Support the Arctic Sea Ice Forum and Blog

Author Topic: Home brew AMSR2 extent & area calculation  (Read 1974143 times)

epiphyte

  • Frazil ice
  • Posts: 387
    • View Profile
  • Liked: 22
  • Likes Given: 21
Re: Home brew AMSR2 extent & area calculation
« Reply #2200 on: June 30, 2016, 09:38:56 AM »
Has anyone else noticed that of late the (IMO fatally simplistic) "extent" horserace always seems to converge to a dead-heat on or about July 16th?


Flocke

  • New ice
  • Posts: 20
    • View Profile
  • Liked: 1
  • Likes Given: 1
Re: Home brew AMSR2 extent & area calculation
« Reply #2201 on: June 30, 2016, 09:42:48 AM »
For a key to the colors of the melt extent map, see http://forum.arctic-sea-ice.net/index.php/topic,382.msg80472.html#msg80472 (hint ... ;-)

Darvince

  • Frazil ice
  • Posts: 318
    • View Profile
  • Liked: 31
  • Likes Given: 7
Re: Home brew AMSR2 extent & area calculation
« Reply #2202 on: June 30, 2016, 12:14:31 PM »
Has anyone else noticed that of late the (IMO fatally simplistic) "extent" horserace always seems to converge to a dead-heat on or about July 16th?
This is likely only an artifact of the small size of the dataset for that date, with only years 2013-2015.

jdallen

  • Young ice
  • Posts: 3410
    • View Profile
  • Liked: 650
  • Likes Given: 244
Re: Home brew AMSR2 extent & area calculation
« Reply #2203 on: June 30, 2016, 12:17:32 PM »
Has anyone else noticed that of late the (IMO fatally simplistic) "extent" horserace always seems to converge to a dead-heat on or about July 16th?
No surprise - it's around then I expect that most of the ice in peripheral seas and lower latitudes has melted out.
This space for Rent.

Tor Bejnar

  • Young ice
  • Posts: 4606
    • View Profile
  • Liked: 879
  • Likes Given: 826
Re: Home brew AMSR2 extent & area calculation
« Reply #2204 on: June 30, 2016, 02:58:32 PM »
Referencing Wipneus' Reply #2201 above

I find it interesting that A-Team's big floe in the Beaufort Sea that looks as solid as a rock [Aside: to a geologist, natural ice makes a monomineralic rock.] in 'true color' images and shows as having 70-90% concentration in others (attached) due to melt ponds (I'm sure). (attached is an extract from one of Wipneus' images)
Arctic ice is healthy for children and other living things because "we cannot negotiate with the melting point of ice"

Shared Humanity

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2205 on: June 30, 2016, 03:15:55 PM »
One persistent feature of this years melt season is the lack of transport through the Fram, movement of the ice towards the Atlantic overall is low. The low extent in the Greenland Sea is one result, as is the ice edge so close to the pole on the Atlantic side. This is good news for ice retention overall. I would pay very close attention as to whether a high forms over Greenland with continued lows over the CAB. We do not want Fram transport to take off. With the ridiculously warm waters on the Atlantic side, the ice would melt quickly.

etienne

  • Nilas ice
  • Posts: 2045
    • View Profile
    • About energy
  • Liked: 309
  • Likes Given: 23
Re: Home brew AMSR2 extent & area calculation
« Reply #2206 on: June 30, 2016, 03:31:39 PM »
Hello,

I updated my Sea Ice Extend average graph and added below a second graph showing the difference between the trend and the average value.


We are now almost exactly on the trend, but since 2016 is quite lower than 2015, the average will go down faster than the trend.

I got no reaction on this graph, should I make a monthly update ?

Thanks, best regards,

Etienne

Maybe I should explain a little bit where the idea of the graph comes from. I am maintaining a facility that is heated by the heat of an IT room. In 2008, I started to make precise electricity consumption analysis, and daily averages was something very useful because it allowed me to compare consumption on days with sun, with clouds… this allowed me for example to see that the lowest energy use was with an average outside temperature of 5°C. The problem I had was to check if optimisations where helpful or not, and for this, I needed longer averages, but averages on 3-4 weeks are not acceptable because for example May and June don’t have the same climatic conditions, so I made yearly average which allowed me to have a weather independent measurement of our electricity need. I didn’t care for a trend line because normal consumption is flat, optimisations bring the average down, and load increase brings the average up. The issue was not to save energy, because more energy use could mean better business, but to avoid the waste of energy. The main objectives are to see if we are above or below average, and if average goes up or down, and for both cases I have to check if this is normal or not.

