Home brew AMSR2 extent & area calculation

[JimboOmega]

And the surface melting images. The "melt pond extent" has increased again.

Stupid question - but what do the colors on the image mean? And how directly does that translate to melt pond extent

[Wipneus]

Area drop-rate took a bit hit and extent continued its rather slow (for June) decline. 2012 is now only days behind 2016.

Update 20160610.

Extent: -44.5 (-489k vs 2015, -639k vs 2014, -990k vs 2013, -344k vs 2012)
Area: -10.1 (-676k vs 2015, -577k vs 2014, -865k vs 2013, -186k vs 2012)
 
You will find the updated graphs in the top post

By extent Baffin and Hudson declined most: -29k and -21k. Extent increased in the Greenland Sea (+13k) anhd CAA (+12k).

By area again Baffin and Hudson declined most: -30k and -34k. The area increased in Laptev (+32k) and ESS (+29k).

Today the details of Baffin/Newfoundland Bay region. The decline in ice cover seems to be a fast track comparable to 2012. 

[Wipneus]

An animation of Beaufort (again). Ice floes are escaping the center of the gyre into the open water where the are moving in several directions. 

[Wipneus]

The surface melting situation. As expected from the area up-ticks in some regions, the extent in melt (as indicated by ADS/Jaxa map) is decreased.

[Wipneus]

And the surface melting images. The "melt pond extent" has increased again.

Stupid question - but what do the colors on the image mean? And how directly does that translate to melt pond extent

Jim, attached is the legend that is lost when I prepare the image in a standard format. Ice thickness is not reliable in this season due to melting. Melting obviously should be of use now.
 
As discussed previously, I presume "ice melt concentration" is the same as "melt pond concentration". I do not know how useful this data is, regard it as research.


 

[Robert Greer]

That Beaufort animation is pretty frightening. Highly-mobile ice gets "shaken" loose into a bunch of floes sitting in open water. Who needs melt ponds when you can just have interstitial open water sucking up all that insolation?

And the remaining ice in the Beaufort is supposedly some of the thicker ice in the Arctic right now. Doesn't reassure me about the CAB's prospects in the coming months.

[Richard Rathbone]

And the surface melting images. The "melt pond extent" has increased again.

Stupid question - but what do the colors on the image mean? And how directly does that translate to melt pond extent

Jim, attached is the legend that is lost when I prepare the image in a standard format. Ice thickness is not reliable in this season due to melting. Melting obviously should be of use now.
 
As discussed previously, I presume "ice melt concentration" is the same as "melt pond concentration". I do not know how useful this data is, regard it as research.

The average should be decent once the melt pond concentration is high enough but there's a lot of scatter in the correlation between the actual melt pond concentration and the satellite measured brightness temperature index its correlated with so if you aren't averaging over a decent sample size the error bars are large, and if the average concentration is near 0 there's a significant high bias.

[Buddy]

And the remaining ice in the Beaufort is supposedly some of the thicker ice in the Arctic right now. Doesn't reassure me about the CAB's prospects in the coming months.

1)  Temps REMAIN warmer than normal in the Arctic above 80 degrees north latitude.
2)  SST anomalies continue to be warmer than normal
3)  Ice appears to be in much worse shape than 2012 at this date.

While many have thought that the ice melt has "petered out"  this year because of the last 10 days or so.....I still see an ice sheet headed for a new record low in September (unfortunately).

I think "the plunge"....while coming later than I thought....is still coming...and will come to a chewed up ice sheet waiting to be taken out.

3 months of ice melt ahead.....

[Feeltheburn]

The surface melting situation. As expected from the area up-ticks in some regions, the extent in melt (as indicated by ADS/Jaxa map) is decreased.

I was interested in Neven's predictions of warm water entering the Bering Strait and finishing off the ice in Beaufort.  There was even a time when there was continuous open water along the northern coast of Alaska and NW Territories.  But your animation shows the ice has now closed up and for some time now I haven't seen much incursion of open water on the Arctic side of Bering Straight.  What do you make of that?

[Neven]

The surface melting situation. As expected from the area up-ticks in some regions, the extent in melt (as indicated by ADS/Jaxa map) is decreased.

