« Record dominoes 7: Arctic ROOS sea ice extent | Main | ASI 2012 update 10: (wh)at a loss »

Comments

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

Timothy Chase

Seke Rob wrote on September 1st:

Somewhere saw a monthly statistic from observation data of top/bottom loss during different parts of the melt season...
Sorry I didn't see your comment until now.

I am not sure that this is the sort of thing you are looking for, but try:

Figure 11, pg 10 of Michael Steele, Jinlun Zhang, Wendy Ermold (2010) Mechanisms of summertime upper Arctic Ocean warming and the effect on sea ice melt, Journal of Geophysical Research, Vol. 115, C11004

Seke Rob

Thanks, Timothy, you nailed it right on the head. Bookmarked it and remembered the key-words to search in Google Scholar.

Ice mass-balance buoys: a tool for measuring and attributing changes in the thickness of the Arctic sea-ice cover
http://www.ingentaconnect.com/content/igsoc/agl/2006/00000044/00000001/art00034

There's another new term I'd not come across. "Under Ice Melt Ponds": http://www.jstor.org/discover/10.2307/2838118?uid=3738296&uid=2&uid=4&sid=21101166054461.

Seke Rob

NSIDC August charts are up: http://nsidc.org/data/seaice_index/images/n_plot_tmb.png and the report: http://www.nasa.gov/topics/earth/features/earth20120422.html

Lord Soth

What scares me is the amount of first year remaining in the Arctic. Taking the third week of August NSIDC ice age chart, and comparing those areas to the Sept 5 sea ice extent limits, I got roughly 400,000 km of first year ice left in the Arctic, and its not over yet. We had around 11.4 million km of ice at the start of the melt season, so we have already lost 96.5% of first year ice so far.

Add in the transpolar drift, we will see a generation of ice that will be virtually non-existant in March.

But fear not, I still hold out hope for recovery via a super volcano eruption, major asteroid impact, basalt floor eruption or some other extinction level event :)

crandles

>"so we have already lost 96.5% of first year ice so far"

Not sure that quite agrees with

from Arctic sea ice news
http://nsidc.org/arcticseaicenews/

Lord Soth

Crandles, that image does not include Hudson Bay, Labrador Sea, Fox Basin, Davis Strait, etc, etc.

Well not technically ice in the arctic, its part of the southern extent of the arctic sea ice pack in March.

Also, I used the ice edge for Sept 5, which gets rid of half the purple in the August 2012 image again.

A-Team

This is an easy exercise in raster GIS. You can calculate the relative areas using any version of Photoshop. Simply go to the high resolution image linked below, mask out colored regions you do not deem relevant to Arctic sea ice, then touch each square of the legend in turn with the magic wand tool. It will then highlight all pixels of the same color and only those.

Drop the pixel counts in a spreadsheet (not including the legend) and renormalize total pixels to known total sq km of the Arctic (which must corresponds to your choice of masking).

http://nsidc.org/arcticseaicenews/files/2000/08/Figure5.png

Kris

The Lord stated:

Crandles, that image does not include Hudson Bay, Labrador Sea, Fox Basin, Davis Strait, etc, etc

So what?

The ice in these regions always melts away in Summer. Always.
Thus to look there for second-year and older ice would be pretty silly.

crandles

Sorry, I think I was confused.

15.3 extent at maximum * 70% is 10.7 so little to quarrel over regarding your 11.4 FYI extent.

Have we really lost half of that purple area in last two weeks? We have lost 800k extent in that time and most of that probably is purple so probably more. Just 400k of current 3.5M extent therefore does look pretty realistic. So yes, 11M of 11.4M FYI extent gone is 96.5% gone.

Sorry about that.

Lord Soth

When NSIDC releases the Setember Ice age chart in October, I will do the calculations properly using Global Mapper and convert to an equal area projection, if it is not already in one of those projections.

Wipneus

Update Sep 6, new low (-53k):

2012, 09, 01, 3.60776
2012, 09, 02, 3.58558
2012, 09, 03, 3.54015
2012, 09, 04, 3.56411
2012, 09, 05, 3.51108


Tr. five day average: 3,561736 (from 3.597342)

Artful Dodger

NSIDC Daily SIE for Sep 5, 2012: 3.511
M km^2

That's a decrease of -53 K km^2 over the previous day, and a new record.

The 5-day trailing avg SIE is now 3.56 M km^2, and Sep-to-date avg SIE is 3.643 M km^2.

Kris

Artful Dodger typed:

NSIDC Daily SIE for Sep 5, 2012: 3.511
M km²

Whereas according to IARC-JAXA 3.614.219 km2 for September 6th.

Or an impressive difference just over 100.000 km².

Nevertheless, we are talking about the very same Arctic, aren't we? :-)

Seke Rob

A joke of a small difference, not more, but given that JAXA is a prelim number you quote, it's more like 150K. Anywhere a drama to discern?

crandles

Given that multisensor masie is down to 3.686, that is a larger difference between NSIDC and IARC-JAXA two passive microwave products and the difference between IARC-Jaxa and MASIE.

