melt season winds down, and the melt rate slows.

Since the real extent is less than 4.8 million now, many of the upper bounds in your statistical analysis are impossible, and even the likelihood of 4.6 or higher is extremely unlikely. On the other hand, the probabilities of extents below 4.40, and even below 4.25 are much higher than your models suggest.

It might be wise to listen to the advice of the old farmer, who used his skills of observation, to see something the city slicker couldn't.


I'd rather trust a countryman than a townman,
You can judge by his eyes, take a look if you can,
He'll smile through his guard,
Survival trains hard.
I'd rather trust a man who works with his hands,
He looks at you once, you know he understands,
Don't need any shield,
When you're out in the field.

Very interesting stuff you've done, Paul. *takes hat off*

Seke Rob

MASIE dropped 74.3K from 25th > 26th. The kicker is Hudson Bay, with a loss of... drum roll... 67 KmSq. Curiously, Baffin has a hike 1,515 KmSq. Is that freeze up or feeding from the CA/Nares? The current Bremen map shows some spurious stuff down near the Greenland tip, which is all I can make out of that 9.5K I can make out on there.

Kevin O'Neill

Does anyone have an explanation as to why the Cryosphere Today northern hemisphere map http://arctic.atmos.uiuc.edu/cryosphere/NEWIMAGES/arctic.seaice.color.000.png now has the Arctic ringed by shore ice?

At first I thought this must be the beginning of the winter refreeze, but MODIS doesn't show any ice that I can see http://rapidfire.sci.gsfc.nasa.gov/imagery/subsets/?subset=Arctic_r05c05.2011239.terra.250m .

The MODIS image above should be showing the Laptev with the best shore view on the Siberian coast, west of the New Siberian Islands, in the approximate area of Tiksi. I see sand and no ice. Is their algorithm interpreting this brightness as new ice?


UB prelim is out and would probably suggest quite the ice gain extent gain in the Beaufort.

There is no way the ice is refreezing there with the current temps. So maybe the sensors are messed up.

but n area of 100Km2 or more 50-60 concentration ice just showed up.

Seke Rob

Slide down to the noted Greenland tip and way further down... sensors confused. Can't remember to have seen Svalgard so free from the icepack.

Lucia (The Blackboard)

I don't know whether my method is better or worse. I just thought you might be interested in how our specific numbers are comparing.

Does UB post numerical values or are you just eyeballing graphs?


Kevin if you look back over the last month there is a similar bright edge to the arctic all the time. you may be right about the sand as it fluctuates from day to day.

The comments to this entry are closed.