Support the Arctic Sea Ice Forum and Blog

Author Topic: Home brew AMSR2 extent & area calculation  (Read 1515358 times)

BornFromTheVoid

  • Nilas ice
  • Posts: 1071
    • View Profile
  • Liked: 202
  • Likes Given: 78
Re: Home brew AMSR2 extent & area calculation
« Reply #2350 on: July 17, 2016, 02:16:14 PM »
Did you mean 2014? My reading of the graph is that 2013 was close to this year until late in July, when it fell off significantly for the rest of the melt season.

Yes, meant 2014. Will edit, thanks.
I recently joined the twitter thing, where I post more analysis, pics and animations: @Icy_Samuel

nowayout

  • New ice
  • Posts: 20
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2351 on: July 17, 2016, 02:20:27 PM »
Take a look at the topography:

https://en.wikipedia.org/wiki/File:Arctic_Ocean_bathymetric_features.png

and you know where the heat of the north atlantic  ocean current goes. And you know where the heat raking takes places for now.

Maybe next year we'll have the "real drop". It takes it time for the El Nino to propagate through the water.





icy voyeur

  • New ice
  • Posts: 63
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2352 on: July 17, 2016, 05:46:36 PM »
Love the new graph. This year races ahead and then does slip into relative mediocrity. It matches well with my perception based on everything I've been trying to digest in this season.

Mentally, I'm transforming the chart to above or below a moving average, don't really need a new chart for that, except to press the obvious if others do the same. Such a chart, however, would be a bit over derived.

Then the mind wanders into assigning above or below average melting percentage as a value to be potentially correlated with one of many observable to search for predictors, or simply strong correlates. Well, there's lazy Sunday thoughts from the peanut gallery.

RoxTheGeologist

  • Grease ice
  • Posts: 501
    • View Profile
  • Liked: 149
  • Likes Given: 112
Re: Home brew AMSR2 extent & area calculation
« Reply #2353 on: July 17, 2016, 08:14:28 PM »

Adding something like a 5 day and 15 day moving average will give us crossovers. They are standard momentum indicators for tracking stock market changes. Overall they are pretty useless for the stock market (momentum indicators that is). We have a strong underlying signal in the Arctic and that might show up clearly.

A-Team

  • Young ice
  • Posts: 2556
    • View Profile
  • Liked: 432
  • Likes Given: 29
Re: Home brew AMSR2 extent & area calculation
« Reply #2354 on: July 17, 2016, 08:17:37 PM »
Quote
Oren: wish way to resolve the location of the ice front vs. the movement of the ice. The front can be where it is due to ice movement relative to the pole (dispersion/compaction), or due to Atlantic encroachment.   ice moves south yet the front moves north. track individual floes near the front and average their movement, feasible on a short animation but automation for long time-frames and multiple years.
Right, several combinations of advances/melt/dispersion/compaction could result in the same end appearance of the front. In theory there is already automatic tracking in Hycom products; there might be a need for higher resolution and vetting against floe case studies.

Observations about bathymetry in #2349 and #2357 have to be very relevant. I wonder if the current position of the North Atlantic front is near its northernmost maximal extent, ie the drop-off at the edge of the continental shelf which limits the range of warm surface water.

It's a pity that WorldView doesn't have the IBCAO bathymetry as a layer. I recall it being in a different projection (PS with true scale only at 75%) which would be an obstacle to accurate overlays with AMSR2 or WV but it seems to overlay quite well locally, with 257.62% scale increase needed to align the bathymetry shown in #2349 relative to the large format AMSR2 3km.

Wow ... the shelf line-up with the current advance of North Atlantic waters is very impressive suggesting that the front could have already reached a limiting position except nearer to the large Russian island complex Severnaya Zemlya to the east where open water could still come farther polewards with getting beyond the shelf.

The color keys on the IBCAO map do not correspond to colors used to indicate depth because they used faux solar illumination to create relief; in Greenland west coast fjords, the map is known to have made wildly inaccurate guesses as to channels. However Google Earth has implemented a digital IBCAO and offers transect maps of depth (below). Its projection does not rescale so nicely at the edges and the bathymetric color palette is skimpy.

https://www.ngdc.noaa.gov/mgg/bathymetry/arctic/IBCAO_TechnicalReference.PDF

However according to the animation below, an automated strip of AMRSE2 3km to front essentials (last 4 frames), the front is still marching northward as of July 15th. The 16th continues this trend above and to the east of Franz Josef but the 17th (not shown) seems to affirm bathymetrically limited behavior. On the Aqua clear areas (not shown) at 250 m/pxl scale (at 70º lat), advances were averaging 11 km/day or one degree of latitude per ten days or 75 days to reach the pole (not gonna happen).

