Home brew AMSR2 extent & area calculation

[Wipneus]

The ice edge retreat in the Barants/Kara section in an animation. Winds and perhaps currents keep the ice pack compact without much visible (on this scale) open water.

[S.Pansa]

Wipneus, I have a stupid question.

Do you know why there is quite a big gap between the extent from UH AMSR2 3.125 and L3 JAXA AMSR2 10 km since early June (in your extent graph on page 1)?
Before that time the trends were pretty close.
Has this something to do with the greater resolution of UH AMSR2 3.125 and the fact that the ice disperses as the melt progresses (if that is even true)?

Many thanks

Well spotted, not a stupid question at all. It has to do the way Jaxa calculates ice concentration from the AMSR2 microwave measurements, in short the "algorithm".
Jaxa uses an algorithm called Bootstrap, designed by J. Comiso from NASA's Goddard Space Flight Centre. A modification, I think special to Jaxa's calculation, adjusts some parameters during the melting season that compensates for some (but not all) effects due to melting ice surfaces.
That parameter change is fixed on 1st June and ends on 15th October.

The effect of this is I think what you are seeing, the same was visible in 2014.

Resolution has the effect you describe, it is just not as visible in the graphs due to the more gradual increasing effect during the season.

The Jaxa algorithm is described in a document "Descriptions of GCOM-W1 AMSR2
Level 1R and Level 2 Algorithms" available on the GCOM-W1 website https://gcom-w1.jaxa.jp/ (free registration required).

Well that clears things up for me. Thank's a lot - also for the wonderful work you do here and on other threads.

[Wipneus]

An update of the demolition works in the Canadian Archipelago.  Notice that ice around Ellef Ringnes Island is now breaking loose as well.

[iceman]

An update of the demolition works in the Canadian Archipelago.  Notice that ice around Ellef Ringnes Island is now breaking loose as well.

Pretty impressive, though at the same time it's remarkable how much ice has survived the month-long heat wave.  Gives you some idea of how thick and solid it must have become during the abnormally cold winter.  Temps are dropping next week and the wind-driven expulsion of ice into Beaufort will slow or partly reverse.

[plinius]

as discussed in the other thread - the ice itself is not exposed to those high temperature. You can even see in the satellite images that the passages have their own shallow cold layer flowing through them.

[Wipneus]

as discussed in the other thread - the ice itself is not exposed to those high temperature. You can even see in the satellite images that the passages have their own shallow cold layer flowing through them.

The surface melt (ponding) has been extensive as could be seen in MODIS through holes in the cloud cover for a long time. The same can be concluding to some of the worst "torching", dropping of the apparent sea ice concentration (topic of this thread) with the ice cover still solid. That went as far as dropping below 15% concentration, indicating an melt pond concentration of 85% or higher.
I do not doubt that the melting forces on the ice have been ferocious.

[plinius]

Apart from the semantics, I think the discussion is not so much about if there were enhanced melt conditions, but how bad they were. The fog flow through the passage indicates a lot of air motion along the passage and not perpendicular to it, keeping the layer on top and warm air mass over the islands quite nicely separated from the icy channels. Agreed that you still get strong melt ponding, etc., but it makes a difference if you directly hit the ice with 15C airmass, or have mostly downwelling from a clear sky (not cloud cover, the fog is colder/lower down) at 10C.

[Wipneus]

The cyclone over the Beaufort region stirs the ice anti-clock wise concentrating the remaining ice in places with open zones in between. No more hiding for the ice floes deep within the ice pack, the disintegration is everywhere.

[stackmaster]

That looks like a precursor to an extent crash.

[Wipneus]

Stir until it is all gone? Impressive increase of the open water sends the extent downward. Some of the remaining ice gets concentrated in bands, probably not that safe.

(the usual delta image has been augmented with coloring to indicate concentration loss/increase, pinkih indicates over 7% loss, blueish over 7% increase. Solid colors are loss/gain of extent, passing the 15% cut-off)

[Neven]

Cool close-up (first image), Wip.

[Peter Ellis]

Looks like 2006, with that band of persistent MYI near the coast.  Difference is that the water north of the MYI has opened up much earlier this year.  I don't give good odds on the coastal ice surviving the season.

[Shared Humanity]

The most telling aspect of that image is that you can clearly see where concentration of the remaining ice has compacted from the storm but the areas to the outside where this now highly compacted ice is moving towards is either melting out completely towards the shoreline or showing melt towards the pole (the pink shaded areas). There is a lot of warm water doing a lot of damage.

