New CAPIE record

The trend line on one of our in-house charts has broken last year's record:

56.85% vs 57.39% on August 12th 2011.

CAPIE stands for Cryosphere-today Area Per IJIS Extent, and it tells us something about the compactness (official scientific term) of the ice. It all revolves around the different definitions for sea ice area and sea ice extent, which are two ways of calculating the total ice cover. The NSIDC explains it well:

Area and extent are different measures and give scientists slightly different information. Some organizations, including Cryosphere Today, report ice area; NSIDC primarily reports ice extent. Extent is always a larger number than area, and there are pros and cons associated with each method.

A simplified way to think of extent versus area is to imagine a slice of swiss cheese. Extent would be a measure of the edges of the slice of cheese and all of the space inside it. Area would be the measure of where there is cheese only, not including the holes. That is why if you compare extent and area in the same time period, extent is always bigger. A more precise explanation of extent versus area gets more complicated.

Extent defines a region as “ice-covered” or “not ice-covered.” For each satellite data cell, the cell is said to either have ice or to have no ice, based on a threshold. The most common threshold (and the one NSIDC uses) is 15 percent, meaning that if the data cell has greater than 15 percent ice concentration, the cell is considered ice covered; less than that and it is said to be ice free. Example: Let’s say you have three 25 kilometer (km) x 25 km (16 miles x 16 miles) grid cells covered by 16% ice, 2% ice, and 90% ice. Two of the three cells would be considered “ice covered,” or 100% ice. Multiply the grid cell area by 100% sea ice and you would get a total extent of 1,250 square km (482 square miles).

Area takes the percentages of sea ice within data cells and adds them up to report how much of the Arctic is covered by ice; area typically uses a threshold of 15%. So in the same example, with three 25 km x 25 km (16 miles x 16 miles) grid cells of 16% ice, 2% ice, and 90% ice, multiply the grid cell areas that are over the 15% threshold by the percent of sea ice in those grid cells, and add it up. You would have a total area of 662 square km (255.8 square miles).

If we divide area by extent, we get an idea about how spread out the ice pack is or not (which is determined by wind patterns). When the ice pack is compact, for instance during winter, all the 'holes in the cheese' freeze over, and so the area and extent numbers come very close, and CAPIE oscillates around the 95% mark.

At the start of the melting season melt ponds fool satellite sensors into thinking that there is open water where there is none (which get calculated for area, but not for extent). This makes CAPIE go down. As the melting season progresses, melt ponds become less and less of an influence because they drain, or the ice floes they are on break up.

We are now at a point in the melting season where the CAPIE percentage tells us something about how much the ice pack is spread out by 1) winds and 2) thinner ice that melts out in situ and leaves open water behind, when compared to other years. As of today CAPIE is the lowest it has ever been in the 2005-2012 period (and probably in the entire satellite record), and very early in the season as well. This is hardly a surprise as low pressure systems have been dominating the Arctic for quite a while now, and we have seen plenty of evidence of the effects.

I have to add one more possible influence on the CAPIE numbers besides divergence and melt ponds. Up till November last year IJIS used data from the AMSR-E sensor. Due to failing of the sensor IJIS has now switched to the lower-resolution WindSat radiometer. In contrast with previous years bigger jumps have been observed this year in the daily IJIS sea ice extent numbers (up as well as down). For instance, whereas Cryosphere Today sea ice area has decreased very fast in the past couple of days, IJIS sea ice extent numbers slowed down. This of course means that the CAPIE percentage will go down fast as well.

I wouldn't say it's an apples to oranges comparison, but rather a comparison between Golden Delicious and Jonagold. There's plenty of evidence the ice pack is being spread out a lot by diverging winds from low-pressure areas, and I wouldn't be surprised if it went a tad lower in days to come.

Read more about how CAPIE came into being on the ASI Blog, back in 2010 (followed by this blog post). But remember, just like the PICT thickness chart that is used for monthly PIOMAS updates, CAPIE is a crude method that gives us extra context for current developments. Nothing more, nothing less.

To cut a long story short: there's a lot of potential for extent numbers to drop. If a high-pressure system takes over again, its winds will start to compact the ice pack again. And there's a lot to compact. More than ever in fact.

Comments

[Kris]

Neven wrote:

If a high-pressure system takes over again, its winds will

Apparently your request hasn't been granted. :-)

The first of the August tempeste is on it's way:

A series of fronts associated with the low will move across the Arctic coast this weekend and early next week.