Surprising or normal for this time of year?
LY:
MASIE 4.575
NSIDC 4.385
JAXA 4.588

Perhaps 103k difference between NSIDC and JAXA is surprisingly small.

Seke Rob

The 4 way extent chart [DMI/JAXA/JAXA/NSIDC], with zoomed in block of current time section for convenience: http://bit.ly/MASDMI

For NSIDC alone, the Max/Min/Melt chart: http://bit.ly/NSARMX

Since max, 11.78 million km went of 77.04%

r w Langford

Timothy Chase Thanks for that reference it clarifies the ratio of top melt to bottom melt.
"Bottom melting contributes about 2/3 of total volume melt but is geographically confined to the Marginal Ice Zone, while top melting contributes a lesser 1/3 of volume melt but occurs over a much broader area of the ice pack." It is therefor to be expected that bottom melt is very significant in the decrease in ice volume each summer particularly for FYI.
http://psc.apl.washington.edu/zhang/Pubs/Steele_etal_2009JC005849.pdf

Seke Rob

Visually, smack on the 7 sigma deviation for September 5: http://bit.ly/NSAR01 "All is normal" in the willed bliss quarters.

Cartoon time, oldie but goodie: http://voices.washingtonpost.com/tomtoles/2010/06/17/c_06182010.gif

Apocalypse4Real

I anticipate that the wind and warmer air from the SLP moving into the Beaufort Sea and then on into the Canadian Arctic over the next 3 days will have an impact on the ice - both in melt and extent.

http://polarmet.osu.edu/nwp/animation.php?model=arctic_wrf&run=00&var=plot001

The NRLSCC ARC ice motion forecasts of 090512 still model the storm generated motion and ice flow through the Fram.

http://www7320.nrlssc.navy.mil/hycomARC/navo/arc_list_arcticicespddrf.html


Werther

Evening A4R,
Watching the models too. They show lots of movement in the lower troposphere for the next 10 days.
It might contribute to delay the minimum extent/area days. But what I'm really interested in, now that this summer's melt is almost through, is the effect of 6,25 million km2 icefree Arctic Ocean with the SST anomaly now present.
Last year, some interesting fall weather features took form during September.

Although ECMWF/GFS +72 hour and later forecasts can easily change, there are interesting patterns on the +160 hour range.
A very large mesoscale low pressure zone shows upover the N Atlantic. Probably its energy could be boosted by the remains of TC's Leslie and Michael.
At the same time, the polar jet starts looping out of Siberia.

I'll stop here, because it remains to be seen whether this will get relevant.
What seems important to me is that now our focus should be on the consequences of record low sea ice. If there's going to be WACCY weather, let's be prepared to highlight relation with AGW/polar amplification.

Apocalypse4Real

Werther,

I concur that we may see some unprecendented September weather. I too am postulating that all that warm Arctic water will have an impact by increasing volatility.

BTW

Sea Ice Concentration and Thickness plus kmzs are up for Sept 5th

https://sites.google.com/site/apocalypse4realseaice2012/home/sea-ice-concentration-and-thickness-comparison

Chris Reynolds

Werther,

The chances for another WACC pattern this winter look good on the face of it. Siberian snow extent is severely reduced, and as Cohen argues, it is the fast rate of advance of snowfall that induces the sort of pattern seen in winter 2009/10.

With plenty of open water to provide water vapour for systems leaving the Arctic ocean for Siberia, and low snow cover implying that a return to normal conditions will mean fast snowline advance, the chances look good.

I'll email Dr Cohen again later in the autumn to see if he's making a forecast public this year. He didn't last year, but he got it right - no WACC pattern. Sadly I could only tell people after the event. :(

I've blogged on these issues here:
http://dosbat.blogspot.co.uk/2011/12/cold-winters-arctic-connection.html

http://dosbat.blogspot.co.uk/2011/12/cold-winters-snow-advance-index.html

Werther

Evening Chris,
First, I follow your thorough data-sampling and analysis with most interest. From time to time I have an issue with some details. FI '...no WACC-pattern...'(last year).

Symbolizing all intricate dependencies of weather during NH winter WACC stands for 'warm arctic - cold continent, a contradictionary pattern that tricks lots of yeoman to dismiss AGW.

Indeed, The NE United States didn't experience WACC last year. But what happened during the first two weeks of february on the Eurasian continent was at least regionally Wacking-us-out, while the Kara/Laptev region was warming-in-the-dark at the same time.
So maybe we could settle on saying WACC was a regional pattern last winter, while '09-10 and to a lesser degree '10-'11 WACC affected both Eurasia and N America.