[Siffy]

The most telling aspect of that image is that you can clearly see where concentration of the remaining ice has compacted from the storm but the areas to the outside where this now highly compacted ice is moving towards is either melting out completely towards the shoreline or showing melt towards the pole (the pink shaded areas). There is a lot of warm water doing a lot of damage.

If you look on Modis you can see the ice which shows high concentration is covered by thick cloud as well. I'd expect concentration to actually be lower than shown in the Ubremen conc maps. The ice the in Baeufort is going to melt out very rapidly in the next few days.

[seaicesailor]

There is another interesting feature that appears repeatedly  in the vid. Notice two or three 50km diameter holes at the beginning that quickly expand to become wide areas of open water. Close to the end, another hole or two appear.
Are these warm or salty water plumes, pumped from depth or just centers of diverging streamlines? Or what?

A name for it ...

[oren]

The most telling aspect of that image is that you can clearly see where concentration of the remaining ice has compacted from the storm but the areas to the outside where this now highly compacted ice is moving towards is either melting out completely towards the shoreline or showing melt towards the pole (the pink shaded areas). There is a lot of warm water doing a lot of damage.

If you look on Modis you can see the ice which shows high concentration is covered by thick cloud as well. I'd expect concentration to actually be lower than shown in the Ubremen conc maps. The ice the in Baeufort is going to melt out very rapidly in the next few days.

Agreed, that ice is toast.

[jdallen]

The most telling aspect of that image is that you can clearly see where concentration of the remaining ice has compacted from the storm but the areas to the outside where this now highly compacted ice is moving towards is either melting out completely towards the shoreline or showing melt towards the pole (the pink shaded areas). There is a lot of warm water doing a lot of damage.

If you look on Modis you can see the ice which shows high concentration is covered by thick cloud as well. I'd expect concentration to actually be lower than shown in the Ubremen conc maps. The ice the in Baeufort is going to melt out very rapidly in the next few days.

Agreed, that ice is toast.

For when you are musing over SST's...

Pretty much from what I've seen over the last two years, the melt back of the pack is on auto pilot. Most of the heat which will be applied has been taken up. Much of what continues to be applied is balancing radiative loss, with a little extra left over, declining modestly until Sept 1st, then dropping off sharply. Thus we have good justification for the "Momentum" metaphor.

At about -1C, the melt rate for ice starts at just over 1CM/day, and increases by between 1.5 CM/day per 0.5C increase in temperature.  At 1C, the "burn" rate is about 8CM/day.

The Beaufort ice is being blown over water which easily approaches that.

Even at the lower rate - SST at -0.5C - the burn rate is just shy 3CM/day.  30 days of that will translate into 80CM of melt. 

Now, what do our thicknesses look like?  The action of storms sweeping ice over warmer water or warmer water under ice is now critical to just how much more decline will take place.  Even without them, 2007 and 2011's records are going to be challenged.

[Glenn Tamblyn]

I have just been looking through the compare 2 dates feature at CT, looking at what happened from now onwards in 2012. Fascinating.

A large lobe of low concentration ice remained in the Chukchi/ESS and the CAB immediately above them. This vanished in about 10 days so a big extent drop. This left the ice edge as a line roughly between the western entrance to the NWP and the New Siberian Islands.

And the ice at this edge was of lowish concentration, suggesting that the edge would retreat towards the pole.

It didn't!

Essentially that edge locked in place, concentration actually rose.

All subsequent reductions in total extent were essentially in the arc from the Laptev to the East Greenland sea.

Contrast that with today. The Beaufort/Chukchi/ESS arc and the CAB behind it has significantly more ice. But it is dying. If the thickness of the MYI (and weather conditions) doesn't prevent it, melt back to the 2012 line is still plausible. Maybe not as dramatically as in 2012, but what the heck!

If that happens, a result approaching 2012 is still plausible. It all depends on the weather but the Fat Lady sure hasn't sung yet!

[Bruce Steele]

I posted an animation of the Beaufort nearly two weeks ago- http://forum.arctic-sea-ice.net/index.php/topic,382.msg56893.html#msg56893

Here is the sequel, big floes on the outside of the pack keep getting smaller and eventually disintegration. More inside the pack it is only the smaller floes that can seen visibly melting. Some ice is supplied from the M'Clure Strait.