No doubt there are many academic papers on oceanographic circulation here that could have predicted this 2016 development and maybe some did. The situation is somewhat reminiscent of the landfast ice boundary, Mackenzie River discharge effective range and the continental shelf position in the Beaufort Sea.
« Last Edit: July 18, 2016, 06:58:39 PM by A-Team »

A-Team

  • Young ice
  • Posts: 2556
    • View Profile
  • Liked: 432
  • Likes Given: 29
Re: Home brew AMSR2 extent & area calculation
« Reply #2355 on: July 17, 2016, 08:39:55 PM »
Quote
Adding  a 5 day or 15 day moving average could bring out signal.
In math and physics these are called convolutions. Normally they would be centrally weighted rather than unweighted averages. They're easy to do online with graphing tools that have sliders for weighting std deviation, kurtosis and skew plus date range. Wip could attach the data pts as a text file  which is allowed by forum software without generating a tedious giant display.

peterlvmeng

  • Frazil ice
  • Posts: 115
    • View Profile
  • Liked: 31
  • Likes Given: 9
Re: Home brew AMSR2 extent & area calculation
« Reply #2356 on: July 18, 2016, 05:32:06 AM »
The concentration map shows a similar trend as 2012 in August. As the cyclone is still there in two days near the Beaufort Sea, the export of the ice to pacific side will be dramatic. However, the ice area or extent stall recently are at the cost of lower concentration area expansion. I probably think these low concentration areas will be completely melted out in 3 weeks.

S.Pansa

  • Frazil ice
  • Posts: 162
    • View Profile
  • Liked: 36
  • Likes Given: 5
Re: Home brew AMSR2 extent & area calculation
« Reply #2357 on: July 18, 2016, 08:02:37 AM »
I am not sure if it has been mentioned before in this thread (Sorry if it has already been discussed but the forum search didn't show anything). But in case it hasn't. I just saw following quote on the cryosphere page of the Universität Hamburg website (after all those "ACHTUNG +++ ...." remarks):
Quote
Von März auf April 2016 wurde die Prozessierung von DMSP-F17 auf DMSP-f18 umgestellt. Vorläufige Untersuchungen zeigen, dass infolgedessen die Meereisfläche in der Arktis um rund 1-2% niedriger, in der Antarktis aber womöglich um rund 1% höher ist.

According to them, preliminary investigations show, that due to the change in processing from DMSP-F17 to DMSP-f18 (March - April 2016) sea ice area in the Arctic is 1-2% lower as before (in their record I assume); while in the Antarctic it might be 1% higher.
Does this have any influence on the comparisons above (or is this already factored in)?

Thanks!

Rob Dekker

  • Nilas ice
  • Posts: 2386
    • View Profile
  • Liked: 123
  • Likes Given: 119
Re: Home brew AMSR2 extent & area calculation
« Reply #2358 on: July 18, 2016, 08:33:35 AM »
According to them, preliminary investigations show, that due to the change in processing from DMSP-F17 to DMSP-f18 (March - April 2016) sea ice area in the Arctic is 1-2% lower as before (in their record I assume); while in the Antarctic it might be 1% higher.

Mmm. NSIDC did not make any such statement when they switched over from F17 to F18.
I wonder what's going on here.
This is our planet. This is our time.
Let's not waste either.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4086
    • View Profile
    • Arctische Pinguin
  • Liked: 830
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2359 on: July 18, 2016, 08:35:11 AM »
I am not sure if it has been mentioned before in this thread (Sorry if it has already been discussed but the forum search didn't show anything). But in case it hasn't. I just saw following quote on the cryosphere page of the Universität Hamburg website (after all those "ACHTUNG +++ ...." remarks):
Quote
Von März auf April 2016 wurde die Prozessierung von DMSP-F17 auf DMSP-f18 umgestellt. Vorläufige Untersuchungen zeigen, dass infolgedessen die Meereisfläche in der Arktis um rund 1-2% niedriger, in der Antarktis aber womöglich um rund 1% höher ist.

According to them, preliminary investigations show, that due to the change in processing from DMSP-F17 to DMSP-f18 (March - April 2016) sea ice area in the Arctic is 1-2% lower as before (in their record I assume); while in the Antarctic it might be 1% higher.
Does this have any influence on the comparisons above (or is this already factored in)?

Thanks!

The page is also available in English

And no, it is not relevant for this thread because they are different satellites. I am using data from the  AMSR2 instrument on the Japanese built and operated GCOM-W1 satellite.

DMSP-f17 and DMSP-f18 are American satellites, that carry a similar but lower spec instrument (measuring passive microwave radiation) called SSMIS .

Using SSMIS is one thing, how it is used another. NSIDC uses relatively low frequency bands , while UNI Hamburg's calculation is based on the higher frequency (about 90 GHz) bands. That may explain why NSIDC found no difference in sea ice concentration when Uni Hamburg did find a small difference in their SSMIS based product.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4086
    • View Profile
    • Arctische Pinguin
  • Liked: 830
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2360 on: July 18, 2016, 08:42:00 AM »
Here is that big floe breaking further.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4086
    • View Profile
    • Arctische Pinguin
  • Liked: 830
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2361 on: July 18, 2016, 09:19:02 AM »
Extent essentially unchanged and a slow decline in area

Update 20160717.

Extent: +1.4 (-394k vs 2015, -266k vs 2014, -93k vs 2013, +126k vs 2012)
Area: -41.9 (-389k vs 2015, -662k vs 2014, -453k vs 2013, -62k vs 2012)
 
You will find the updated graphs in the top post

Regional extent increased in the Beaufort: +30k. Other regions changed only little, Hudson most with -14k.

Regional area dropped big in the CAB (-96k). For compensation, Beaufort, Chukchi and ESS increased in area (+29k, +27k and +17k).