Hi A4R, I know there are multiple scientific papers on possible drivers of WACC ( temp gradient between N.Japan-Pacific region and Chukchi Sea / Siberian snow gradient / Kara-Barentsz Sea SST's / polar-tropics temp gradient and the fading-looping jet stream to mention some).
I guess they are all interrelated phenomena on a sort of meso-scale.
I'm intrigued how the really larger process of redistributing warmth from the tropics to the poles could drive these mesoscale phenomena. In a coupled ocean-atmosphere system the thermohaline circulation, the ENSO cyclus, atmospheric Rossby-waves and the role of cyclones all contribute to less and less stability. We're going to notice/face the effects.

Apocalypse4Real

Werther,

The predominate pattern last winter was a La Nina with positive Arctic Oscillation for most months.

Nebraska, USA should not have 15-20 C temps for highs in January and February...We had some days with 15-20C positive anomalies.

Do not know if that was a factor elsewhere, but that is what came clear to me in the US.

The last El Nino winter (2009-2010) with negative Arctic Oscillation for most of the winter brought the most snow and coldest temps in 15-20 years.

So I am interested in the same interrelated phenomena.

Espen Olsen

IJIS today 3601875 km2

Artful Dodger

NSIDC SIE Sep 6: 3.488 M km^2

Wipneus

Update Sep 7, new low (-20k):

2012, 09, 02, 3.58558
2012, 09, 03, 3.54015
2012, 09, 04, 3.56411
2012, 09, 05, 3.50857
2012, 09, 06, 3.48834


5 day tr. average: 3.53735 Mm2

Seke Rob

In slower motion:

yyyy m dd extent. chnge #rc 5DayAvg
2012 9 4 3564110 23960 247 3597342
2012 9 5 3511080 -53030 248 3561736
2012 9 6 3488340 -22740 249 3537852

Artful Dodger

High noon at 80° N, its 29.9 F and no sea ice...
"Promise and Deliver"

Roman Polach

Artful Dodger:

It looks like Healy is in Mediterranean Sea :)

Artful Dodger

NSIDC SIE Sep 7: 3.57635 M km^2

This is an increase of 88 K km^2 over the previous day.

Wipneus

5 day tr. average: 3.535504 Mm2

(-2 k)

Artful Dodger

NSIDC SIE Sep 8: 3.53345 M km^2

This is a decrease of 43 K km^2 over the previous day.

Wipneus

Update Sep10, -33k:

2012, 09, 05, 3.50857
2012, 09, 06, 3.48834
2012, 09, 07, 3.57635
2012, 09, 08, 3.53345
2012, 09, 09, 3.49977


5 day tr.average: 3.521296 (from 3.534164)

Artful Dodger

NSIDC SIE Sep 9: 3.49977 M km^2

This is a decrease of 33.7 K km^2 over the previous day.

Apocalypse4Real

Who imagined at the beginning of this melt season that we would be seeing a lt 3.5 million km2 NSIDC SIE?

While this year is "almost" done, one begins to wonder what next year might bring?

We ended up needing a lower category in the SIE vote....

Seke Rob

Not an absolute new minimum, but considering the size of the Arctic, the 9th is practically an ex aequo:

Year M D KMSQRE DAYCHG #Day %DayAvg
2012 9 5 3511080 -53030 248 3561736
2012 9 6 3488340 -22740 249 3537852 < abs.min.
2012 9 7 3576350 88010 251 3536006
2012 9 8 3533450 -42900 252 3534666
2012 9 9 3499770 -33680 253 3521798

Artful Dodger

NSIDC SIE Sep 10: 3.46646 M km²
Decrease of -33 K km²

Seke Rob

Creeping up to the 12 million loss from max to min... and a new day record low.

http://bit.ly/NSARMX
http://bit.ly/NSARMn
http://bit.ly/NSAR01

Wipneus

Update Sep 12 new low (-44.5):

2012, 09, 07, 3.57635
2012, 09, 08, 3.53345
2012, 09, 09, 3.49977
2012, 09, 10, 3.46646
2012, 09, 11, 3.42193

5 day tr. av. 3.499592(from 3.512874)

Espen

Yes and it looks like there is even room for a little bit more, when watching the weather forecasts?

Wipneus

Update Sep 13, up 7k:

2012, 09, 08, 3.53345
2012, 09, 09, 3.49977
2012, 09, 10, 3.46646
2012, 09, 11, 3.42193
2012, 09, 12, 3.42945

5 day tr. average: 3.470212 (was 3.49959)

Artful Dodger

There, see? Our handsomest politicians have dropped an ice cube in the ocean, thus solving the problem of global warming forever. FOREVER! ;^)

To be fair, they're our handsomest business executives who block action on climate change.