I was interested in Neven's predictions of warm water entering the Bering Strait and finishing off the ice in Beaufort.  There was even a time when there was continuous open water along the northern coast of Alaska and NW Territories.  But your animation shows the ice has now closed up and for some time now I haven't seen much incursion of open water on the Arctic side of Bering Straight.  What do you make of that?

Welcome, Feeltheburn (what's with the Bernie Sanders pic?  ). I didn't predict warm water to finish off the ice near the coasts of the Beaufort Sea. I expected just enough to wind to push it away, but the winds turned. Then I thought that maybe higher air temps and solar radiation would do the trick, but the wind picked up some more (check out my latest blog post on the ASIB).

So now I just wait and see what the Arctic will do with that ice and how soon things open up. Remember, it wasn't until the first week of July that the coast was entirely clear of ice on the Uni Bremen sea ice concentration map, in 2009 and 2011.

[Shared Humanity]

Hasn't there been research that shows the water entering into the Arctic from the Pacific takes so long that it loses much of its heat?

[Sleepy]

This study from last week might be worth following:
http://forum.arctic-sea-ice.net/index.php/topic,1454.msg79434.html#msg79434
The CP-nino region mentioned (a bit larger than the Nino 4) will probably not cool off any time soon.

[Wipneus]

Belated report of my extent and area calculations (based on UH AMSR2 hi-res sea ice concentration data).

Also this data sees an uptick in area, bringing 2016 area's lead to 2012 back to nearly nothing. Extent is also nearly the same as in 2012, may take a day or two to loose the lead.

Update 20160611.

Extent: -21.4 (-468k vs 2015, -601k vs 2014, -992k vs 2013, -226k vs 2012)
Area: +49.2 (-587k vs 2015, -509k vs 2014, -777k vs 2013, -75k vs 2012)
 
You will find the updated graphs in the top post

The area increase is explained by the ESS: +47k. Also Greenland Sea SIA gained: +28k. The Baffin region keeps its steady path and lost -17k.
Extent  in Baffin lost -18k, gained +19 in the Greenland Sea.

Attached the delta map of Beaufort. Red and blue colors shows the movement of the ice floes from day to day clearly with lots of open water around it.

[Wipneus]

The animations shows ESS and Laptev. Melt and restoration are changing daily. The last few days melting is hitting the fast ice in Laptev hard. The pitch black ice in the last frame has less than 15% concentration.

(needs a click to start)

[Wipneus]

The melting area has increased a negligible amount. Especially in the central basin, melting is back below the threshold.

[Feeltheburn]

Welcome, Feeltheburn (what's with the Bernie Sanders pic?  ). I didn't predict warm water to finish off the ice near the coasts of the Beaufort Sea. I expected just enough to wind to push it away, but the winds turned. Then I thought that maybe higher air temps and solar radiation would do the trick, but the wind picked up some more (check out my latest blog post on the ASIB).

Neven, sorry if I misstated what you had said.  As I indicated I have been a lurker and therefore probably lose some of the nuance and also probably form an incorrect impression.

I guess my picture and name are a double entendre?  I nabbed the picture from the internet!

[Wipneus]

Update 20160612.

Extent: -44.3 (-414k vs 2015, -582k vs 2014, -979k vs 2013, -110k vs 2012)
Area: -120.4 (-601k vs 2015, -552k vs 2014, -843k vs 2013, -135k vs 2012)
 
You will find the updated graphs in the top post

The century area drop is enough to just stay ahead of 2012. The lead in extent with its slower decline has now nearly evaporated.

Regional extent declined in Kara (-16k), Greenland Sea (-16k) and Beaufort (-12k). Extent in the CAB increased by +14k.
Area lost -31k in the Greenland Sea, -22k in Hudson and -21k in the Beaufort. Smaller declines in ESS, Laptev and Kara (-17k, -13k and -13k).

Today's highlighted region is the Greenland Sea. The colorful composition in red and blue is a area with ice concentration hovering around the 15%.

[Wipneus]

Animated Kara sea ice concentration. The decline is obvious.

[Wipneus]

The surface melting situation. "melt ice extent" has hardly changed, some reduced in ESS, some gained in Chukchi and Beaufort.