It well could be the remaining ice in the Chucksi and East-Siberian Seas will be reduced to bits and pieces ...

[Neven]

I mean, AFTER this cyclone (that is forecasted to be huge), Kris. :-P

[Frankd 1977]

Thank you Neven for the insightful graph. It's updates like this that make your blog my favorite.
Another big drop for the SIA to the tune of over 157,600 km2. That makes the four day total (July 31st to Aug 3rd) over 606,000 km2 of ice loss!
For comparison, that's the size of The Ukraine.

[Dave Leaton]

The 60-day CTA daily average is once again over 100k, the 15th such period this melt season (all of 1 such period occurred prior).

[Frankd 1977]

Correction: the 157K loss should be 127K loss (typo) making for a 576,000 km2 four day loss which is the size of Madagascar :)

[Neven]

Thank you Neven for the insightful graph. It's updates like this that make your blog my favorite.

Thanks a lot, Frank. I hope I've explained CAPIE well. It's very handy, especially in this phase of the melting season.

[idunno]

Hi all,

In addition to a low CAPIE, there seems to me to be a difference between large lumps and small.

Imagine a grid of 10x10...

If it has one lump of ice measuring 5x10, the ice concentration is 50%

If it has 50 lumps of ice measuring 1unit squared each, the ice concentration is also 50%.

2 aspects of this are unsatisfactory to my mind.

First the ice in scenario 2 has a much greater surface area. Melt can only really happen on an exposed surface, so surface area is perhaps the most significant measurement. It is at least arguably on a par with volume.

Also, I stuspect that the size of the lumps is also a good proxy measurement for thickness...

Compare for example these two images, looking at the area of the exit of the McClure Strait...

About a month ago, large lumps of ice...

http://lance-modis.eosdis.nasa.gov/imagery/subsets/?subset=Arctic_r04c02.2012179.terra

A couple of days ago, with the ice having presumably declined in thickness, the lumps are much smaller...

http://lance-modis.eosdis.nasa.gov/imagery/subsets/?subset=Arctic_r04c02.2012214.terra

Thinner ice breaks into smaller lumps.

[Kevin McKinney]

"First the ice in scenario 2 has a much greater surface area."

Eh? 50 units^2 versus 50 units^2, surely?

[R. Gates]

Neven (er, Günther),

Were you over messin' with poor Anthony again? You know he's got enough on his hands with is "unprecedented" paper and all, he sure doesn't need you poking another stick in his eye. But did you notice he used a chart that absolutely told the least possible interesting things about the big difference between this melt year and any other year? Certainly not as interesting as your CAPIE, or even just looking at the basic Cryosphere today extent?

But, FWIW, I think the chart to watch over the next few weeks is this Arctic Basin anomaly:

http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.1.html

This represent of course the core of what's left when all the more southerly latitudes melt out over the next few weeks. How quickly the anomaly approaches that key 2.5 million sq. km. mark and how far below it it goes will tell us us a great deal about how what the low mark will be this season for the ice, but also about the health of this core region going forward.

But again, please leave Anthony alone. There's not much happening in the Arctic anyway, and he's got his "unprecedented" paper to complete.

[Seke Rob]

My variation on CAPIE, based on NSIDC monthlies also gave a record for July, at 63.1%. See http://bit.ly/NSIDCC . Computed how much a monthly CAPIE would be of CT-IJIS which produces this comparison.

Metric CAPIE NSIDC
Extent 7,684 7.94 Million km^2
Area--- 5,039 5.01 Million km^2
Concnt 65.6% 63.1%

Was surprised to find them so close together, considering the different methodology on extent between JAXA and NSIDC

[idunno]

Hi Kevin,

No, I'm counting the sides as "surface area" as well. "Exposed surface" might be a better phrase.

50 lumps of 1 have a much greater exposed surface than a single lump of 50.

[Alberto Silva]

Have you noted that NSIDC extent, unlike IJIS, is at a record low?

http://nsidc.org/data/seaice_index/images/daily_images/N_stddev_timeseries.png

It would be interesting to make a CAPNE (Cryosphere-today Area Per NSIDC Extent)and compare it with CAPIE.

[Paul Klemencic]

I agree with dunno completely. The heat transfer is driven by the surface area in contact with water during the "bottom melt" season. In reality, "bottom melt" has become dominated by "side melt". The photos of the ice floes taken by the USCGC Healy research vessel last year showed this really well.