Cheers,
Lodger

Wipneus

Update sep 14, up 21.5k:

2012, 09, 09, 3.49977
2012, 09, 10, 3.46646
2012, 09, 11, 3.42193
2012, 09, 12, 3.42945
2012, 09, 13, 3.45096

5 day tr. avg: 3.453714 (from 3.470212)

Wipneus

Update Sep 15, new low (-32k):

2012, 09, 10, 3.46646
2012, 09, 11, 3.42193
2012, 09, 12, 3.42945
2012, 09, 13, 3.45096
2012, 09, 14, 3.41878

5 day tr. average: 3.437516 (was 3.453714 )

Seke Rob

New record, in low extent but not for lateness with NSIDC's 1 day minimum. Sep.22, 1989 it was, but then there was 6.89M km^2 extent. Same day then it was 6.97M i.e. 0.08M more. Is there a boding in there what still could be?

George Phillies

Accuracy of these numbers? See the popular science report

http://www.guardian.co.uk/environment/blog/2012/sep/13/less-arctic-sea-ice-satellites

proposing that the amount of ice may be being overstated by satellite reconnaissance, to that there is actually a bit less ice than some numbers indicate.

crandles

Hopefully they are consistently wrong and therefore are a useful guide.

2010 managed to lose 221k in next 5 days.

There is quite a bit of low concentration ice for the time of year. So if we get high pressure causing compaction, extent goes down. If we get low pressure causing dispersal, concentration goes below 15% and extent still goes down.

(There is still the possibility of having neither compaction nor dispersal in which case the low concentration ice could melt down to the 15% if there is enough heat in mixed water or freeze up if not.)

221k is the record loss from this latest data's date, and still seems a lot despite large low concentration areas.

Bob Wallace

"proposing that the amount of ice may be being overstated by satellite reconnaissance, to that there is actually a bit less ice than some numbers indicate."

Might this arise from a misunderstanding of what extent and area maps/numbers represent?

If the "ice"/"no ice" threshold is set at 15% then one should not be surprised when moving into areas mapped as frozen contain up to 85% open water.

LRC

@Bob: "Might this arise from a misunderstanding of what extent and area maps/numbers represent?"
I do believe that is a great discredit to those who work and study in the Arctic on a regular basis. The reality is, no matter what system you put into a satellite, they do not really 'see' ice. They are now finding that the data received by the sensors can actually be a variety of different ice types and the default interpretation is be coming more and more the wrong one.

Apocalypse4Real

The NCOF/Godiva2 Sept 14 sea ice concentration and thickness imagery is up.

https://sites.google.com/site/apocalypse4realseaice2012/home/sea-ice-concentration-and-thickness-comparison

Bob Wallace

LRC - I'm having a problem finding where I spoke of satellites seeing the ice.

What I was pointing out that when you are measuring the presence or absence of ice in a grid square there is no information other than "15%" or "not 15%". Area and extent do no discriminate five meter thick solid ice from a foot of rotten slush.

Wipneus

Update Sep 16, new low -25.5k:

2012, 09, 11, 3.42193
2012, 09, 12, 3.42945
2012, 09, 13, 3.45096
2012, 09, 14, 3.41878
2012, 09, 15, 3.39324

5 day tr. aver.: 3.422872 (was 3.437516)

Peter Ellis

Bob: No. The satellite data (after processing) gets translated into a percentage concentration. Julienne Stroeve (of NSIDC) is on the ship, and is quite clear that she is in a grid where the satellite shows >80% concentration, but the reality on the surface is <50% concentration

Seke Rob

A rather lengthy week ending Sept.8 log entry from the Polarstern. There's deepsea life.

The report closes with an accident recording and the need to return to Norway.

But at the very end of the week, there are also less good news to report. During a routine safety exercise with our rescue boats, an accident occurred and several people were hurt. Besides some smaller injuries, one scientist broke his arm and now needs to be transferred to a hospital on land. Hence, Polarstern is now on the way back to Norway. We expect to pick up work again in a week from now, but will of course continue to report more from our research next week.

The date of last log end implies there will be another one in the next few days ending Sept.15.

If Julienne Stroeve were on this ship [IIRC commented somewhere on this blog], no mention of her presence. A quick google search suggests though she is on the Arctic Sunrise departing from same Tromsø port where the Polarstern was heading: http://iceedge2012.wordpress.com/

Espen

Now I understand why Polarstern went very close to Kirkenes (Norway), and then turned back toward the north.

Seke Rob

It's btw good to read that Dr. Stroeve actually gave a potential reason why the sats were seeing more ice... dense fog.

Timothy Chase

Seke Rob wrote:

It's btw good to read that Dr. Stroeve actually gave a potential reason why the sats were seeing more ice... dense fog.
For the sake of those that didn't follow your earlier link and find it at http://iceedge2012.wordpress.com :
According to the satellite data, we should have already reached nearly 100 percent ice concentration, yet at 83N, the ice concentration remains less than 40 percent. It could be that the heavy fog has resulted in an overestimation of the ice concentration from the passive microwave satellite observation.

