Please support this Forum and Neven's Blog

Author Topic: Home brew AMSR2 extent & area calculation  (Read 848182 times)

Wipneus

  • ASIF Governor
  • Posts: 3170
    • View Profile
    • Arctische Pinguin
Re: Home brew AMSR2 extent & area calculation
« Reply #2900 on: March 19, 2017, 10:53:09 AM »
Plenty of related along the ESS-Laptev fast ice edge.

Jim Hunt

  • ASIF Governor
  • Posts: 3015
    • View Profile
    • The Arctic sea ice Great White Con
Re: Home brew AMSR2 extent & area calculation
« Reply #2901 on: March 19, 2017, 03:17:51 PM »
Perhaps it is fixed now

My first link leads to the NetCDF files, but if you click through my second link to the AMSR2 images I can still only access the SSMIS ones. The "Ice-Conc-Amsr" product menu option doesn't currently exist as far as I can tell.
Reality is merely an illusion, albeit a very persistent one - Albert Einstein

epiphyte

  • ASIF Citizen
  • Posts: 298
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2902 on: March 19, 2017, 09:25:56 PM »
If you have panoply installed you can plot from the .nc file. E.g. here's one from from ice_conc_nh_polstere-100_amsr2_201703181200.nc (yesterday) with the scale set to show maximum contrast in the range 88-100% (i.e. everything under 88% is dark blue)...


Jim Hunt

  • ASIF Governor
  • Posts: 3015
    • View Profile
    • The Arctic sea ice Great White Con
Re: Home brew AMSR2 extent & area calculation
« Reply #2903 on: March 19, 2017, 11:15:17 PM »
If you have panoply installed you can plot from the .nc file.

I do have Panoply installed, so sure. I've even been beta testing some improvements. Nonetheless it does still seem as though the OSI-SAF web site is somewhat broken in the AMSR2 department.
Reality is merely an illusion, albeit a very persistent one - Albert Einstein

Wipneus

  • ASIF Governor
  • Posts: 3170
    • View Profile
    • Arctische Pinguin
Re: Home brew AMSR2 extent & area calculation
« Reply #2904 on: March 23, 2017, 09:35:46 AM »
In case anybody wonders: second day without the usual data update from Uni Hamburg.

Jim Hunt

  • ASIF Governor
  • Posts: 3015
    • View Profile
    • The Arctic sea ice Great White Con
Re: Home brew AMSR2 extent & area calculation
« Reply #2905 on: March 23, 2017, 09:44:41 AM »
I'd noticed that too Wipneus!

Meanwhile OSI-SAF tell me:

Concerning the ice type product, there is as you point out a systematic problem that gives multi-year ice in some young ice areas, and we are working on improving that.

and

Concerning the new AMSR2 product, we will provide a quicklook archive of product images as for SSMIS. This should be available in a couple of weeks.
Reality is merely an illusion, albeit a very persistent one - Albert Einstein

Wipneus

  • ASIF Governor
  • Posts: 3170
    • View Profile
    • Arctische Pinguin
Re: Home brew AMSR2 extent & area calculation
« Reply #2906 on: March 29, 2017, 12:48:48 PM »
Data stream is back with an address change to:

ftp://ftp-projects.cen.uni-hamburg.de/seaice/AMSR2/


Wipneus

  • ASIF Governor
  • Posts: 3170
    • View Profile
    • Arctische Pinguin
Re: Home brew AMSR2 extent & area calculation
« Reply #2907 on: March 29, 2017, 12:50:31 PM »
The crack in the ESS is getting real wide. The Laptev fast ice is torched.

DrTskoul

  • ASIF Middle Class
  • Posts: 731
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2908 on: March 29, 2017, 01:45:45 PM »
The crack in the ESS is getting real wide. The Laptev fast ice is torched.