In the case of Sea Ice Extend, normal evolution is a yearly reduction of Sea Ice Extend (about 60’000 km^2 per year), so you can not compare the average with a flat line. It is important to find out what the trend line is in order to be able to check if we are above the trend (low melting year) or below the trend (high melting year).  This is why I made the second graph which only shows the difference between the extend and the trend. If my graphs are right, we are now just on the trend, so we are a normal melting year, the main issue is that a year with such a low extend can be normal.

The problem of this presentation is that, in order to avoid seasonal variations, averages are done on 365 days, which mean that it is more a long term analysis tool than for short variations.
« Last Edit: June 30, 2016, 08:43:10 PM by etienne »

magnamentis

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2207 on: June 30, 2016, 03:44:02 PM »
One persistent feature of this years melt season is the lack of transport through the Fram, movement of the ice towards the Atlantic overall is low. The low extent in the Greenland Sea is one result, as is the ice edge so close to the pole on the Atlantic side. This is good news for ice retention overall. I would pay very close attention as to whether a high forms over Greenland with continued lows over the CAB. We do not want Fram transport to take off. With the ridiculously warm waters on the Atlantic side, the ice would melt quickly.

once there will be zero ice, export through the fram will cease :-)

what sounds like self-evident or like sarcasm is meant to remind users that there are currently 2 additional reasons why there is little export.

a) the ice that would usually be exported is and has been melting almost throughout the entire winter north of svalbard and adjacent regions.

b) the total amount of ice that could be exported is very reduced and that reduced amount of course will melt
again before reaching fram and in the wake count as export

c) both of the above factors are boosted by the fact that the ice that would normally be exported is much thinner and more fragmented due to lower temps in winter.

d) yes there remains the wind and circulation patterns that are currently not in favour of export, that's the
part which has been there before while a-c are a bit on the exceptional side this year.

BenB

  • Frazil ice
  • Posts: 283
    • View Profile
  • Liked: 51
  • Likes Given: 13
Re: Home brew AMSR2 extent & area calculation
« Reply #2208 on: June 30, 2016, 03:49:01 PM »
One persistent feature of this years melt season is the lack of transport through the Fram, movement of the ice towards the Atlantic overall is low. The low extent in the Greenland Sea is one result, as is the ice edge so close to the pole on the Atlantic side. This is good news for ice retention overall. I would pay very close attention as to whether a high forms over Greenland with continued lows over the CAB. We do not want Fram transport to take off. With the ridiculously warm waters on the Atlantic side, the ice would melt quickly.

I'm not sure I agree with that. There was quite a lot of transport in the early melt season, which is why it was so striking that Greenland Sea extent fell so quickly and remained extremely low in spite of that.

BenB

  • Frazil ice
  • Posts: 283
    • View Profile
  • Liked: 51
  • Likes Given: 13
Re: Home brew AMSR2 extent & area calculation
« Reply #2209 on: June 30, 2016, 04:23:42 PM »
Just as a follow up, on 6 May Neven wrote in a comment to his blog post:

Extremely warm/not-cold winter, relatively thin ice on the Pacific side of the Arctic, massive ice movement during the transition period when open water no longer freezes up to a meaningful thickness, continued high temps, high export through Fram Strait, extremely rapid loss of snow cover (there's a correlation there as Rob Dekker has shown through the years, something which Dr. Andrew Slater from the NSIDC is also working on).

And in his blog on 27 May he wrote:

Given the high pressure over the Beaufort Sea, the Beaufort Gyre has been doing its thing for a couple of weeks now, which means ice on the Atlantic side of the Arctic gets pushed into the Greenland Sea, via Fram Strait.

So I don't think that I'm imagining that there was quite high Fram export near the start of the melt season.

bbr2314

  • Nilas ice
  • Posts: 1817
    • View Profile
  • Liked: 158
  • Likes Given: 53
Re: Home brew AMSR2 extent & area calculation
« Reply #2210 on: June 30, 2016, 07:06:56 PM »
Just as a follow up, on 6 May Neven wrote in a comment to his blog post:

Extremely warm/not-cold winter, relatively thin ice on the Pacific side of the Arctic, massive ice movement during the transition period when open water no longer freezes up to a meaningful thickness, continued high temps, high export through Fram Strait, extremely rapid loss of snow cover (there's a correlation there as Rob Dekker has shown through the years, something which Dr. Andrew Slater from the NSIDC is also working on).