Most ice floes contain underwater shelves, with water above and below the ice shelf. As the floe loses ice mass, the underwater shelves expose an enormous amount of ice surface area to seawater melt. This is how melt rates of ice floes can easily exceed 20 cm daily, and even approach 100 cm per day in storm conditions, versus bottom melt rates of only 1-2 cm daily for stationary pack ice.

[idunno]

Another stab at describing this...

A square metre of a continuous 1 metre thick slab of ice has an exposed 1 m^2 top surface exposed to air, and an exposed 1m^2 bottom surface exposed to water.

A floating cubic metre of ice, which is 90% submerged also has an additional 0.4 m^2 exposed to air, and an extra 3.6m^2 exposed to water.

P.S. WUWT Sea Ice News, referenced by R Gates above, is well worth a visit:

http://wattsupwiththat.com/2012/08/04/sea-ice-news-volume-3-number-9/

[Paul Klemencic]

Here is a picture of smaller ice floes from NOAA library. Notice that while some floes have 80-90% of the floe area above water, many have 80% or more of the floe area below water. In all cases, there is significantly more heat transfer area exposed to the seawater than simple block models.

Larger floes (say 5 km or larger) would have much less ice area below water, but as they break up, the amount of the floe area below the water should increase substantially.

[crandles]

>"It would be interesting to make a CAPNE (Cryosphere-today Area Per NSIDC Extent)and compare it with CAPIE."

If anyone want to do this and would prefer the data in 34 years by 365 rows:

I have collected NSIDC extent, PIOMAS Volume and Cyrosphere today data into

https://docs.google.com/spreadsheet/ccc?key=0AjpGniYbi4andGxEQllJaENMclhnYXBIVENMa0FyQWc#gid=0

I will try and keep that reasonably up to date.

Simply copy data into your favourite spreadsheet, divide area by extent numbers and graph.

NSIDC extent numbers have been going a long time and may well be better than JAXA numbers because of those changes in variability when satelites changes have been made.

Shouldn't be too difficult to graph say 5 year averages 1979-1983, 1984-1988, 1989-1993, 1994-1998, 1999-2003 as well as individual years from 2004 onwards for a longer term view of how things have changed.

Similarly thickness as PIOMAS Volume/CT Area
should also be easy.

I will get around to it eventually if nobody beats me to it. ;-)

[dabize]

Hi Neven,

Just a heads up -I sent you a movie of ice floe movement in the Chukchi in July, where you can see that the ice moves a bit, but stays surprisingly in place, even when storms are about.

Bottom melt in situ FTW..........

[Twemoran]

dabize

Looking forward to the movie. I'd expect the ice to be slowly moving away from low pressure areas in a radial pattern as opposed to following the swirling pattern of the clouds or following the storm track.

Terry

[dabize]

Just sent it to AmWX - you can see it there if you can stand the speed of the GIF!

If Neven posts it, it will be suitably slowed down, I'm sure, but I haven't the technology to do that to a GIF from here.....

[Werther]

Having a holiday in France I have only spent some lurking time last week to keep up. This evening I started with a glance at MODIS. Through all clouds I sensed a change in most patternless floe areas (r03c04, r04c04). As if they were finally melting out. Realising the decline I supposed eight days ago.
CT and UB confirmed it. The final reel seems to have begun for any ice outside the last stand (Arctic Basin 3,8 mkm2). Not yet as dramatic as 200K a day, but getting extent close to 2007. And volume is going even faster.
To enhance the storm into world news (soon), I think most Greenland outlet glaciers show a parallel behaviour. Though I miss my CAD aid out here, I'm pretty sure 79/Zachariae are losing front ice as well as the Humboldt. All of this is indicating warmer SST’s, a feature developing since 2010 and now yielding it’s serious consequences.

[dabize]

Wie Gott im Frankreich, Werther, I'm envious.

I think that idunno has it pegged. CAPIE is essentially a proxy for the reciprocal of the surface area/volume ratio. Almost anything that happens to that ice now (storms, high pressure etc) will involve its movement through water that is warm enough to add significant heat to it. The constant movement only ensures that the boundary layer (where thermal equilibration might occur in a static situation) will be constantly removed, maintaining a maximal thermal gradient between water and ice and thus a maximal rate of melting. The more surface area, the more melting.

Kind of like the how a convection oven works - and how breathing keeps us from asphyxiating.......

Eventually the rate of melting becomes non linear, reaching ruinously high rates (as a percentage of remaining volume) before final meltout occurs.