In search of large ice floes, by Julienne Stroeve (from her blog Ice Edge 2012, September 11, 2012)

Timothy Chase

Speaking of Arctic experts, at the Guardian's Arctic sea ice hits record low: live updates and reaction Walt Meier is quoted:

"It used to be that it all hung together like a big herd of animals almost protected. Now there are pieces that are becoming more isolated, almost separated from the herd, and they can be more easily melted," he told Suzanne Goldenberg.
I have written on the importance of metaphor before. Joe Romm has written a book that spotlights its importance in communication. I would submit that this is a great metaphor. (Or if one wishes to be exact, simile.) It really helps to get across what is happening to the multiyear ice and helps to explain the process through which it is being destroyed in terms that the layperson can understand.

Wipneus

Update Sep 17, new low (-24k7):

2012, 09, 12, 3.42945
2012, 09, 13, 3.45096
2012, 09, 14, 3.41878
2012, 09, 15, 3.39324
2012, 09, 16, 3.36855


5 day tr. avg: 3.412196 (was 3.42287)

Seke Rob

Shakes head [on the litany of seemingly never letting off new low records], and updates charts, the question being: What will be for 2013... even lower or "quadruple recovery" (in the 4th dimension, for those who think that time is on their side).

Whilsts, IJIS has as yet not moved for a 16th number.

Wipneus

September 1-16 average now 3.491 Mm2

Since we are now about half way, there is a good chance the final September average will about that or a lower figure. If only the decline would stop...

Lord Soth

This is getting crazy. Everytime I draw a line in the ice; it melts !

The Average DMI temperture North of 80, is only -2 and is a good 6 degrees celsisu above normal. You can barely make Nilas in perfect condtions.

Here is the image from the North Pole Web Cam (which should be called the almost entering Fran Strait Web Cam)

http://psc.apl.washington.edu/northpole/NPEO2012/7.jpg

Just another rainy day at 81.7 degrees North

Wipneus

Update Sep 18, up +66k:

2012, 09, 13, 3.45096
2012, 09, 14, 3.41878
2012, 09, 15, 3.39324
2012, 09, 16, 3.36855
2012, 09, 17, 3.43451

5 day tr. avg: 3.413208 ( was 3.412196), first uptick here?

Seke Rob

The 17 day September average, of the 5 day trailing [used by NSIDC to compute the monthly] does still go down... 3522726 it's now. By comparison to the 2007 September all time record of 4.3 million 18% less. Of course the present average wont stick with the uptick having started across all the daily sea ice reporters [MASIE bumped up 122K]. Last year the 18th to 30th NSIDC extent increase was 423K as an reference. Adding half to the present average and we're looking at 3.75 million km square or thereabouts, 550K below 2007. We'll know in 2 weeks if the Arctic behaves "new normal".

Nightvid Cole

Seke Rob,

Are you sure the monthly average for September is the 30-day mean of the 5-day running mean? They average twice?

crandles

3.36855
3.43451
3.51352

We are now 145k above that 3.36855 and no year has managed a drop of that amount nor even more than 70k from this data point time onwards.

So it is certainly looking likely that 3.36855 is the one day NSIDC minimum.

I think that is probably just about enough margin to risk calling it.

So, 3.36855 is the one day NSIDC minimum for 2012.

Seke Rob

Nightvid Cole, I came within few Ks of their monthly number and tested this method on a half dozen months. Simply speaking they're not doing a monthly on the single days series but on something else close to the 5 day trailing averages.

BTW, just checking, NSIDC seem to have reprocessed August, where before they showed 4.98 (extent) over 2.60 (area). It now shows for monthly.

2012 8 NRTSI-G N 4.72 2.52 (without the blind spot of 0.31M)

Taking the latest file dailies the extent average works out as 4.807M

The 5 day trailing works out as 4.987M

What the heck, that's both ways a considerable difference. Bummer, why is that.

There's also a September 18 update, provisionally declaring the 16 as the minimum at 3.41M http://nsidc.org/arcticseaicenews/

Seke Rob

Addendum: The 3.41M of the 16th in their report ties with my 5 day trailing of 3,412,196 for same date.

Seke Rob

Dang, reprocessing needed as per the footnote in the Sept.18 update :(

Note that the dates and extents of the minimums have changed since we originally posted in 2007; see our Frequently Asked Questions for more information.
Wipneus

Update Sep 20, up (+16k):

2012, 09, 15, 3.39324
2012, 09, 16, 3.36855
2012, 09, 17, 3.43451
2012, 09, 18, 3.51352
2012, 09, 19, 3.52948

5 day tr. average: 3.44786 (was 3.42572)

Wipneus

Update Sep 21, down -4k:

2012, 09, 16, 3.36855
2012, 09, 17, 3.43451
2012, 09, 18, 3.51352
2012, 09, 19, 3.52948
2012, 09, 20, 3.52524

5 day tr. avg: 3.47426 (was 3.44786)

Sept 1-20 avg: 3.492810 or 3.521923 (based on 5 day averages)

Wipneus

Update Sep 22, up +57k:

2012, 09, 17, 3.43451
2012, 09, 18, 3.51352
2012, 09, 19, 3.52948
2012, 09, 20, 3.52524
2012, 09, 21, 3.58193

5daytr avg: 3.516936

Sept 1-21 avg: 3.497054 or 3.521686

Seke Rob

Wipneus, how do you compute that 5 day average based average for September? We agree on the single day average i.e. there is no discrepancy in the 3.497054, but I have 3.515757 for the trailing MTD [simply the average even at start of month include the last August]. It's just a very minor difference, so no sweat.