And in his blog on 27 May he wrote:

Given the high pressure over the Beaufort Sea, the Beaufort Gyre has been doing its thing for a couple of weeks now, which means ice on the Atlantic side of the Arctic gets pushed into the Greenland Sea, via Fram Strait.

So I don't think that I'm imagining that there was quite high Fram export near the start of the melt season.
You are not, but the high export has met equally obscene temps, so exported ice has disappeared.

The main evidence of significant recent export is the mobility of the CAB; the key isn't that the entirety has been pushed to FRAM/the ATL, it's that there has been enough to allow the pack to lose structural integrity and break apart in the wake of the ATL/FRAM export.

seaicesailor

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2211 on: July 01, 2016, 12:20:26 AM »
Just as a follow up, on 6 May Neven wrote in a comment to his blog post:

Extremely warm/not-cold winter, relatively thin ice on the Pacific side of the Arctic, massive ice movement during the transition period when open water no longer freezes up to a meaningful thickness, continued high temps, high export through Fram Strait, extremely rapid loss of snow cover (there's a correlation there as Rob Dekker has shown through the years, something which Dr. Andrew Slater from the NSIDC is also working on).

And in his blog on 27 May he wrote:

Given the high pressure over the Beaufort Sea, the Beaufort Gyre has been doing its thing for a couple of weeks now, which means ice on the Atlantic side of the Arctic gets pushed into the Greenland Sea, via Fram Strait.

So I don't think that I'm imagining that there was quite high Fram export near the start of the melt season.
You are not, but the high export has met equally obscene temps, so exported ice has disappeared.

The main evidence of significant recent export is the mobility of the CAB; the key isn't that the entirety has been pushed to FRAM/the ATL, it's that there has been enough to allow the pack to lose structural integrity and break apart in the wake of the ATL/FRAM export.

According to satellite imagery, it was strong until end of May. No more since then. There was export toward Svalbard during June though, but that petered out a couple of weeks ago as well. All this is evident as long as one believes satellite pics are not fabricated

epiphyte

  • Frazil ice
  • Posts: 387
    • View Profile
  • Liked: 22
  • Likes Given: 21
Re: Home brew AMSR2 extent & area calculation
« Reply #2212 on: July 01, 2016, 07:02:01 AM »
Has anyone else noticed that of late the (IMO fatally simplistic) "extent" horserace always seems to converge to a dead-heat on or about July 16th?
No surprise - it's around then I expect that most of the ice in peripheral seas and lower latitudes has melted out.

...Quite! Which makes all the variation in extent prior to that pretty much moot, don't you think?

epiphyte

  • Frazil ice
  • Posts: 387
    • View Profile
  • Liked: 22
  • Likes Given: 21
Re: Home brew AMSR2 extent & area calculation
« Reply #2213 on: July 01, 2016, 07:29:39 AM »
Has anyone else noticed that of late the (IMO fatally simplistic) "extent" horserace always seems to converge to a dead-heat on or about July 16th?
This is likely only an artifact of the small size of the dataset for that date, with only years 2013-2015.

in this sense,
Quote
Artefact: noun: "something observed in a scientific investigation or experiment that is not naturally present but occurs as a result of the preparative or investigative procedure."

Yes, likely an artefact. But is it likely that a useful predictive metric converges to the same point at the same time, in not two but three widely differing melt years, before going on to again diverge into three very different trajectories?

Isn't it more likely the case that, around the middle of July especially, instantaneous extent purports no useful information w.r.t the prior or subsequent melt progression?

Rob Dekker

  • Nilas ice
  • Posts: 2386
    • View Profile
  • Liked: 120
  • Likes Given: 119
Re: Home brew AMSR2 extent & area calculation
« Reply #2214 on: July 01, 2016, 09:23:07 AM »
I got no reaction on this graph, should I make a monthly update ?

Etienne, it seems to me that your graph shows a trend that is apparent in every long term NSIDC extent graph for any month, and is half a year behind any "current" developments.
This is our planet. This is our time.
Let's not waste either.