Regional delta map is the Arctic Basin where the cyclone is showing its presence.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4086
    • View Profile
    • Arctische Pinguin
  • Liked: 830
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2362 on: July 18, 2016, 09:26:12 AM »
Animation of the Greenland Sea ice cover. A little Fram Strait export may be visible, but the ice does not come very far.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4086
    • View Profile
    • Arctische Pinguin
  • Liked: 830
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2363 on: July 19, 2016, 10:20:51 AM »
After today's changes 2016 is still leading in area and second to 2012 (in my limited data set).

Update 20160718.

Extent: -51.8 (-313k vs 2015, -260k vs 2014, -47k vs 2013, +162k vs 2012)
Area: -98.6 (-360k vs 2015, -644k vs 2014, -441k vs 2013, -156k vs 2012)
 
You will find the updated graphs in the top post

Rather small changes in regional extent, ESS was highest with -18k.

Regional area changes are bigger, CAB (-29k) and ESS (-27k) declined most while Greenland Sea increased by +28k.

Delta map is from the Beaufort section. Lots of reds and blue's from moving floes. Open water is appearing deeper into the pack.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4086
    • View Profile
    • Arctische Pinguin
  • Liked: 830
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2364 on: July 19, 2016, 10:25:30 AM »
Animation is centered over Chukchi. Big increases in extent happened here in the last few days but seem to have ended here. The reason can be seen to bee dispersion, the ice in the ice pack is getting more and more open.

Steven

  • Grease ice
  • Posts: 618
    • View Profile
  • Liked: 194
  • Likes Given: 17
Re: Home brew AMSR2 extent & area calculation
« Reply #2365 on: July 19, 2016, 07:00:12 PM »
After today's changes 2016 is still leading in area and second to 2012 (in my limited data set).

Update 20160718.

Extent: -51.8 (-313k vs 2015, -260k vs 2014, -47k vs 2013, +162k vs 2012)
Area: -98.6 (-360k vs 2015, -644k vs 2014, -441k vs 2013, -156k vs 2012)
 
You will find the updated graphs in the top post

Looking at the legends of the extent and area graphs in the top post of this thread:



your data for 2012 are from a different source than the 2013-2016 data.  How exactly does this affect your comparisons with 2012?

I searched through the comments on the first few pages of this thread and found a few comments from June 2013 about this issue, e.g. Reply #108.  This suggests to me that the difference of resolution (3.125km vs. 12.5km) doesn't matter much for extent, but it matters more for area.  Changing the resolution (from 3.125km to 12.5km) could easily increase the calculated sea ice area by 0.2 million km2 or so during the melt season. 

Is there a UH SSMIS 12.5km dataset available for the 2016 melt season?  If so, how this it compare to your calculations with the UH AMSR2 3.125km data for 2016?

AmbiValent

  • Frazil ice
  • Posts: 136
    • View Profile
  • Liked: 7
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2366 on: July 19, 2016, 07:33:33 PM »
AMSR2 data starts near the end of July. Will you use those numbers from then on?
Bright ice, how can you crack and fail? How can the ice that seemed so mighty suddenly seem so frail?

RoxTheGeologist

  • Grease ice
  • Posts: 501
    • View Profile
  • Liked: 149
  • Likes Given: 112
Re: Home brew AMSR2 extent & area calculation
« Reply #2367 on: July 19, 2016, 08:12:41 PM »


Observations about bathymetry in #2349 and #2357 have to be very relevant. I wonder if the current position of the North Atlantic front is near its northernmost maximal extent, ie the drop-off at the edge of the continental shelf which limits the range of warm surface water.

Wow ... the shelf line-up with the current advance of North Atlantic waters is very impressive suggesting that the front could have already reached a limiting position except nearer to the large Russian island complex Severnaya Zemlya to the east where open water could still come farther polewards with getting beyond the shelf.


A-Team, that is an amazing observation! It would suggest that the Atlantic waters drops well below the Arctic Ocean surface. The bathymetry changes from 300 m on the shelf to over 3000 m at the abyssal plane in a few km. The obvious conclusion is this is the source of at least some of the Atlantic Water in the Arctic Basin (the layer at 200-900m). The Russian Island complex is 'protected' by the St Anna Trough, where the Atlantic waters flow into the Arctic before they reach the Siberian shelf.

http://polardiscovery.whoi.edu/arctic/circulation.html

http://www.nature.com/scitable/knowledge/library/arctic-ocean-circulation-going-around-at-the-102811553


A-Team

  • Young ice
  • Posts: 2556
    • View Profile
  • Liked: 432
  • Likes Given: 29
Re: Home brew AMSR2 extent & area calculation
« Reply #2368 on: July 19, 2016, 08:13:03 PM »
The animation looks further at what is going on at the polar front between the ice pack and North Atlantic, which is advancing/retreating, and whether there is rapid melt at the front. The dates shown are 13-18 July 2016. The front has been fairly stable, the ice pack has ceased advancing, melt can be seen but floe obliteration is modest.

Since the front now lies directly over the bathymetric plunge of the continental plunge, the default hypothesis would be that denser North Atlantic water descends downslope, leaving nothing to move the melting front poleward. In this scenario the polar front is at its maximal position.

However the oceanographic literature provides meagre support for this. The focus of research has been almost entirely on currents in the Fram Strait (East Greenland, West Spitzbergen, Atlantic Returning currents) and bottom topographically driven flow on the Yermak Plateau (4th frame).