I guess it starts with a bang... that is quite the drift
“You can know the name of a bird in all the languages of the world, but when you're finished, you'll know absolutely nothing whatever about the bird... So let's look at the bird and see what it's doing -- that's what counts.”
― Richard P. Feynman

Darvince

  • ASIF Citizen
  • Posts: 146
    • View Profile
    • NSIDC Daily
Re: Home brew AMSR2 extent & area calculation
« Reply #2909 on: March 29, 2017, 02:19:19 PM »
Does this image have enhanced contrast or is it really that bad?

slow wing

  • ASIF Middle Class
  • Posts: 512
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2910 on: March 29, 2017, 02:32:59 PM »
The crack in the ESS is getting real wide. The Laptev fast ice is torched.

Wow! So what is happening here?

 EDIT: U. Bremen's AMSR2 concentration map for 2017-03-28 is also showing the loss - with around 75% concentration in the fast-ice region of the Laptev Sea that has gone dark in Wipneus' graphic.

Is it melt ponds or is there actually no ice over around 25% of that region?

Options:
A. We're only just past the equinox, there presumably won't be enough sunlight to cause melt ponds, fair?
B. But it's plausible that winds could have melted some snow cover?
C. Alternatively, the ice could have been in such a poor state that the winds tipped it over the boundary to where much of it is no longer classified as ice cover?
D. Or it might be tearing in the wind and now full of micro-cracks?

Anyone more knowledgeable have a take on which of the options is most likely?

The screenshot is from Nullschool at 12:00 UTC on 2017-03-27.
The illustrative chosen point in the Laptev Sea with the green circle is showing 25 kph winds at -2.6 degrees C.
« Last Edit: March 29, 2017, 02:52:06 PM by slow wing »

Tor Bejnar

  • ASIF Upper Class
  • Posts: 1112
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2911 on: March 29, 2017, 02:53:24 PM »
The crack in the ESS is getting real wide. ...
Last year about this time (and basically for two months) winds were pushing Beaufort Sea ice toward the East Siberia Sea (ESS), opening the Beaufort early.  It is starting to look like the reverse may be happening this year.

This is probably premature, but would an ESS opening early or Beaufort Sea opening early affect the CAB differently?  (I'm thinking in terms of a "Laptev Bite"-like phenomenon).
Arctic ice is healthy for children and other living things.

Wipneus

  • ASIF Governor
  • Posts: 3170
    • View Profile
    • Arctische Pinguin
Re: Home brew AMSR2 extent & area calculation
« Reply #2912 on: March 29, 2017, 03:08:52 PM »
Yes, these images are enhanced in contrast so it is not as bad as it looks.

For the new readers here: we see this torching in the AMSR2 ASI images every year but a completely satisfactory explanation is not given (to me at least).

Melt ponding: unlikely at the current temperatures; NSIDC is normally the most sensitive to surface whetting, yet nothing is to be seen there.
Clouds/water vapor: unlikely as it would increase (whiten) apparent sea ice concentration.

Most likely (IMHO) explanation is some form of snow fall and/or a malfunctioning of the ASI algorithm here.

Not withstanding that we do not know exactly what it causes, we have learned to associate these "torchings" with warming events (but not necessary enough for melting )

seaicesailor

  • ASIF Upper Class
  • Posts: 1244
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2913 on: March 29, 2017, 03:12:51 PM »
There might have been some rain or a brief period with conditions near zero, which at least affects the snow blanket. Not as bad as contrasting implies, but it may qualify as pre-pre-conditioning (half kidding)
Edit. Well there you go, thank you Wipneus

slow wing

  • ASIF Middle Class
  • Posts: 512
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2914 on: March 29, 2017, 04:18:12 PM »
Wipneus and SeaIceSailor,

  Thanks for your comments on the potential causes of the concentration loss observed in the Laptev Sea.

  Yes, a snow/rain connection looks reasonable: toggling to the "3-hour precipitation" option on Nullschool and stepping through the 3-hourly displays does indeed show precipitation of order 1 mm (which could be ~1 cm if it fell as snow) drifting over the region around that time.