Rob Dekker

  • Nilas ice
  • Posts: 2386
    • View Profile
  • Liked: 120
  • Likes Given: 119
Re: Home brew AMSR2 extent & area calculation
« Reply #2215 on: July 01, 2016, 09:59:44 AM »
Has anyone else noticed that of late the (IMO fatally simplistic) "extent" horserace always seems to converge to a dead-heat on or about July 16th?

That is a myth.
For example, look at the IJIS "extent" development around July 16 over the past decade :


No such thing as a "dead-heat" around July 16.
This is our planet. This is our time.
Let's not waste either.

Richard Rathbone

  • Nilas ice
  • Posts: 1730
    • View Profile
  • Liked: 387
  • Likes Given: 24
Re: Home brew AMSR2 extent & area calculation
« Reply #2216 on: July 01, 2016, 11:44:50 AM »
You are looking at the wrong data. That's a graph of ice extent, not melt pond extent.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2217 on: July 01, 2016, 01:12:24 PM »
A particular slow decline in area, rankings unchanged though.

Update 20160630.

Extent: -88.1 (-579k vs 2015, -26k vs 2014, -607k vs 2013, +33k vs 2012)
Area: -6.1 (-510k vs 2015, -111k vs 2014, -449k vs 2013, -107k vs 2012)
 
You will find the updated graphs in the top post

Starting with area, the Greenland Sea and ESS had rather big increases: +38k and +26k. Kara and CAB declined -22k and -17k in area.

In regional extent, the Greenland Sea increased as well (+18k). ESS declined thoug: -22k, Kara -14k.

Regional delta map is ESS that dropped in extent and increased in area. Large length of the coast is getting ice free.


Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2218 on: July 01, 2016, 01:24:31 PM »
The animation is of the Barents section. Any ice that is too far from the edge can be seen melting. Fast ice on the north coast of Greenland is breaking off.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2219 on: July 01, 2016, 01:35:38 PM »
And the surface melting situation, a jump in melt extent. Mostly in the Chukchi/ESS regions and along the ice edge north of Svalbard.

Shared Humanity

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2220 on: July 01, 2016, 03:40:10 PM »
Based on this latest animation, while it is clear that any ice that strays from the main pack towards the Fram melts quickly, it is just as clear that there is little if any transport through the Fram or east of Svalbard into the Barents. As seaicesailor has pointed out, this has been the case for most of June and it has contributed to the slow down in melt we have seen this past month. What is remarkable in this image is the integrity of the ice edge north of Svalbard, given the fragmented and fragile state of the ice pack. This is only because the pack is relatively stationary. I still believe the Atlantic side holds the key to the final numbers for this years melt season. If a strong high were to set up over the CAA/Greenland in the second half of the melt season and get transport moving, it will be a race to the bottom as the ridiculously warm waters devour the increasingly fragile ice. As most of us realize, the ice is more fragmented and mobile than ever before. We all better hope it doesn't get moving rapidly through the Fram and east of Svalbard into the Barents.
« Last Edit: July 01, 2016, 04:07:19 PM by Shared Humanity »

Rob Dekker

  • Nilas ice
  • Posts: 2386
    • View Profile
  • Liked: 120
  • Likes Given: 119
Re: Home brew AMSR2 extent & area calculation
« Reply #2221 on: July 02, 2016, 09:53:57 AM »
You are looking at the wrong data. That's a graph of ice extent, not melt pond extent.

No, Richard. I was looking at the right data.
Remember where this came from ?

Quote
: epiphyte  June 30, 2016, 09:38:56 AM
Has anyone else noticed that of late the (IMO fatally simplistic) "extent" horserace always seems to converge to a dead-heat on or about July 16th?

No suggestion of "melt pond extent" there.
This is our planet. This is our time.
Let's not waste either.

Richard Rathbone

  • Nilas ice
  • Posts: 1730
    • View Profile
  • Liked: 387
  • Likes Given: 24
Re: Home brew AMSR2 extent & area calculation
« Reply #2222 on: July 02, 2016, 12:55:01 PM »
You are looking at the wrong data. That's a graph of ice extent, not melt pond extent.

No, Richard. I was looking at the right data.
Remember where this came from ?