The Fram plays an outsize role because it is the only entrance to the Arctic Ocean with a deep sill and dominates the sverdrups circulating below the ice pack. The shallow Barents has a modest Bear Island current east of Svalbard, whereas Kara Sea is relatively out of the loop. Thus the position of the polar front in these regions may merely reflect waters warmed by seasonal sunlight and atmosphere, rather than a boundary defined by descending currents.

Still, surface waters at the edge of ice pack are freshened by its melt and so buoyant relative to saline waters of the North Atlantic. The Hycom view of salinity dilution (frame 9) captures this effect quite plausibly. Wave action would mix melt waters to fairly shallow depths, increasing density as offset by temperature expansion, quite warm according to Hycom sea surface temperature (frame 10).
« Last Edit: July 19, 2016, 09:02:43 PM by A-Team »

FishOutofWater

  • Grease ice
  • Posts: 831
    • View Profile
  • Liked: 441
  • Likes Given: 151
Re: Home brew AMSR2 extent & area calculation
« Reply #2369 on: July 19, 2016, 08:34:36 PM »
The "warm" Atlantic water layer pretty much covers the Arctic ocean at a depth of 300m with a thickness of hundreds of meters. The Atlantic water enters via the Fram. Clearly, the Atlantic water entering at the Fram isn't just cooling and sinking into the depths of the Arctic ocean although there is some evidence in the literature of intermediate-deep water formation increasing along that shelf margin. It certainly isn't going to sink at this time of year when mixing tends to freshen it and the weather is warming it.

be cause

  • Nilas ice
  • Posts: 1317
    • View Profile
  • Liked: 575
  • Likes Given: 420
Re: Home brew AMSR2 extent & area calculation
« Reply #2370 on: July 19, 2016, 08:40:24 PM »
dear A-team is there any way to stall this medley of images so that a slow brain like mine can appreciate them rather than wanting to scream at the overload delivered in a few seconds ? Otherwise love your contributions ... bc .
2007 + 5 = 2012 + 4 = 2016 + 3 = 2019 + 2 = 2021 
 (phew)

A-Team

  • Young ice
  • Posts: 2556
    • View Profile
  • Liked: 432
  • Likes Given: 29
Re: Home brew AMSR2 extent & area calculation
« Reply #2371 on: July 19, 2016, 09:01:46 PM »
Quote
slow animation?
Sure. They're now slowed to 2400 ms from 1800ms. The sequence of AMSR2 3k is the same to retain the sense of motion there, it is repeated a second time within an animation run. About half the Modis half-frames were too cloudy to have any value to microwave interpretation.

Better, download the gif and study individual frames at your leisure. Even though ms delay is part of the .gif standard, it may not be implemented as specified on some devices.

The Google Earth implementation of IBCAO bathymetry is a little sketchy on the colors but it does seem to allow accurate point and transect measurements of sea floor depth which the IBCAO poster does not. The 'txt' file attached shows the template for the second transect -- just replace the coords with something built in a spreadsheet and change to 'kml' to open in GE.

The GE projection can unsurprisingly not be successfully extended over the whole Arctic Ocean to align with that of WorldView and AMSR2; on the other hand, some satellite portals provide kml export of imagery that imports into GE. Below the north poles and Bering Strait were used to set the rescaling but elsewhere the agreement is so-so.

Quote
Intermediate waters

Right, incoming water sinks only to density-appropriate depth. The Arctic Ocean has various basins such as the adjacent Nansen where water is trapped by sills on both side. It seems counter-intuitive that Atlantic Water could circulate for years at intermediate depth before exiting more or less unmixed (eg through Nares Strait) but the experimental evidence for that is persuasive.
« Last Edit: July 19, 2016, 09:40:57 PM by A-Team »

ktonine

  • Frazil ice
  • Posts: 363
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2372 on: July 19, 2016, 11:44:00 PM »


Observations about bathymetry in #2349 and #2357 have to be very relevant. I wonder if the current position of the North Atlantic front is near its northernmost maximal extent, ie the drop-off at the edge of the continental shelf which limits the range of warm surface water.

Wow ... the shelf line-up with the current advance of North Atlantic waters is very impressive suggesting that the front could have already reached a limiting position except nearer to the large Russian island complex Severnaya Zemlya to the east where open water could still come farther polewards with getting beyond the shelf.


A-Team, that is an amazing observation! It would suggest that the Atlantic waters drops well below the Arctic Ocean surface. The bathymetry changes from 300 m on the shelf to over 3000 m at the abyssal plane in a few km.

The effects of both the Lomonsov Ridge and the Gakkel Ridge have been discussed here and on the blog before; back around 2014 -- notably by Chris Reynolds and Rob Dekker IIRC.

Arctic Ocean boundary currents are explained on the Univ. of Washington's Polar Science Center website.