Nice

Sea Ice Sailor,  If we go back three weeks to when the moth holes first appeared as shown in the animation Wipneus posted you can see flow from the McClure Strait meeting southbound current at
Cape Manning  then moving West towards the center of the Beaufort.  I believe this convergence carried relatively warm surface water out under the pack ice . ITP 85 and it's microcat are located to document this water profile and they show fresh surface water between -1.2 and -.6 for almost a month now. 
 The recent cyclone blew right over these buoys and there is no evidence of salty water. Surface waters have freshened and cooled with a surface cold freshwater lens reforming the last couple days at least that's what happened at the one buoy giving us both water column temperatures as well as salinity profiles.
 

[Wipneus]

I will be on the road today, so no weekly overview today.

In the mean time two animations. North of Greenland something permanent seems to building again (contrast enhanced to bring out the feature).

[Wipneus]

Second animation are the past few days in the Beaufort where ice that seemed to have disappeared reappears again in the last frame. That include the thin sliver of ice between the pack and the Alaskan coast, a sliver that was there in 2014 as well.
The reappearance of the ice is causing an increase in extent in both Jaxa's as well as the one based on Uni Hamburg. Area is hardly affected.

[Andreas T]

I will be on the road today, so no weekly overview today.

In the mean time two animations. North of Greenland something permanent seems to building again (contrast enhanced to bring out the feature).

which Feature do you have in mind_
On MODIS http://1.usa.gov/1Dw9Ksw I see areas of darker ice in two major bands stretching across the north of the Lincoln sea and north east of it. I see also an area where the pack is loosening with some divergence.
I have taken the darkened ice as ice containing algae or sediment of the kind seen by Obuoy9 recently
The second image is from yesterdays MODIS showing part of the dark ice band breaking up in the northeast polynya where Obuoy9 floats
PS actually that is seen a bit further east but  it shows the fairly uniform darkness of the ice irrespective of floe size and the outline which to me shows a pattern of this ice having a common history/ origin with subsequent deformation.

[oren]

I will be on the road today, so no weekly overview today.

In the mean time two animations. North of Greenland something permanent seems to building again (contrast enhanced to bring out the feature).

It's interesting to see the Lincoln see emptying near the entrance to Nares Strait. I still can't decide whether it's export of ice down the Nares on the southbound current, or wind coming up the strait and pushing the ice to the north. What do the experts say?

[seaicesailor]

Second animation are the past few days in the Beaufort where ice that seemed to have disappeared reappears again in the last frame. That include the thin sliver of ice between the pack and the Alaskan coast, a sliver that was there in 2014 as well.
The reappearance of the ice is causing an increase in extent in both Jaxa's as well as the one based on Uni Hamburg. Area is hardly affected.

That's the "peek-a-boo effect"

[BornFromTheVoid]

I will be on the road today, so no weekly overview today.

In the mean time two animations. North of Greenland something permanent seems to building again (contrast enhanced to bring out the feature).

It's interesting to see the Lincoln see emptying near the entrance to Nares Strait. I still can't decide whether it's export of ice down the Nares on the southbound current, or wind coming up the strait and pushing the ice to the north. What do the experts say?

A bit of both it seems. There has certainly been some southerly winds blowing up through Nares in recent days.

10m winds for yesterday



ice drift data for the 3rd and 4th.

[Neven]

That's the "peek-a-boo effect"

 

[Wipneus]

I was asked (paraphrased) why my graphs only include the last 4 years and not 2007 a much more interesting year than 2013.
Not discussing whether 2013 is interesting or not, the fact is that AMSR2 based graphs are only available since 2012 for the most basal reason. Also no proper homogeneous SIC series exist that use the AMSR2 for the last 4 years. Reason is (in short) that the problem is difficult and without it a whole bunch of apple and orange issues exist.

Thet is not true for the NSIDC data, actually the NSIDC data is the only useful homogeneous series that I know of. There are very promising data from OSISAF, but they are experimental and I have yet to look at them.

In the mean time I have added 2007 to the "basin only" graphs. 



LINK



LINK

[oren]

Thank you Wipneus.

[Yuha]

Second animation are the past few days in the Beaufort where ice that seemed to have disappeared reappears again in the last frame. That include the thin sliver of ice between the pack and the Alaskan coast, a sliver that was there in 2014 as well.
The reappearance of the ice is causing an increase in extent in both Jaxa's as well as the one based on Uni Hamburg. Area is hardly affected.

That's the "peek-a-boo effect"

This might be what the peek-a-boo ice looks like:

[anotheramethyst]

Thank you Wipneus.

i agree, and also that was a very elegant solution to the apples/oranges problem

[greatdying2]

Interesting difference in the CAB in recent days between JAXA and UH. Is this due to the grid size difference? 3.125 km grid increasing rapidly, 10 km grid plummeting, but converging in value. Can someone explain it?