Wipneus

Seke Rob:

The first number is based on single day data, the second for the 5 day trailing averages.

If it was established that NSIDC does the first thing then I will be more than happy to drop the doubly averaged number (I was not sure).

Seke Rob

Well, I tested the simply 5 day trailing average trailing on half a dozen some weeks ago [see other thread posts], and they were all within a few K, but then they revised August'12, and neither method comes close to the 4.72 they have as monthly number. For August I had originally a trailing of 4.987 and a single of 4.807. Both far removed from the 4.72 final: 2012 8 NRTSI-G N 4.72 2.52

The July test was

Single: 7.756
5DayTrl: 7.947
FinalMth: 7.94

Did this for said half dozen, so was satisfied with the approach to get a very close on the 1st of the month. All confidence out the window. Usually I check the last 5 years for revisions of a month series file, as often they change prior years. Given that through 2010 is declared final [for dailies], would only expect a change to 2011 when there's a new month out. What ever the interpolations they do [the temp files classic], sometime decades back there are small changes. Ever so slightly, maybe old shipping logs they're putting their hands on, or a site relocation with date going back years, which then requires a spacial correction over a longer period... don't know, but guess that's the nature of the beast... continuous reprocessing needed. Lucy at the Blackboard for long ran these comparisons of changes. The blog bored me [and placing her where she stood became dubious GW], so haven't been there in probably 18 months if not longer.

Wipneus

Update Sep 23, small up +9k8:

2012, 09, 18, 3.51352
2012, 09, 19, 3.52948
2012, 09, 20, 3.52524
2012, 09, 21, 3.58193
2012, 09, 22, 3.59177

5 day tr. avg: 3.548388 (was 3.516936)

Sep 1-22 avg: 3.501359 (based on single day data)

Seke Rob

Who's going on about the Antarctic and it recording a 1 day record for the 22nd, for extent to be precise? Curiously CT-SIA for the Antarctic dropped... 3 days in a row, which begs the question... are we looking at rotten ice at the Antarctic too?

2012,7123 1,031687 16,053446
2012,7151 0,938474 15,983150
2012,7178 0,854205 15,932101
2012,7206 0,799209 15,891980

Time to set up a CANTPIE and see if there's a trend, of any significance.

Wipneus

Update Sep 24, up 59k5:

2012, 09, 19, 3.52948
2012, 09, 20, 3.52524
2012, 09, 21, 3.58193
2012, 09, 22, 3.59177
2012, 09, 23, 3.65133


5 day tr. average: 3.57595 (was 3.548388 )
Sep 1-23 average: 3,507880 (based on single day data)

Wipneus

Update Sep 25, up 37k:

2012, 09, 20, 3.52524
2012, 09, 21, 3.58193
2012, 09, 22, 3.59177
2012, 09, 23, 3.65133
2012, 09, 24, 3.68844

5 day tr. avg: 3.607742 (was 3.57595)
Sep 1-24 average: 3.515403 (based on single day data

Seke Rob

We're starting to see multiple indicators that the monthly averages are likely/possibly going to be below the single day past records... a paradigm shift I'd venture to think.

Wipneus

Update Sep 26, up 34k5:

2012, 09, 21, 3.58193
2012, 09, 22, 3.59177
2012, 09, 23, 3.65133
2012, 09, 24, 3.68844
2012, 09, 25, 3.72289

5 day tr. avg: 3.647272 (from 3.607742)
Sep 1-25 average: 3.523702 (based on single day data)

5 days to go, final September average less than 3.6 ??

Seke Rob

Wipneus,

You can add the average daily growth [either single of 5 day average] of last 3-5 years for these dates and add that to present... it's likely.

A little sound-boarding needed. Developing a step chart for NSIDC extent same as for IJIS. The questions are:

1) on single day or on 5 day running average? (Which smooths the decline effect)
2) on first time below or last time below? (E.g. a series of 13.1M 12.98M 12.97M 13.05M 12.96M There's some longer stretches where extent goes below, then above again for a week, then below a step level again. My inclination is first time (which for the 5 day averages is lots less likely).

Thanks [for opine and motivation of]

Wipneus


Nut very strong of opinion, but since you ask:

1) stick with what NSIDC is using: the 5 day average. No need for explanations when the instant value crosses eg 3.0 Mm2, and in NSIDC reports it did not. Using a moving average instead of trailing makes comparing with IJIS more meaning full.
2) I'd say first time below. Once you cross them, you have crossed them.