And from Nghiem et al, Seafloor Control On Sea Ice, 2009 we read:

Quote
"The seafloor structures can dictate where perennial sea ice resides across the Arctic Ocean: over the deep basin off the continental shelf of the Laptev Sea, the Kara Sea, and Barents Sea, and over the continental shelf of the Beaufort Sea. This distribution of perennial sea ice is coincidentally located along the primary path of the Transpolar Drift. In strong Polar Express events driven by strong wind forcing supported the atmospheric dipole anomaly (Nghiem et al., 2007), an excessive transport of perennial sea ice out of the Arctic via the Fram Strait can results in massive ice loss since perennial ice is the older and thicker type of sea ice compared to first year or seasonal sea ice. While the Polar Express is driven by anomalous winds, the direction ofthe ice transport along the Transpolar Drift is not arbitrary as it is limited in the deep transpolar basins such as the Nansen Basin, and aligned along Lomonosov Ridge across the Arctic Ocean (Alkire et al., 2007). Such bathymetry control on the location of perennial sea ice sets up a condition conducive to an abrupt reduction of perennial sea ice, which in turn preconditions the large decrease of the total SIE during summer."

seaicesailor

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2373 on: July 20, 2016, 12:45:39 AM »
Quote
slow animation?
...
The Google Earth implementation of IBCAO bathymetry is a little sketchy on the colors but it does seem to allow accurate point and transect measurements of sea floor depth which the IBCAO poster does not. ...
The similarity of the Sep 2012 Extent with the limits of the continental shelves is striking!
- The only two regions where large extents of open water lie over deep ocean are Beaufort sea and Laptev Sea. Laptev sea ice is thin after winter due to the transpolar drift that forces ice to refreeze constantly. Beaufort sea is also prone to developing open water due to Beaufort High in spring, and also imports heat from the Pacific transported via atmosphere and ocean. Ocean currents out of Bering and entering Chukchi and Beaufort is a world of science itself, trying to understand the basics. I suspect currents have much to do with ice loss especially when the sea ice opens up soon (but that requires more digging).
- The only region where sea ice stays over shallow shelf is Greenland Sea, the reason being the steady import it receives from the Arctic proper (at least in years within 2007-2012) (Edit: CAA and a band of CAB along the coast of Greenland are also shallow shelves).

So, is it the September 2012 extent a boundary that will be very difficult to cross due to this fact and why? No idea. We however have had this boundary being crossed in the past (2007 went way further into the CAB from Chukchi, 2014 Laptev bite made its way to the 85N latitude). 2007 was exceptional in its sustained heat transport from the Pacific side, and 2014 Laptev bite was also exceptional. We can expect a very exceptional melting season in all the fronts eventually, but can the melt front go very far into the basins and why or why not?

Press and blogs in general do not mention this fact and I find it extremely revealing.
« Last Edit: July 20, 2016, 01:01:54 AM by seaicesailor »

seaicesailor

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2374 on: July 20, 2016, 02:03:30 AM »
This Sea Ice News and Analysis Nov 2012 described briefly this observation

http://nsidc.org/arcticseaicenews/tag/bathymetry/

Rob Dekker

  • Nilas ice
  • Posts: 2386
    • View Profile
  • Liked: 123
  • Likes Given: 119
Re: Home brew AMSR2 extent & area calculation
« Reply #2375 on: July 20, 2016, 09:15:12 AM »
Thank you seaicesailor, that is a great summary.
From that NSIDC article :

Quote
Research by our colleagues Jamie Morison at the University of Washington Seattle and NASA scientist Son Nghiem suggests that bathymetry (sea floor topography) plays an important role in Arctic sea ice formation and extent by controlling the distribution and mixing of warm and cold waters. At its seasonal minimum extent, the ice edge mainly corresponds to the deep-water/shallow-water boundary (approximately 500-meter depth), suggesting that the ocean floor exerts a dominant control on the ice edge position.
which supports the observations by you and A-Team that the ice edge on the Atlantic side is bounded by the bathymetry, due to the ocean currents.



On the Pacific side, such bathymetry-driven currents are less pronounced, and do not include external heat input by ocean currents, which suggests that the Pacific side may be more vulnerable to weather and AGW than the Atlantic side.
« Last Edit: July 20, 2016, 09:38:11 AM by Rob Dekker »
This is our planet. This is our time.
Let's not waste either.

seaice.de

  • New ice
  • Posts: 86
    • View Profile
  • Liked: 32
  • Likes Given: 3
Re: Home brew AMSR2 extent & area calculation
« Reply #2376 on: July 20, 2016, 10:46:25 AM »
Is there a UH SSMIS 12.5km dataset available for the 2016 melt season?  If so, how this it compare to your calculations with the UH AMSR2 3.125km data for 2016?

Not exactly at UH but IFREMER does the processing with the same algorithm (Kaleschke et al. 2001, see link below) and their results should be very similar:

ftp://ftp.ifremer.fr/ifremer/cersat/products/gridded/psi-concentration/data/arctic/daily/netcdf/2016/

https://icdc.zmaw.de/fileadmin/user_upload/icdc_Dokumente/kaleschkeetal_SSMISeaIceRemoteSensing_CJRS_2001.pdf

Bill Fothergill

  • Frazil ice
  • Posts: 278
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2377 on: July 20, 2016, 11:20:32 AM »
@ Seaicesailor,

I would like to echo Rob's appreciative comments. Not sure how it happened, but I don't think I even saw that issue of Sea Ice New & Analysis.

There are obviously many pay-walled articles on this topic, but there is an open-access article on the NASA site entitled "Seafloor Control on Sea Ice"...
http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20120003985.pdf

Quoting briefly from the Summary...