Neven

  • Administrator
  • ASIF Governor
  • *****
  • Posts: 3409
    • View Profile
    • Arctic Sea Ice Blog
Re: Home brew AMSR2 extent & area calculation
« Reply #2915 on: March 29, 2017, 04:49:16 PM »
There might have been some rain or a brief period with conditions near zero, which at least affects the snow blanket. Not as bad as contrasting implies, but it may qualify as pre-pre-conditioning (half kidding)

That's exactly right. It's called melt onset, and once the pre-pre-condition snow/ice is hit with shortwave radiation (sunshine) things can go really fast.

So, the worse thing that can happen on the Siberian side of the Arctic, is for all the anomalous heat, clouds (downwelling longwave radiation) and early opening up, is for a high to move in next.
Il faut cultiver notre jardin

JimboOmega

  • ASIF Citizen
  • Posts: 115
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2916 on: April 04, 2017, 06:14:34 PM »
BTW... under what circumstances does Wipneus provide the numbers? Or are they no longer relevant/being produced for some reason?

I used to look here first, to see the ice volume of the day...

Jim Williams

  • ASIF Citizen
  • Posts: 348
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2917 on: April 04, 2017, 06:26:11 PM »
There might have been some rain or a brief period with conditions near zero, which at least affects the snow blanket. Not as bad as contrasting implies, but it may qualify as pre-pre-conditioning (half kidding)

That's exactly right. It's called melt onset, and once the pre-pre-condition snow/ice is hit with shortwave radiation (sunshine) things can go really fast.

So, the worse thing that can happen on the Siberian side of the Arctic, is for all the anomalous heat, clouds (downwelling longwave radiation) and early opening up, is for a high to move in next.
I was looking at nullschool with the high in the west and low in the east that are attempting to setup....and thinking that would be worse.  Not if they are transient, but if they get stuck.....


RikW

  • ASIF Lurker
  • Posts: 11
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2918 on: April 04, 2017, 09:22:35 PM »
BTW... under what circumstances does Wipneus provide the numbers? Or are they no longer relevant/being produced for some reason?

I used to look here first, to see the ice volume of the day...

At least for PIOMAS aftr they are uploaded which happens once a month in the first couple of days of the month on the webpage of the science center that produces those numbers. And they uploaded them a couple of hours ago I think, so I guess Wipneus will update his graphs in a couple of days/ hours

Lord M Vader

  • ASIF Middle Class
  • Posts: 952
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2919 on: April 04, 2017, 09:34:11 PM »
PIOMAS for March:


JimboOmega

  • ASIF Citizen
  • Posts: 115
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2920 on: April 11, 2017, 05:52:01 PM »
BTW... under what circumstances does Wipneus provide the numbers? Or are they no longer relevant/being produced for some reason?

I used to look here first, to see the ice volume of the day...

At least for PIOMAS aftr they are uploaded which happens once a month in the first couple of days of the month on the webpage of the science center that produces those numbers. And they uploaded them a couple of hours ago I think, so I guess Wipneus will update his graphs in a couple of days/ hours

I misspoke when I said volume - I meant ASMR2 Extent & Area (what this thread is for).

I used to look forward to the numbers - including the breakdown by basin. A little tick on a NSIDC chart is not nearly as nice :).

oren

  • ASIF Upper Class
  • Posts: 1057
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2921 on: April 11, 2017, 06:46:38 PM »
You can find the full data on Wipneus' site, arctischepenguin or something like that. I have it bookmarked.