[binntho]

Today's highlighted region is the Greenland Sea. The colorful composition in red and blue is a area with ice concentration hovering around the 15%.

I'd like to "unlurk" to thank Wipneus for all his fantastic images and graphs. And to point out that Iceland has not had sea ice along its shores in June for at least some centuries (except for the occasional large iceberg in late summer, calvings from the great Greenland glaciers).

In winter, drift ice from the far north occasionally settles in the fjords in the north-west, the last time this happened was in January 2007 (as far as I know).

The changes in sea ice around Iceland in Wipneus' image  is perhaps a reminder that satellites are not perfect?

[Wipneus]

Today's highlighted region is the Greenland Sea. The colorful composition in red and blue is a area with ice concentration hovering around the 15%.

I'd like to "unlurk" to thank Wipneus for all his fantastic images and graphs. And to point out that Iceland has not had sea ice along its shores in June for at least some centuries (except for the occasional large iceberg in late summer, calvings from the great Greenland glaciers).

In winter, drift ice from the far north occasionally settles in the fjords in the north-west, the last time this happened was in January 2007 (as far as I know).

The changes in sea ice around Iceland in Wipneus' image  is perhaps a reminder that satellites are not perfect?

The satellites are what they are. The ice around Iceland (and other coasts) is not counted (by my computation) in the area and extent figures as they are. A, what I call "coastal ice filter", gets rid of most of it.

[Wipneus]

2016 is still ahead of 2012.

Update 20160613.

Extent: -58.4 (-372k vs 2015, -581k vs 2014, -934k vs 2013, -98k vs 2012)
Area: -119.2 (-604k vs 2015, -610k vs 2014, -790k vs 2013, -176k vs 2012)
 
You will find the updated graphs in the top post

Most of the extent drop are in Kara and Hudson (-23k and -21k), reliable decliners.
Area, as always very volatile, went down especially strong in CAA (-51k), across the Arctic Laptev  area increased by +33k. The CAB, Kara, Baffin, Hudson and Chukchi declined some -20k an average.

Today's regional delta map is from the CAA , parts of it seem to be baking in the summer.

[Wipneus]

An animation of the declining ice in the Hudson region.

[Wipneus]

And the extent of "melting ice" situation. Little overall change, some shift from Chukchi to Beaufort. Almost nothing in the central basin, but the melting in the CAA has deepened somewhat.

[Lord M Vader]

Wipneus: looking at CCI_Reanalyzer, forecasts are calling for the warm conditions to persist over CAA, Kara and Hudson the next 5 days. By the way, you can add some rain onto the ice in CAA and ESS.

The ice in Hudson and Kara should see a fairly big drop during the next 7 days. The same might be true for Baffin as quite warm conditions will be there too.

[ChrisReynolds]

Wipneus,

As much as I appreciate the AMSR2 data you produce I'm finding the short time scale available frustrating. So I've just updated from your NSIDC concentration based data. I note that the massive errors in April have stopped. Is the source concentration data now from the uncalibrated F18 stream?

[Wipneus]

Wipneus,

As much as I appreciate the AMSR2 data you produce I'm finding the short time scale available frustrating. So I've just updated from your NSIDC concentration based data. I note that the massive errors in April have stopped. Is the source concentration data now from the uncalibrated F18 stream?

They are from now calibrated f18 data.

http://nsidc.org/arcticseaicenews/2016/06/satellite-data-transition-complete/

[Wipneus]

2016 manages to stay ahead of 2012 (in the race to the bottom).

Update 20160614.

Extent: -91.3 (-386k vs 2015, -637k vs 2014, -982k vs 2013, -95k vs 2012)
Area: -77.8 (-630k vs 2015, -646k vs 2014, -797k vs 2013, -122k vs 2012)
 
You will find the updated graphs in the top post

Kara, Hudson and Baffin took care of the drop in extent (-35k, -22k, -19k).

The same three declined significant as well (-31k, -18k, -16k), and were helped by CAA (-23k). Ice area did inrease in Chukchi and ESS (+19k,+12k).