"Bathymetry plays an important role in the formation and evolution of sea ice from local areas to
extensive regions in the Arctic Ocean. In summary, we have found that:

1. In the Bering Sea, the maximum SIE reaches the shelf break in spring (case study in
March 2008). Sea ice forms and remains over the shallow bathymetry on the continental
shelf. The ice edge aligns with the shelf-break bathymetry spanning over 1200 km.

2. In the Chukchi and Beaufort Seas, the sea ice edge is close to the shelf break during the
freeze-up transition from the summer to the fall season (case study in September to
November 2009). In this case, sea ice preferentially remains over the deep bathymetry as
opposed to the case of the Bering Sea. As the sea ice survives the summer melt and
becomes perennial ice, the boundary of the perennial ice and other ice classes continues
to conform to the vicinity of the shelf break across an extensive length of 1670 km.

3. In the Barents and Greenland seas, the bathymetry control on the sea ice cover is evident
from sea ice maps during the decade of 2000s. For the Barents Sea, two distinctive cases
are discovered from the sea ice patterns: Sea ice edges in the north of the Central Bank in
years 2000, 2001, and 2005-2009, and sea ice edges over the Central Bank and extended
further south in years 2002-2004. Additionally, in 2003 and 2004, the Odden ice tongue
was extensive and well defined within the limit of Jan Mayen fracture zone and the Mohns Ridge. Overall, sea ice resides preferentially over the continental shelf in the Greenland Sea.

4. Across the Arctic Ocean, perennial ice occupied the shallow region of the Greenland
continental shelf, spanned across the deep channel in the Fram Strait, and covered the
deep basin off the shelves of the Barents, Kara, and Laptev Seas. This conformation
spans over an extensive distance of 3700 km across the Arctic Ocean."



Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4086
    • View Profile
    • Arctische Pinguin
  • Liked: 830
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2378 on: July 20, 2016, 01:41:27 PM »
This suggests to me that the difference of resolution (3.125km vs. 12.5km) doesn't matter much for extent, but it matters more for area.  Changing the resolution (from 3.125km to 12.5km) could easily increase the calculated sea ice area by 0.2 million km2 or so during the melt season. 

Is there a UH SSMIS 12.5km dataset available for the 2016 melt season?  If so, how this it compare to your calculations with the UH AMSR2 3.125km data for 2016?

Extent should be most sensitive to resolution, area very little. But there are always more differences. Here the satellite/instruments are different, slightly different microwave frequencies. Most data sets have different opinions about what is land and what is ocean (different landmasks).

Attached are the Uni Hamburg's SSMIS and ASMR2 extent and area in 2013 and 2014.
2014 is when the data stream stopped. The mismatches during winter 2013/14 are caused by wrong masks.

@ seaice.de

Thanks, I did not know IFRAMER. I will have a look.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4086
    • View Profile
    • Arctische Pinguin
  • Liked: 830
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2379 on: July 20, 2016, 01:41:52 PM »
Did 2016 find the fast lane again? The lead in area is growing again.

Update 20160719.

Extent: -102.9 (-274k vs 2015, -340k vs 2014, -123k vs 2013, +78k vs 2012)
Area: -175.9 (-448k vs 2015, -763k vs 2014, -542k vs 2013, -392k vs 2012)
 
You will find the updated graphs in the top post

Extent drops were largest in ESS (-30k), CAB (-21k) and Baffin (-18k).

CAB area fell -84k, Baffin and ESS follow at -28k and -21k.

Regional delta map is Baffin, all the ice in the south has gone. Near Baffin Island some ice that is fast going away.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4086
    • View Profile
    • Arctische Pinguin
  • Liked: 830
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2380 on: July 20, 2016, 01:49:34 PM »
Todays animation is the East Siberian Sea. Ice is both moving and melting, also here the ice pack is not very compact.


Richard Rathbone

  • Grease ice
  • Posts: 766
    • View Profile
  • Liked: 123
  • Likes Given: 10
Re: Home brew AMSR2 extent & area calculation
« Reply #2381 on: July 20, 2016, 01:50:42 PM »
Thats a big bite out of the CAB if it follows through.

wanderer

  • Frazil ice
  • Posts: 156
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2382 on: July 20, 2016, 02:33:40 PM »
Well, 400k areawise below 2012... that's a statement!

A-Team

  • Young ice
  • Posts: 2556
    • View Profile
  • Liked: 432
  • Likes Given: 29
Re: Home brew AMSR2 extent & area calculation
« Reply #2383 on: July 20, 2016, 05:30:32 PM »
Nice find by seaicesailor. Both those SV Nghiem articles on bathymetry control of surface melt are free full text. However the authors run into the same problem we have, correlation is not causality:

The continental shelf forms the southern perimeter and the ice melts from south (where it's warmer) inwards to the cold pole. So naturally there will be good alignment of the ice front with bathymetry even without any physical process connecting them.

In terms of warm saline North Atlantic waters mixing with cold fresh Arctic waters, a crash course (deep dive?) into oceanography soon turns up double diffusion, cabelling and the Taylor arctan that connects them, notably in the Fram where the southbound East Greenland Current is adjacent to the northbound West Spitsbergen.

Because isopycnic lines are curved on the temperature vs salinity plot, mixing waters of equal density but different salinities and temperatures results in a lower density (ie lies along a chord of the isopycnic arc) which then sinks. It's not clear however that this would be operative at the polar front north of Svalbard and Severnaya Zemlya.