[Artful Dodger]

Sorry Lads, that's got to be 'NECAP'... ;^)

Seke Rob,

The value of NSIDC Area/Extent for July 2012

( 5,039 / 7,684 = 65.58% )

is in good agreement with my CAP2E value of 66.96% for July 2012.

Note: my CAP2E metric uses a two day average of CT SIA, to better match the IJIS 2-day data averaging practice.

So good job with the NSIDC data. Any chance you will release an annual chart?

Cheers,
Lodger

[FrankD]

CAP2E - nice refinement, Lodger. Tell, me is that prononced "cap-tooey"? Sounds like a man chawin' baccy...

8^P

BTW - Just to head off any confusion in having two FrankD's on the one site (what are the odds?), I should point out that frankd1977 and I are different people.

So, "Hi" to new and improved frankd. For a start, his 'tache is far more impressive than mine. :-)

[Artful Dodger]

FrankD, it's true: if you travel far enough, eventually you will meet yourself ;^)

Since you *asked*, here's the CAP2E money $hot: I've used it along with 78N Insolation to graph 'Solar Gain', with the 100% value normalized to the heat absorbed by ice-bound region of the Arctic Ocean on the solstice, June 20, 2012.

As the Solar Gain chart shows, on Aug 2 the region covered with pack ice received approximately 120% of the solar heat (insolation) that it receive this year on the solstice.

This is the fingerprint of Arctic amplification: the albedo flip leads to longer Summers, and accelerating melt.

[Artful Dodger]

hmm, should have said "pack ice absorbed approximately 120% solar heat compared to the solstice"... ;^)

[k eotw]

I expect NSIDC extent is going to plummet over the next two weeks faster than 2007, reaching 5 million within a week. All that ice over the east siberian sea is disconnected and low concentration which is hitting the area numbers at the moment and the extent numbers soon. After than I suspect the decline will slow down and it'll look increasingly touch and go whether 2012 beats the 2007 minimum.

[Mike Constable]

The size of ice floes will affect melting rates but anything seen on a satellite image will have much more "top & bottom" area than "edge" area, so size of floe will have a very limited (but not zero) effect.
The nearest thing to ice "mush" occurs in the Foxe Basin, and this drifts round in pretty swirls for most of the summer because there is no supply of warm water to melt it from underneath.(The ice there also appears to be a creamy colour, which should make it melt more quickly anyway!).
Just my 2p's worth!
Mike

[Wouterlefebvre]

Neven, I used some of your excellent graphs on a post on Arctic sea ice on the Dutch-speaking weerwoord forum (http://www.weerwoord.be/includes/forum_read.php?id=1659808&tid=1659808)

[Neven]

Thanks, Wouter!

[Seke Rob]

Lodger wrote:

So good job with the NSIDC data. Any chance you will release an annual chart?

. I have an annual chart, I have quarterlies... just have it set up that 1-12 as a chart-dataset var. for month of year, 13 is year and A,B,C,D is the quarters [meteorological] DJF/MAM/JJA/SON, to coincide also with the snow charts. For SSTs I've even gone into allowing a from-to on months, so can for instance run a series that picks Jan-May for all years on record, so there is always a way to have a chart through the latest available data. Here's FTM the http://i137.photobucket.com/albums/q210/Sekerob/Climate/ArcticSIA-SIE%20July.png and NSIDC Annual (through 2011): http://i137.photobucket.com/albums/q210/Sekerob/Climate/ArcticSIA-SIE_Annual.png (Charts originally designed for the Moncktonian clique, who are awe struck with pretty backgrounds, and of course will continue to reject SIA and have no grasp of the water content in extent)

P.S. What is the CT-SIA true days number(?) as in how many days go into the day value to correct for sat disturbances? NSIDC uses 5, and given the very close proximity of 5.01 and 5.039 I'm inclined to think that CT is using a multi-day average methodology (No attempts taken to find their data construction background).

[Wouterlefebvre]

You're the one to be thanked Neven, you made the graphs, I used them.

[Neven]

But it's good to know I'm not making them for nothing. :-)

[Artful Dodger]

Seke Rob asked "What is the CT-SIA true days number(?) as in how many days go into the day value to correct for sat disturbances?"

AFAIK, it's a one day data release. However, for cells with missing observations, the last available observation is used.

So there is no one right answer, which of course is why NSIDC reports daily extent, and monthly area.