Quote
: epiphyte  June 30, 2016, 09:38:56 AM
Has anyone else noticed that of late the (IMO fatally simplistic) "extent" horserace always seems to converge to a dead-heat on or about July 16th?

No suggestion of "melt pond extent" there.

Just look at Wipneus's graph. Its obvious.

Peter Ellis

  • Grease ice
  • Posts: 619
    • View Profile
  • Liked: 33
  • Likes Given: 14
Re: Home brew AMSR2 extent & area calculation
« Reply #2223 on: July 02, 2016, 01:31:17 PM »
Just look at Wipneus's graph. Its obvious.
Obvious that there is a convergence in extent when you look at only four years' data, using one particular algorithm for measuring extent.

If you use other algorithms, there is no such convergence (check out NIPR or Charctic).  If you look at a wider time frame, there is no such convergence.

Ergo, this is just a coincidence affecting those four years in that one specific data set, and is not scientifically interesting.

seaicesailor

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2224 on: July 02, 2016, 01:54:36 PM »
Just look at Wipneus's graph. Its obvious.
Obvious that there is a convergence in extent when you look at only four years' data, using one particular algorithm for measuring extent.

If you use other algorithms, there is no such convergence (check out NIPR or Charctic).  If you look at a wider time frame, there is no such convergence.

Ergo, this is just a coincidence affecting those four years in that one specific data set, and is not scientifically interesting.

True.
In fact 2016 did depart during April - May from the statistically narrower distribution of extents (well, at least 30-year statistics). That was interesting

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2225 on: July 02, 2016, 06:02:28 PM »
2016 extent is now behind 2012 and 2014 (by very small amounts). 2016 area is still the lowest.

Update 20160701.

Extent: -55.5 (-568k vs 2015, +18k vs 2014, -519k vs 2013, +24k vs 2012)
Area: -109.8 (-594k vs 2015, -189k vs 2014, -514k vs 2013, -134k vs 2012)

Regional extent: only Kara is worth mentioning -17k.

Regional area: Laptev (-39) and ESS (-26k) declined most. CAB increased 17k.

regional delta map is from Kara where most of the remaining ice is now staying in the east corner.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2226 on: July 02, 2016, 06:07:06 PM »
Animation is Hudson, how long?

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2227 on: July 02, 2016, 06:18:51 PM »
And the surface melting situation, which did not change much.

Lord M Vader

  • Nilas ice
  • Posts: 1406
    • View Profile
  • Liked: 60
  • Likes Given: 39
Re: Home brew AMSR2 extent & area calculation
« Reply #2228 on: July 02, 2016, 06:21:10 PM »
Wipneus: I give all that green/yellow/red blob at most 10 days.

A-Team

  • Young ice
  • Posts: 2977
    • View Profile
  • Liked: 944
  • Likes Given: 35
Re: Home brew AMSR2 extent & area calculation
« Reply #2229 on: July 03, 2016, 01:36:59 AM »
Quote
give all that green/yellow/red blob at most 10 days.
The 85º-87º band around the north pole is one of the most intensively surveyed regions on the planet because of the constraints on polar orbiting satellites (they just about all pass overhead every day, some multiple times). That did not mean a great deal for us until the launch of Sentinel 1A because the region has been packed with clouds for months, meaning only active radar has adequate imaging resolution.

The first image below shows an entire swath of ice between the pole and CAA to be as solid as a rock on 02 Jul 16. And that seems to be case for the entire circumpolar band. It requires a click to display at full width (and even that is only 1/4 of the available resolution).

It is very convenient to get this imagery from http://www.polarview.aq/arctic; we could be doing a lot more in terms of tracking ice condition, bulk movement beneath the clouds  and formation of independent floes, eg north of Svalbard.
« Last Edit: July 03, 2016, 01:50:44 AM by A-Team »

slow wing

  • Grease ice
  • Posts: 823
    • View Profile
  • Liked: 155
  • Likes Given: 546
Re: Home brew AMSR2 extent & area calculation
« Reply #2230 on: July 03, 2016, 01:45:06 AM »
Thanks A-team, that is spectacular resolution and very informative, seeing through the cloud cover.

bbr2314

  • Nilas ice
  • Posts: 1817
    • View Profile
  • Liked: 158
  • Likes Given: 53
Re: Home brew AMSR2 extent & area calculation
« Reply #2231 on: July 03, 2016, 01:58:30 AM »
Pretty sure the donut hole is outside those bounds but nevertheless useful/interesting.