Today's delta map is from ESS and Laptev. The darkest shades are getting brighter.

[Wipneus]

Animation of the Nares Strait. A vague outline of the upper arch in the Lincoln Bay is visible in the last frame.

[Wipneus]

The "melting extent" did not change since yesterday. Yet from in the animation you can see that the melt of the ice in the CAA and Laptev is intensifying (darker shades of blue).

[Shared Humanity]

Looking at these animations, it would not surprise me if the CAA is virtually ice free by the end of the melt season.

[Richard Rathbone]

Looking at the graph on page 1, 2012 extent seems lower than 2016, not higher. How do you get it as still being higher? Surely once the 10km footprint calculation separates from the 2.5km one, the 12.5km value has to be at least as high as the 10km one, not with the 2.5km one? .

[Peter Ellis]

The 10k and 2.5k are from the same satellite - the 12.5k is a different satellite.

On both like-with-like comparisons (NSIDC and JAXA), 2016 is still somewhat below 2012.

[Wipneus]

Looking at the graph on page 1, 2012 extent seems lower than 2016, not higher. How do you get it as still being higher? Surely once the 10km footprint calculation separates from the 2.5km one, the 12.5km value has to be at least as high as the 10km one, not with the 2.5km one? .

Attached is a blow-up of the extent graph. The black line (2016) is still under the orange line (2012).
I did fix a one-day off error in this graph (leap day problem) yesterday, so I am surprised you are  still seeing it.

The Jaxa (10km) separates because of their change in algorithm from dry ice to wet ice. Starts 1 June and takes about 10 days.

I choose to show the 12.5 km SSMIS ASI calculation for 2012, because it has excellent agreement during 2013 and 2014 as well. The 3.125 km or 12.5 km do not seem to matter that much here.

[ChrisReynolds]

Wipneus,

As much as I appreciate the AMSR2 data you produce I'm finding the short time scale available frustrating. So I've just updated from your NSIDC concentration based data. I note that the massive errors in April have stopped. Is the source concentration data now from the uncalibrated F18 stream?

They are from now calibrated f18 data.

http://nsidc.org/arcticseaicenews/2016/06/satellite-data-transition-complete/

That's such good news, and far quicker than I'd thought it would be. The previous assessment suggested it could be months.

Thanks, I'm off to update what I updated yesterday.

[Wipneus]

The numbers indicate a very slow day and 2012 taking the lead. However the image (attached) shows that there are problems, mostly on the Atlantic side.
So I skip discussion of numbers and animations today, unless an interim update will arrive (did happen before).

[Wipneus]

ADS/Jaxa melting data appears to be unaffected. Surface melting extent went down and is now below 2013-2015 for the day of the year. No "intensifying" either.

[slow wing]

Thanks Wipneus, I really appreciate these plots.

Question: why do the high thickness areas dance around so much from day to day on the ice melt plot? Is the day to day thickness measurement mostly noise then?

(Reminds me of those cosmic microwave background measurement plots that showed mostly noise but (reasonably) won a Nobel Prize anyway.)

[Wipneus]

The short answer is that the thickness is not reliable in the presence of melt ponds:

The product of sea ice thickness and the melt pond concentration which is shown in the VISHOP is calculated from AMSR-E and AMSR2 data by using a research algorithm1) developed by K. Tateyama (Kitami Institute of Technology) and others in the Arctic research projects utilizing the IARC-JAXA Information System (IJIS) and satellite imagery (1st - 4th generations). This product is an essentially experimental and research product. This product has the effectiveness in the relative dry freezing seasons such as autumn, winter and spring (September – May), but cannot provide the accurate sea ice thickness in melting wet season (June - August) because the sea ice surface is covered by melt ponds.

See: https://ads.nipr.ac.jp/vishop/vishop-extent.html?N

There is of course a longer answer but that would require  you to have a look at:

Krishfield, R. A., A. Proshutinsky, K.Tateyama, W. J. Williams, E. C. Carmack, F. A. McLaughlin, and M.-L. Timmermans (2014), Deterioration of perennial sea ice in the Beaufort Gyre from 2003to 2012 and its impact on he oceanic freshwater cycle, J. Geophys. Res.Oceans, 119, doi:10.1002/2013JC008999.