Looking at actual temperature, temperature anomaly and ocean current data (OSCAR @ nullschool as a check on Hycom), even though 80ºN seems to be a cutoff, doesn't reveal a mixing mechanism. The one notable head-on current displayed may be river discharge which should receive support from sea surface salinity but does not.

Looking directly at Modis for potentially bathymetric-driven mixing processes such as currents across the front, there should be eddies seen in streamers of brash ice coming off the sea ice. We've seen eddies way south off Greenland, eddies in Petermann ice, surfaces eddies off the Alaskan coast, and Taylor columns above Herald Shoal but there's no indication of them along the polar front.

The AMSR2 3k animations above show a lack of correspondence (negative correlation) of sea ice surge features with shallow bathymetric features such as the Franz Victoria Trough, Hinlopen Basin and the Bear Island obstruction (though the latter is said to have an ice tongue caused by local bathymetry).

There's no bottom line, just tracking this interesting phenomenon to see what develops along the melting front over the rest of the summer.

http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20120003985.pdf
http://linkingweatherandclimate.com/ocean/waterdensity.php
http://www.ngu.no/glacipet/photos/internal/pdfs_of_articles/kleiber_knies_niessen2000.pdf
http://onlinelibrary.wiley.com/doi/10.1111/j.1751-8369.2007.00024.x/full
http://onlinelibrary.wiley.com/doi/10.1002/2015JC011290/full
http://www.whoi.edu/fileserver.do?id=189964&pt=2&p=194729
http://www-odp.tamu.edu/publications/151_IR/VOLUME/CHAPTERS/ir151_01.pdf
http://sam.ucsd.edu/sio210/lect_5/lecture_5.html
http://www-pord.ucsd.edu/~sgille/pub_dir/TOS_Gille_etal_low_res.pdf
http://onlinelibrary.wiley.com/doi/10.1002/2015JC011375/abstract
http://journals.ametsoc.org/doi/pdf/10.1175/2010JPO4371.1

Phil.

  • Frazil ice
  • Posts: 354
    • View Profile
  • Liked: 46
  • Likes Given: 1
Re: Home brew AMSR2 extent & area calculation
« Reply #2384 on: July 20, 2016, 05:44:15 PM »
Thanks, I did not know IFRAMER. I will have a look.

Just to avoid future problems it's ifremer.

Andreas T

  • Nilas ice
  • Posts: 1143
    • View Profile
  • Liked: 23
  • Likes Given: 4
Re: Home brew AMSR2 extent & area calculation
« Reply #2385 on: July 20, 2016, 06:21:36 PM »
....
 and the Bear Island obstruction (though the latter is said to have an ice tongue caused by local bathymetry)......

?????
look again, the ice tongue is a feature I have seen frequently. You show the temperature anomaly associated with that trough, and have pointed to it this winter. When there is ice south of Svalbard as there was occasionally this winter the southern edge follows that line fairly clearly.
image from 26th March 2016 http://go.nasa.gov/29Z08gT You'll recognize AMSR2 ice concentration in worldview.

What we need to look for is water profiles north of Spitsbergen I think. I have seen profiles from the Northeast corner of Greenland which show the pattern found in the central arctic of cold water over more saline, warmer water at depth of over 100m .
The fact that there are no eddies north of Spitsbergen has possibly to do with the sinking of the warmer saltier altlantic water there. I can imagine the shear layer being subducted with that sinking water taking momentum out of potential eddies. Ice drift is westward along the edge at times but often also dominated by wind, that was my impression when I spent some time earlier this year looking at floes there (see Svalbard thread)

RoxTheGeologist

  • Grease ice
  • Posts: 501
    • View Profile
  • Liked: 149
  • Likes Given: 112
Re: Home brew AMSR2 extent & area calculation
« Reply #2386 on: July 20, 2016, 07:02:22 PM »

I found a couple of interesting papers on how the Barents sea, the later discusses the Atlantification of

http://onlinelibrary.wiley.com/doi/10.1002/rog.20017/full

and

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.686.9429&rep=rep1&type=pdf






FishOutofWater

  • Grease ice
  • Posts: 831
    • View Profile
  • Liked: 441
  • Likes Given: 151
Re: Home brew AMSR2 extent & area calculation
« Reply #2387 on: July 20, 2016, 07:20:03 PM »
The warm salty water of the Atlantic water layer moves into the Arctic ocean along the margin of the continental shelf at a depth of about 300m. It's quite warm as it moves out of the Fram strait but cools as it flows towards the European Arctic. However, the incoming water retains its salinity so its salinity is an easy way to trace the new Atlantic water. This year the water flowing in has been warmer than normal, leading to rapid surface melting of the ice that is pushed over it by the transpolar drift. The Coriolis effect tends to keep the inflowing Atlantic water close to the shelf break.

Over the central Arctic the Atlantic water layer has been mixed with colder fresher waters already in the Arctic so it has both lower salinity and lower temperature than the new Atlantic water flowing in along the shelf break. It's going to take a lot of solar energy to melt ice out over the water nearer the pole because it isn't going to get much heat from the water below. The ice and water at the end of the transpolar drift, about to depart through the Fram, is enthalpy depleted.

Thanks for the references. Good discussion.