Per DMI/HYCOM we are about to see big holes opening up across 80N in terms of concentration... but combined with the storms over the next few days, thickness should also begin taking big hits. DMI shows large areas starting to drop below 1M thickness and is also now picking up on the concentration hits.




Tigertown

  • Nilas ice
  • Posts: 1678
    • View Profile
  • Liked: 24
  • Likes Given: 20
Re: Home brew AMSR2 extent & area calculation
« Reply #2232 on: July 03, 2016, 04:04:01 AM »
I think they are starting to believe you, as they are no longer tearing you a new one everytime.
"....and the appointed time came for God to bring to ruin those ruining the earth." Revelation 11:18.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2233 on: July 03, 2016, 08:14:07 AM »
Solid centuries for both area and extent. 2016 is leading again (with small margin) in my data set limited to recent years.

Update 20160702.

Extent: -136.4 (-627k vs 2015, -38k vs 2014, -477k vs 2013, -67k vs 2012)
Area: -134.1 (-622k vs 2015, -285k vs 2014, -497k vs 2013, -224k vs 2012)
 
You will find the updated graphs in the top post

Highest extent loss in the Hudson: -36k.

Highest area loss in ESS: -49k, Hudson was just -21k.

Regional delta is Laptev. Laptev is now definitely the slowest in my limited data set. Will it become ice free as it did in previous years?

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2234 on: July 03, 2016, 08:22:45 AM »
The animation is of the Canadian Archipelago. A large piece of ice in the Barrow Strait gets mobile.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2235 on: July 03, 2016, 08:28:01 AM »
And the surface melting situation. Total extent did not chnage muc, there are some regional shifts.

bbr2314

  • Nilas ice
  • Posts: 1817
    • View Profile
  • Liked: 158
  • Likes Given: 53
Re: Home brew AMSR2 extent & area calculation
« Reply #2236 on: July 03, 2016, 08:43:40 AM »
Comparing sat images of Beaufort over the past week, it seems like things are now quickly beginning to go poof... bet big block splits in the next day or three.

Darvince

  • Frazil ice
  • Posts: 318
    • View Profile
  • Liked: 31
  • Likes Given: 7
Re: Home brew AMSR2 extent & area calculation
« Reply #2237 on: July 03, 2016, 10:20:39 AM »
Looks like Hudson to go next week and Baffin the week after, based on the melt concentration.

magnamentis

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2238 on: July 03, 2016, 01:38:51 PM »
Comparing sat images of Beaufort over the past week, it seems like things are now quickly beginning to go poof... bet big block splits in the next day or three.

hey, looks like our expectations come true, i wouldn't be totally surprised if we saw a 200k+ daily loss during the next 2-3 weeks. someone else came up first with the "poof" term but i like it because this year this could well be what makes the difference, there are a few large areas where extent could go "poof" within a day or two, especially if the wind patterns that are actually predicted come true. keep goin'

A-Team

  • Young ice
  • Posts: 2977
    • View Profile
  • Liked: 944
  • Likes Given: 35
Re: Home brew AMSR2 extent & area calculation
« Reply #2239 on: July 03, 2016, 05:06:36 PM »
Looking now at Arctic Ocean coverage by the newly available ASTER imagery (green, red, near infrared; 15 m resolution 8-bit precision) at EarthExplorer, it is very similar to Landsat-8 in that satellite admin decided to only keep near-coastal imagery in the belief no one is interested in images of sea ice. This could amount to shutting down acquisition, download, storage, or provisioning.

The first image shows recent swaths through the Beaufort Sea. There are some nice images of the Mackenzie delta, land-fast ice, and floes on the shelf. Coverage does not extend to very much of the Beaufort Gyre; for example, Big Block floe is well out of range (red star 73.3º, -147.3º).

The second image shows one of the outer floes at full resolution -- which reveals many tiny pieces of sea ice that would not be visible with Modis and are very unlikely to be used in calibrating extent, area, or volume products.

The small round 'blue' circle may be a melt pond. There is no way however to make 'natural color' since the satellite has no blue channel. It might be approximated by duplicated the green channel (which will be highly correlated) and discarding the infrared. The 4th image shows various channel combinations with the potential to resolve this.