Artful Dodger

Hi Seke Rob, hi Wipneus,

There is a step change in the Arctic that has not been fully described or subjected to statistical forecasts. That is, the distribution of sea ice by age.

How is this a 'step change', and why is this important?

The molar enthalpy of fusion (the specific heat of melting) varies by sea ice age. First year sea ice (containing liquid brine, air bubbles) requires about 62 kcal/g of heat energy to melt. Multi-year (low salinity) sea ice takes about 72 kcal/g to melt.

Additionally, the seasonal cycle in the Arctic forces sea ice into annual categories (ice age is a discreet integer value), according to the number of winters it has existed.

So we get a glimpse of a potential way to predict sea ice extent with a statistical model: A multiple regression treating categories of sea ice by age.

Ice age is important because the varying enthalpy of fusion partially explains the pattern of melting observed in Summer 2012: New sea ice melted first in the Central Basin. Old sea ice advected from the Canadian Archipelago persisted in the South, with the last of it melting in the Chukchi sea near Wrangel Is. after the annual minimum.

Sea ice age data can be extracted from the lower chart here: (click for a full-size image)

However, ice age is expressed as a percentage of March sea ice extent, rather than a raw extent value.

This information is easily calculated give NSIDC monthly SIE data, available from this site:

ftp://sidads.colorado.edu/DATASETS/NOAA/G02135/Mar/

I'll do an mget on the ftp files, and post a table of March SIE by year.

Wipneus, I wonder if you could employ your bitmap to vector technology to extract the March sea ice age data from the image above.

I believe focusing on trends in multi-year sea ice is the best way to break through the annual noise, and create an accurate prediction for the year of the first ice free Summer.

Cheers,
Lodger

Artful Dodger

NSIDC March SIE: (1983-2012)
year extent
1983 16.10
1984 15.62
1985 16.06
1986 16.08
1987 15.95
1988 16.13
1989 15.52
1990 15.88
1991 15.50
1992 15.47
1993 15.88
1994 15.58
1995 15.32
1996 15.13
1997 15.58
1998 15.66
1999 15.40
2000 15.27
2001 15.61
2002 15.44
2003 15.49
2004 15.05
2005 14.74
2006 14.43
2007 14.65
2008 15.22
2009 15.14
2010 15.11
2011 14.56
2012 15.21

Twemoran

Lodger

I think you're on to something by somehow quantifying the changes in the ages of MYI.

The losses this year through the CAA should IMO lead to a very fast melt out next year on the Pacific side.

Hope you can get the new charts up while there's still some MYI to work with.

Terry

sofouuk

Lodger, apologies if this is a silly question, but is the point of this to calculate total energy required to melt all the ice at the start of each melting season? if so could you post the results because that would be very interesting :) if not how do you plan to accurately predict the first ice-free summer?

Wipneus

Lodger, quick job, should be checked:

n year 1-year 2-year 3-year 4-year

1 1983 43.67543 35.06001 27.403612 18.788193
2 1984 45.74951 36.65546 29.318149 21.979089
3 1985 46.54723 36.17682 29.637239 22.936357
4 1986 49.09995 41.44180 32.668590 25.967709
5 1987 47.18541 37.61273 31.711321 24.531805
6 1988 46.38769 38.72954 31.392232 25.489074
7 1989 45.11133 35.69819 29.158604 23.095902
8 1990 41.92043 31.39223 24.212716 19.585917
9 1991 41.60135 32.82813 24.212716 17.990469
10 1992 39.68681 29.31815 23.095902 16.873655
11 1993 43.35634 31.07314 22.298178 16.873655
12 1994 41.12271 32.18996 22.936357 15.437752
13 1995 42.55861 28.83951 21.819544 15.278207
14 1996 35.69819 27.56316 18.947738 13.684513
15 1997 42.55861 25.01044 19.426372 12.886789
16 1998 37.77227 28.83951 17.990469 13.684513
17 1999 43.51588 31.23269 23.255447 14.320938
18 2000 37.45318 29.63724 22.138633 16.395021
19 2001 37.45318 25.80816 20.702730 14.640028
20 2002 38.56999 26.76543 18.628648 13.842304
21 2003 37.77227 27.56316 19.107282 13.365423
22 2004 38.09136 26.76543 20.383641 14.320938
23 2005 38.72954 27.56316 19.266827 14.480483
24 2006 36.01728 26.28680 18.309558 12.886789
25 2007 34.42183 22.45772 16.235476 10.812706
26 2008 21.02182 14.00185 9.217258 6.664542
27 2009 24.69135 11.13180 8.259989 6.026362
28 2010 27.08452 12.56770 7.302721 5.228638
29 2011 28.20134 14.64003 5.069094 3.314101
30 2012 24.85089 13.84230 6.983631 2.356832

Artful Dodger

Hi sofouuk,

Not a silly question at all, in fact quite the point! Here's a silly question for you:

How does one pronounce 'sofouuk'? :^)

Hi Terry,

I'm hopeful we can save the WINTER sea ice! To quote your countryman James Cameron, 'the Summer sea ice is toast'.