[Wipneus]

Interesting difference in the CAB in recent days between JAXA and UH. Is this due to the grid size difference? 3.125 km grid increasing rapidly, 10 km grid plummeting, but converging in value. Can someone explain it?

You are showing area. Grid size has only a modest influence on computing area compared with extent. I suspect the different sensitivities of the algorithms for marginal ice compared to consolidated ice is a factor here. The effectivity  of the weather filters is another, given the passage of cyclones.

[greatdying2]

Interesting difference in the CAB in recent days between JAXA and UH. Is this due to the grid size difference? 3.125 km grid increasing rapidly, 10 km grid plummeting, but converging in value. Can someone explain it?

You are showing area. Grid size has only a modest influence on computing area compared with extent. I suspect the different sensitivities of the algorithms for marginal ice compared to consolidated ice is a factor here. The effectivity  of the weather filters is another, given the passage of cyclones.

Hmmm... ok. Not sure if that makes it less interesting, or more. Thanks.

[greatdying2]

Ok, sudden correction of the UH trend -- I guess it was some sort of artifact, so not very interesting.

[Wipneus]

Due to circumstances I have no time to do one of my weekly overviews today. Perhaps on an other day, but no promise.

The CAA is dropping in area and extent fast. The ice in all branches is mobile now, they are open to import and export of ice. One of the more southern alternatives of the NW passage is on the brink of opening (as seen by these satellite derived images).

[Wipneus]

Much of the extent decline in the last few days is due to the retreat of the ice edge in the Laptev section.

[slow wing]

Thanks Wipneus, these gifs are great!

As well, the ice pack has suddenly gotten full of holes in the region shown below the Laptev Sea in the direction of the North Pole. It will be fascinating to see how that develops!

[Wipneus]

Continuing with the CAA. The main NWP channel is getting cleared of ice. Ice, presumably of the thick MYI variety, can be seen entering the channels from the Arctic Basin.

[AmbiValent]

To me it seems like only Peary Channel is seeing actual CAB ice moving in, while other CAB ice floats by keeping its distance due to the telltale gap. The ice moving into the other northern CAA openings seems to be CAA ice that broke off, drifted a bit north and now came back.

[cats]

Another nice animation - thanks!!  It also looks like some ice from the Lincoln Sea is making its way south thru the Nares Strait

[Wipneus]

The combination of lots of marginal ice with a strong cyclone is nothing less than spectacular. Some individual ice floes can be seen speeding across the Beaufort Sea.
The effect of the weather filter can be seen in action: some of the details are blurred due to the use of lower frequency microwave bands that have a lower resolution (wider field of view) than the 89GHz channels. Also some of the immediate extent loss is restored later: the delta image shows that the "arm" grew again in the latest image. I expect the same to happen for some of yesterday's loss in the main pack.

[Wipneus]

Late summer expansion of the Laptev open water. Most of this summer large scale ice movements where absent. The detached bit of ice can be seen melting in place. There may be some melting involved in the retreat of the main pack's edge, but it is clearly overwhelmed by transport by winds and currents.

(needs a click to start)

[seaice.de]

Animation of AMSR2 sea ice data from May to August:

[slow wing]

Nice animation! Thanks... 

[seaicesailor]

Animation of AMSR2 sea ice data from May to August:

That's awesome!

[Pmt111500]

Thanks for the animation,

[iceman]

« Reply #1721 on: July 18, 2015, 01:15:24 PM »

   ....
It looks like some stubborn ice on the coasts of Baffin and in the Foxe Basin (north of the Hudson Bay) will hold out for a while.
   ....

Good call.  I would not have expected ice south of the Arctic Circle to hang on by its frigid fingertips through the end of the melt season.

[Neven]

Animation of AMSR2 sea ice data from May to August:

Awesome animation! What an amazing contrast between Pacific and Atlantic side.

Will this be put on Youtube, or should I upload it?

[Neven]

I've taken the liberty of uploading the video to YouTube, as I wanted to use it for my latest blog post. Hope that's okay!

[Wipneus]

With high pressure over the central arctic, I was curious about possible large scale movements visible in the UH AMSR2 concentration maps. With ice concentrations above 95% almost everywhere, the contrast had to be increased considerable: dark grey's do not mean that the ice is melting!

Spinning wheel (when clicked) , with in some corners some fast ice.