The 3rd image below decomposes the colors of the supposed pond into their RGB channels. The 'red' channel seems to provide the most detail; it is probably band 1 (red wavelengths, second image). The 'green' channel then corresponds to band 2 green wavelengths and the blown-out contrast to band 3 infrared.

The interesting thing about ASTER is the 15m infrared channel. The nearest counterpart on Landsat-8 is the 30 m band 5 which is farther out in the infrared: 0.85–0.88 vs 0.76-0.86 µm. We have done very little on these forums with band 5. There is mass confusion over how to interpret anything not in the visible.

On ASTER, band 3 has both nadir and backward looking options, designated 3N or 3B. The purpose of 3B is unclear; it involves a time lag.

There's quite a good interactive applet tutorial on color mixing at the links below, helpful in understanding mixing of monitor color channels and human color perception. The ASTER infrared channel is somewhat outside the gamut of what the eye can see. The gaussian-type peaks represent the likelihood that a photon of the given wavelength will register with the sensor. That's most likely at the peak wavelength but falls off on the shoulders.

https://graphics.stanford.edu/courses/cs178-10/applets/colormatching.html

https://graphics.stanford.edu/courses/cs178/applets/colormixing.html

ASTER images typically come in a HDF-EOS file format but can be converted to Geotiff files in several projections.Tools for carrying out data transformations can be found at https://lpdaac.usgs.gov/tools. though this seems not to go anywhere useful to ordinary users.

Level 3 ASTER images classify each pixel of polar scenes into one of eight classes: water cloud, ice cloud, aerosol/dust, water, land, snow/ice, slush ice, and shadow. Sounds promising but where do we get them? EarthExplorer has only the DEM and L1T images.

http://wiki.landscapetoolbox.org/doku.php/remote_sensor_types:aster

https://en.wikipedia.org/wiki/Advanced_Spaceborne_Thermal_Emission_and_Reflection_Radiometer
« Last Edit: July 03, 2016, 09:34:00 PM by A-Team »

seaicesailor

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2240 on: July 03, 2016, 05:32:09 PM »
I think they are starting to believe you, as they are no longer tearing you a new one everytime.
Lol (because that was good, no offense intended)
In my case people stopped insisting in the massive amount of bottom melt and I stopped insisting it was not.
However I acknowledged that predictions corresponded to polynya being created by storms before this thing even started.
http://forum.arctic-sea-ice.net/index.php/topic,1493.msg79498.html#msg79498
The coming storm is going to create more damage and now bottom melt is kicking in in the CAB because temperatures are sustained over zero (not so much because of solar radiation over open sea).
But as said by others with more experience, perhaps compaction reduces these holes before becoming anything really extraordinary.
« Last Edit: July 03, 2016, 05:38:15 PM by seaicesailor »

A-Team

  • Young ice
  • Posts: 2977
    • View Profile
  • Liked: 944
  • Likes Given: 35
Re: Home brew AMSR2 extent & area calculation
« Reply #2241 on: July 03, 2016, 06:26:09 PM »
Quote
The animation is of the Canadian Archipelago. A large piece of ice in the Barrow Strait gets mobile.
Yep, opening up. that large piece is ~113 km in length.

A-Team

  • Young ice
  • Posts: 2977
    • View Profile
  • Liked: 944
  • Likes Given: 35
Re: Home brew AMSR2 extent & area calculation
« Reply #2242 on: July 03, 2016, 06:32:42 PM »
Quote
Pretty sure the donut hole is outside those bounds but nevertheless useful/interesting.
No. The latitude and longitude lines are precisely computed at the source and irrevocably embedded in the imagery. The S1A imagery indeed extends beyond 87ºN leaving a rather small hole about the pole that lies outside its swath.

I looked at tiling up a complete ring of those rectangles around the north pole but this a couple of issues: first, it would represent a span of ~ six weeks of dates (30 days are shown below) rather than a snapshot on a fixed date; second the file size would get very unwieldy at a decent resolution.

That leaves an animation tour around the ring which would need locator insets on each frame. Under forum constraints it may not be feasible to improve much on the 3.125 km AMSR2 scenes that wipneus is providing. While S1A provides a more familiar 'picture', both arise from microwave imaging.