Bedankt, Wipneus

It's good to have vrienden!

Cheers,
Lodger

Peter Ellis

Lodger: I think the March ice age graph covers a more restricted domain than the total extent figures (e.g. I'm not sure Hudson Bay is included for the ice age graph, while it is included for the extent figure).

It would be better to take the MYI values from the September figure.

Artful Dodger

Hi Peter,

Thank you for your advice. I was thinking about those issues even as I posted the original comment.

Certainly an easy way to find March extent for each ice age is to start with the distribution at the September min. Then add 6 months to the age, and categorize any new ice as < 0.5 yrs old, since no sea ice survives Summer outside the Maslanik domain.

I'm not sure if this approach is totally accurate since March ice age allows for Winter drift and divergence. Leads also create new ice inside regions of old ice, so the Sep distribution is distorted (but maybe not the SIE/Age values?)

Given that any cell with > 15 % area of ice aged N is counted as age N, I think the six months of ice evolution may be important. Clearly I need to do Sep 2011 to Mar 2012 both ways and compare as a double check.

I do need to find a reference defining the domain used for Maslanik sea ice age. From Fig 5 above, it looks like Hudbay is included, but not displayed. The C.A. is NOT included, but Bering Sea is. Okhotsk? Who knows? Anybody's help is appreciated (maybe I'll start here ;^)

http://www.agu.org/pubs/crossref/2011/2011GL047735.shtml

Maslanik, J., J. Stroeve, C. Fowler, and W. Emery (2011), Distribution and trends in Arctic sea ice age through spring 2011, Geophys. Res. Lett., 38, L13502, doi:10.1029/2011GL047735.

So my approach does require the September numbers, since I will attempt a statistical fit to Seasonal ice loss in each category. But I have a little prep work to do first.

Cheers,
Lodger

Wipneus

Update Sep 26, up 32k:

2012, 09, 22, 3.59177
2012, 09, 23, 3.65133
2012, 09, 24, 3.68844
2012, 09, 25, 3.72289
2012, 09, 26, 3.75511

5 day tr. avg: 3.681908 (was 3.548388)

Sep 1-26 avg: 3.532603 (based on single day data)

Seke Rob

First pass step chart using NSIDC extent data.

http://bit.ly/NSTP1D

It's not completely done [the blotted out column on right]

There will be 4 charts:

1) The single day raw data.
2) A standard 5 day trailing
3) A centered moving average
4) A 2 day moving average, to match the IJIS average.

For present, *Do Not* compare to the one for IJIS. You're warned... it's not apples-apples.

P.S. Is what happened in 2010 showing?

Wipneus

Update Sep 27, up 54k5:

2012, 09, 23, 3.65133
2012, 09, 24, 3.68844
2012, 09, 25, 3.72289
2012, 09, 26, 3.75511
2012, 09, 27, 3.80965


5 day tr. avg: 3.725484 (was 3.681908)

Sep 1-27 avg: 3.542864 (was 3.532603 based on single day data)

Three day to go in September, pointing to about 3.58 final value for the average.

Seke Rob

The MTD average based on the 5 day trailing sits at 3.541M ... no 3.6 will be passed. Plugging that into the monthlies chart and you have a Bastardi/Goddard extra extra topping blog-science ordered special recovery.

http://i137.photobucket.com/albums/q210/Sekerob/Climate/AntarcticSIASIESeptember.png

Oops, need to extend the Y-Axis down... the Area falls off. :(

Seke Rob

Anomaly again greater than negative 2.5M

2012.7369 -2.4821250 2.6062055 5.0883307
2012.7397 -2.5004177 2.6334040 5.1338215

Seke Rob

Wrong thread. This was CT-Atmos data

Wipneus

Update Sep 27, up 75k:

2012, 09, 24, 3.68844
2012, 09, 25, 3.72289
2012, 09, 26, 3.75511
2012, 09, 27, 3.80965
2012, 09, 28, 3.88454

5 day tr. avg: 3.772126 (was 3.725484)

Sep 1-28 avg: 3.555066 (was 3.542864 based on single day data)

Wipneus

Update Sep 30, up 9k:

2012, 09, 25, 3.72289
2012, 09, 26, 3.75511
2012, 09, 27, 3.80965
2012, 09, 28, 3.88454
2012, 09, 29, 3.89439

5 day tr. avg: 3.813316 (was 3.772126)

Sep 1-29 avg: 3.566767 (was 3.555066 based on single day data)

One day to go, estimated final Sep average 3.580 Mm2

(no updates from me tomorrow, I will be on the road)

The comments to this entry are closed.