You can see that thickness is derived from the polarization ratio of the emissions in the 36GHz band. Thickness has a very nonlinear relation to the said ratio. That means slight deviations from the model have a big influence on the result. Think melt ponds, but also ice with large thickness variation will not give you a good estimate of the average thickness.

As they say in the link above:

This product is opened to the public for the usages of research and validation of algorithms.

[slow wing]

Thanks Wipneus for the informative answer.

Certainly not a criticism and appreciate K. Tateyama et al. providing the tool - it's very helpful in its intended purpose of informing us where the surface melt is taking place.

[Wipneus]

Today's sea ice concentration seems to be OK. Since I still not trust yesterday's data, I would not pay too much attention to the one-day difference reported.
What we can see is that the totla figures are now very  close (just above) 2012. There is a couple of days of slow 2012 declines ahead, so 2016 may take the lead again.Temporary or not.

Update 20160616.

Extent: -114.8 (-419k vs 2015, -589k vs 2014, -972k vs 2013, +22k vs 2012)
Area: -1.2 (-520k vs 2015, -515k vs 2014, -579k vs 2013, +50k vs 2012)
 
You will find the updated graphs in the top post

In case you still wondered where the century extent drop came from, it is Kara that declined -74k.

[Wipneus]

In the animation of the Greenland Ice, the fast ice can be seen getting loose.

[Wipneus]

The surface melting situation based on ADS/Jaxa data. Only a little increase today. Some of that is near the Fram Strait.

[Neven]

Thanks again for these melt pond extent maps and graph, Wip. I've used the latter in my latest ASIB blog post (made an animation of the former myself): 2016 melting momentum, part 1

Here's the animation, showing changes since June 1st:

[Rob Dekker]

In the animation of the Greenland Ice, the fast ice can be seen getting loose.

Thanks Wipneus.
Your animation suggests that ice being pushed into Fram strait and the entire area North of Svalbard is melting out almost as fast as it is being pushed in.

[Wipneus]

The declines in extent and area are enough to take over 2012 again.

Update 20160617.

Extent: -98.6 (-480k vs 2015, -690k vs 2014, -991k vs 2013, -87k vs 2012)
Area: -71.3 (-525k vs 2015, -632k vs 2014, -528k vs 2013, -18k vs 2012)
 
You will find the updated graphs in the top post

Extent dropped in Kara, CAB, Baffin and CAA: -27k, -17k, -17k and -15k.
Area dropped in Kara, Baffin and ESS: -21k, -18k and -14k.

The regional delta map is from the Canadian Archipelago. The ice concentration apparently drops below 15% in some places in the main channel. I don't know how realistic that is.

[Wipneus]

The ice in Kara is disappearing fast. I imagine this is what it will look like in the central pack when the time has come.

[Wipneus]

The state of surface melting. The total "melt pond extent" (from ADS/Jaxa AMSR2 data) did increase. From the animation we see that this did happen almost everywhere except in ESS and Laptev where the melting lessened. Also the intensity (melt pond concentration) intensified in many places, the same spot in the CAA that I noticed above is deep blue here.

[Shared Humanity]

With the outrageous SST anomalies and the explosion of melt ponds in this animation, I am keeping my eye on the Atlantic side of the Arctic.

[Wipneus]

Slow day both extent and area, yet the lead with 2012 remains, gets a bit bigger even.

Update 20160618.

Extent: -34.0 (-492k vs 2015, -642k vs 2014, -985k vs 2013, -152k vs 2012)
Area: -21.0 (-420k vs 2015, -511k vs 2014, -632k vs 2013, -94k vs 2012)
 
You will find the updated graphs in the top post

The only region where extent drooped significantly is Greenland Sea (-25k).
The area measures are more in a flux, declines in Hudson, Greenland Sea and CAB (-28k, -28k and -19k). The Laptev and ESS increased in area (+39k and +13k).

Since NSIDC and Jaxa data are missing, I attach the whole delta map. Notice the concentration drop over the Lincoln Sea  and the (apparent) cold on the Asian side.

(no surface melting reported today, the ADS server seems to be in maintenance)