schnitm

  • ASIF Lurker
  • Posts: 5
  • Data Geek
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2922 on: April 13, 2017, 12:08:48 AM »
Was poking around Wipneus' site.  He shows an 8 sigma event in Sea Ice Area Anomaly last year.  8 sigma is astronomical, probability in the rage of 10^-15.  Really just doesn't happen in stable systems.  Of course the system is not stable.  there is clearly a slope and considerably more 2 and 3 sigma events than there should be since 2006 and there is clear autocorrelation.  So it might be a 6 or 7 sigma event in a system in decline.  Even than, this graph is the most terrifying thing this statistician has seen here, a high negative sigma event in an autocorrelated declining system.  All things are possible but that looks to me like a crash is around the corner and this is global!

https://14adebb0-a-62cb3a1a-s-sites.googlegroups.com/site/arctischepinguin/home/sea-ice-extent-area/grf/nsidc_global_area_normanomaly.png?attachauth=ANoY7cp3xOtJ_nLfDpNZdtu7-xMj-WeKmBBkX1cZ2oA23rK_tsFAoTTMYLP4oEvlhyjruLrOVo1c9qKKhng9bTtH6rB1N25DEIb2a_VJmmAzboSL72bB6kZpsrQe895zExE6wVgqxjNIxgK2lwu0bJhqTVZWxSWj0KOFja65JVx3QMyob4sr1GeDaFMbxIC8xUEH0HeZgJ1XG21T8dcm23FA1oNiui68U9GLX0VkHUgudVwJD_hjcc7FUAn3j4FE9520ep65ZieczoGMOQGcZuOICLL43s9RfhGlbQVSbi9GFAOgq6EteRA%3D&attredirects=0

Dundee

  • ASIF Lurker
  • Posts: 39
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2923 on: April 13, 2017, 05:20:10 AM »
As a bored, retired, and concerned quantitative analyst, I took a deep dive (well, deep for a home PC) thru NSIDC extent last year. Because 'standard deviations' against anomalies relative to a long term average in the presence of a known long term declining trend are problematic, I regressed day-wise and looked at residuals. Record minimums were not striking, because the natural variation at the end of the melting season is at its highest. What really stood out was the marked dip in as melting took hold. Dispersion has historically been very low that time of year - the residuals peaked higher than they'd been thru the entire satellite record, remained when short term variation was filtered out, and stayed high for weeks. Turned out that 'high sigma' (no real justification for treating the distribution as normal, it isn't really) negative excursions were much more frequent than expected - there was a secondary peak in the distribution.

I bounced the timing of the excursions against a number of factors and influences and came up completely blank for a common driver. It appeared negative excursions are becoming more frequent, and less likely to be coincide with isolated extreme weather events (as at least one of the extreme record minimums was).

I haven't cranked in the last six months of data, but am confident if I did the picture would not be pretty. I have no idea why things are unfolding as they are, but my gut feel (how's that for a statistical term of art) is that we are picking away at a wall between a well behaved noisy decline, and a new set of behavior driven by newly triggered rules. There has been a lot of discussion about just what constitutes an "ice free" Arctic - I've maintained when it comes, it will be obvious and there will be no need to wonder if five years or whatever are appropriate to confirm the event is not just system noise. I don't expect this year to be ice free, but would be surprised if next autumn we end up looking back on a boring, back-to-the-trend line melting season.

Back to topic, the great thing about NSIDC is its deliberately consistent approach across nearly forty years of data. I see AMSR2 becoming critical - in the current environment it is very possible it (and eventual successors) will be all we have to work with. I recall the gap in U.S. operational weather satellite coverage after the first Shuttle loss. We are potentially looking at something more profound, but adopted as policy rather than thru accident.

ktonine

  • ASIF Citizen
  • Posts: 354
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2924 on: April 13, 2017, 05:43:52 AM »
I haven't cranked in the last six months of data, but am confident if I did the picture would not be pretty. I have no idea why things are unfolding as they are, but my gut feel (how's that for a statistical term of art) is that we are picking away at a wall between a well behaved noisy decline, and a new set of behavior driven by newly triggered rules.

Yes, that seems most likely scenario.

I recall the gap in U.S. operational weather satellite coverage after the first Shuttle loss. We are potentially looking at something more profound, but adopted as policy rather than thru accident.