« Last Edit: July 03, 2016, 06:43:47 PM by A-Team »

epiphyte

  • Frazil ice
  • Posts: 387
    • View Profile
  • Liked: 22
  • Likes Given: 21
Re: Home brew AMSR2 extent & area calculation
« Reply #2243 on: July 03, 2016, 07:33:16 PM »
Just look at Wipneus's graph. Its obvious.
Obvious that there is a convergence in extent when you look at only four years' data, using one particular algorithm for measuring extent.

If you use other algorithms, there is no such convergence (check out NIPR or Charctic).  If you look at a wider time frame, there is no such convergence.

Ergo, this is just a coincidence affecting those four years in that one specific data set, and is not scientifically interesting.


Well I did say "of late". The graph shows the line converging toward/away from the same point for five out of the past five years. To say that's not scientifically interesting is to suggest that there is nothing to be learned from it.

Which is more likely? a) crandles' observation is correct - since 2012/3, every year, the easy ice has all gone before that point in time, and none of the difficult ice has yet succumbed, or b) it's a random coincidence which is just noise in the context of a longer timeframe.

- if a) is true it would suggest that there actually is some consistency to the vagaries of extent in the post-myi era, at least at that point. If so, then the closer we get to that point, the more significant a departure from a trajectory which is converging toward it would be. Conversely, the *less* significant the instantaneous slope of the line as long as it is aimed toward that point.

- if, otoh, b) is true, and it is just a random coincidence, then the wide variation in final outcomes relative to that point would reinforce the notion (which I have always entertained in the past) that at this time of year it is the absolute value of extent which is not scientifically interesting.



Rob Dekker

  • Nilas ice
  • Posts: 2386
    • View Profile
  • Liked: 120
  • Likes Given: 119
Re: Home brew AMSR2 extent & area calculation
« Reply #2244 on: July 04, 2016, 07:41:15 AM »
The animation is of the Canadian Archipelago. A large piece of ice in the Barrow Strait gets mobile.

Thank you Wipneus ! In the same animation, on the other side of the Canadian Arctic, the remainder of Amundsen Gulf is seen going mobile at the same time, all the way through Dolphin and Union Strait deep into Coronation Gulf.

The ice-covered part of the NorthWest Passage just became a whole lot shorter in the span of a few days. Quite amazing..
This is our planet. This is our time.
Let's not waste either.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2245 on: July 04, 2016, 01:21:06 PM »
2016 is still leading with small margins.

Update 20160703.

Extent: -85.0 (-633k vs 2015, -95k vs 2014, -430k vs 2013, -57k vs 2012)
Area: -79.9 (-609k vs 2015, -347k vs 2014, -367k vs 2013, -173k vs 2012)
 
You will find the updated graphs in the top post

CAA regional extent dropped most (-24k), followed by Hudson (Hudson (-23k) and Kara (-16k).

Regional area declined big way in the CAB (-49k), CAA (-35k) and Hudson (-26k). It increased in the Beaufort (+21k).

The daily regional delta map is CAA with big drops in both area and extent.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2246 on: July 04, 2016, 01:26:26 PM »
And an animation of the ice north of greenland. The fast ice can be seen detaching, but is in no way going fast to the exit. Ice in the Fram Strait is not going anywhere.
There is a curious dark blotch near the North Pole in the last frame. My guess it is rain associated with a cyclone in that neighborhood.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2247 on: July 04, 2016, 01:32:05 PM »
And the surface melting situation. No net change, most of the thick ice north of the CAA and Greenland is now in melt.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2248 on: July 05, 2016, 04:38:05 PM »
declines are enough to stay ahead of 2012 by a small margin.

Update 20160704.

Extent: -112.5 (-666k vs 2015, -188k vs 2014, -338k vs 2013, -65k vs 2012)
Area: -91.9 (-558k vs 2015, -408k vs 2014, -334k vs 2013, -95k vs 2012)
 
You will find the updated graphs in the top post

Most of the extent drop is by Hudson (-38k) and Baffin (-28k). ESS drops a further -18k.

Area losses are largest in Greenland Sea (-31k), Baffin (-26k) and Laptev (-19k).

Regional delta map is Baffin. The ice near Labrador is now almost all gone and in the north is is just a matter of time.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2249 on: July 05, 2016, 04:42:17 PM »
Animation is of the East Siberian Sea. Any remaining doubt about the mobility of the fast ice has been removed now.