Problematic, but the EU will probably not let that happen - and one has to wonder if the Chinese have something they'd contribute that we don't know about.

Dundee

  • ASIF Lurker
  • Posts: 39
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2925 on: April 13, 2017, 06:33:08 AM »
No, this one is all on the U.S.

I am not comfortable drawing any statistical inference from a cyclic data set until I have something like 30 cycles to work with (100 is better, at 500 I can be specific). The NSIDC data is a rare bird - a quality data set with detailed global coverage covering nearly 40 years. Even so, I would not yet hazard a guess whether the declining trend is linear or just what.  If I do 30 year cuts at five year intervals, in two more years I will have only three NSIDC data points to work with . . . .

The current administration can draw a line under NSIDC data with a stroke of a pen. It would have to actively support the program to get the next set of sensors on orbit - something that so far as I know is not currently programmed - in order for us to say not only is ice declining, but state with any confidence how it is declining, on the basis of 50-60 years of data. It hurts that, just as NASA/Icebridge is beginning to do detailed work to validate and improve understanding of the fine structure of satellite measurements, the proposed budget would essentially end the whole line of research (that is to say polar ice in it's entirety, not just Icebridge specifically).

There is no good way to splice AMSR2 back to NSIDC data. If JAXA keeps sensors up, and can make them directly comparable across generations, we are still looking at 30 years before we can say things about JAXA data that we can today about NSIDC trends.

It is very probable that by the time we can say with confidence the rate of JAXA area/extent drop is linear, increasing, or decreasing, there will no longer be summer ice to measure and the question will be moot.

Human activities to influence the trend are, as I believe you are saying, another matter. The U.S. was long the largest economy in the world. It is now number three, and despite the bluster, there is no sign the trend of its influence will turn around any time soon. The current administration is on a path to last at best four years and in any case, the rest of the world is continually gaining leverage to influence U.S. behavior. There is hope (but way off topic here).

schnitm

  • ASIF Lurker
  • Posts: 5
  • Data Geek
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2926 on: April 13, 2017, 08:35:37 PM »
My gut agrees with Dundee's gut.  Doesn't mean it's going to happen, you know what they say about statisticians.  However, this data looks like what I've seen in data when looking back prior to a large state change.  Like the rules are about to shift dramatically.  Predicting future states that have no historical precedent is very hard though.  So who knows, maybe -  maybe not.

Wipneus

  • ASIF Governor
  • Posts: 3170
    • View Profile
    • Arctische Pinguin
Re: Home brew AMSR2 extent & area calculation
« Reply #2927 on: April 18, 2017, 08:30:31 AM »
Chukchi-ESS 2017 compared with 2016. Although some characteristics appear similar, the 2017 images show more breaking and "torching".

Click that image to start animating.

iceman

  • ASIF Citizen
  • Posts: 213
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2928 on: April 18, 2017, 01:18:34 PM »
Not to get too anthropomorphic and all, but it does look like 2017 is staring in gape-mouthed, sweaty-browed horror at the condition of the Bering Sea ice.

Wipneus

  • ASIF Governor
  • Posts: 3170
    • View Profile
    • Arctische Pinguin
Re: Home brew AMSR2 extent & area calculation
« Reply #2929 on: April 19, 2017, 09:32:47 AM »
Similar comments can be made for the situation in the Canadian Archipelago.

Click for an animation.

oren

  • ASIF Upper Class
  • Posts: 1057
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2930 on: April 19, 2017, 12:31:51 PM »
Nares has suddenly cleared.

Bill Fothergill

  • ASIF Citizen
  • Posts: 190
    • View Profile
Re: Home brew AMSR2 extent & area calculation
« Reply #2931 on: April 20, 2017, 08:59:54 PM »
Nares has suddenly cleared.

In the post-BREXIT UK, some might see that as another export opportunity.