Note that upwelling of the Atlantic water layer generally occurs along the shelf break when strong high pressure sets up over the Beaufort sea or over the pole for extended periods of time. Strong upwelling this spring helped begin this year's melt season early.

seaicesailor

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2388 on: July 20, 2016, 09:41:47 PM »
Thank you very much Bill, A-Team, and Rob! 
@Rob related to this (and very off-topic so I keep it short) I was convinced that a big clash of ocean currents was happening in Beaufort because of these eddies that A-Team shows and we have been looking for days (and there are more in the picture, plus those shown by HYCOM). I was convinced that the open water around the defunct Big Block was caused not only by insolation but also by warmer water from the Pacific. But in this case in particular the simplifications that one usually entertains won't work at all, so I have to keep my suspicions for myself.
I have been looking these days for material in layman terms. There is this nice presentation on the Bering inflow by R. Woodgate, http://psc.apl.washington.edu/HLD/Bstrait/Woodgate_AONSeattleNov2015_17thNov2015.pdf , then lecture notes on the different types of currents along shelves http://www.mt-oceanography.info/ShelfCoast/chapter09.html
Also a very instructive page on the Chukchi-Beaufort currents, http://dm.sfos.uaf.edu/chukchi-beaufort/ I suspect that ironically this research might be funded by oil money, if you want to drill you better know your shelfbreak jets.

Glenn Tamblyn

  • Frazil ice
  • Posts: 128
    • View Profile
  • Liked: 1
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2389 on: July 21, 2016, 02:54:15 AM »
Looking at Wipneus' graphs, it really seems to be the Laptev vs all the rest. All the other regions are at or ahead of usual. And in the Laptev it is the most southerly, coastal part that seems to be lagging.

The naive view might be that the peripheral seas need to melt out before the CAB can start to change. However this year is turning that on its head. The CAB is seeing action across a large area, even the melt in the Laptev is still producing open water nearer the boundary with the CAB. If all that action in the CAB translates to significant concentration drops while the sun is still shining, allowing CAB waters to warm, we might see some dramatic changes around mid August with the coastal part of the Laptev melting out late, but it will still have contributed to giving the CAB a decent start.

Passing 2012 is still quite possible.

oren

  • Moderator
  • First-year ice
  • Posts: 5894
    • View Profile
  • Liked: 1995
  • Likes Given: 1767
Re: Home brew AMSR2 extent & area calculation
« Reply #2390 on: July 21, 2016, 09:15:16 AM »
Passing 2012 is still quite possible.

Agree with your analysis, and the fact that this is even possible given the mediocre melting weather speaks for itself.

A-Team

  • Young ice
  • Posts: 2556
    • View Profile
  • Liked: 432
  • Likes Given: 29
Re: Home brew AMSR2 extent & area calculation
« Reply #2391 on: July 21, 2016, 05:28:47 PM »
Quote
Laptev vs all the rest
The animation below looks at UH AMSR2 3k over 16-20 Jul 2016 to see if the Laptev and other areas showing 100% sea ice concentration are really in that condition or rather consistently affected by heavy cloud cover artifacts.

The method here removes clouds by stacking the daily images in chronological order (ie July 20th on top), then defining a relationship with the date below by pixel multiplication. The resultant view is defined as (top*bottom)/255 which results in a darker pixel unless one of the images is pure white at that pixel (255 grayscale or 100% concentration).

This has the effect of taking out the clouds by substituting the minimum sea ice concentration, that of the date least obscured. Over the short time frame of five days, effects of ice movement, compaction and dispersion will contribute blur, overstating the final product which becomes in effect just a lower bound or worst case scenario of sea ice concentration.

The final frame shows regions of the Arctic Ocean (notably the Laptev, Prince Patrick and polar front area north of Svalbard) that were 100% concentration (red) over the entire five day period. The north pole and CAA seem in remarkably poor condition for this date when not viewed with rose tinted glasses as upper bound (uncorrected AMSR2 3k).
« Last Edit: July 21, 2016, 11:47:47 PM by A-Team »

be cause

  • Nilas ice
  • Posts: 1317
    • View Profile
  • Liked: 575
  • Likes Given: 420
Re: Home brew AMSR2 extent & area calculation
« Reply #2392 on: July 21, 2016, 06:09:43 PM »
A-team .. your final frame shows remarkable consistency with the bremen AMSR2 image I was looking at earlier . Looks like both are worth trusting . Sadly both show most thicker ice destined to dissappear and my May 12th prediction of an ice-free North Pole later this season is looking increasingly likely .
2007 + 5 = 2012 + 4 = 2016 + 3 = 2019 + 2 = 2021 
 (phew)

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4086
    • View Profile
    • Arctische Pinguin
  • Liked: 830
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2393 on: July 21, 2016, 06:55:34 PM »
We are getting now in the time of year where extent of the years in my limited data set are very close together (within 400k or less). 2016 falls nicely in that range.

Update 20160720.

Extent: -99.2 (-271k vs 2015, -362k vs 2014, -182k vs 2013, +20k vs 2012)
Area: -82.1 (-450k vs 2015, -752k vs 2014, -572k vs 2013, -238k vs 2012)
 
You will find the updated graphs in the top post

Regional declines are nowhere very remarkable, Chukchi, CAA, ESS and Hudson all dropped around -18k.

Regional area declines are biggest in Greenland Sea (-42k) and CAA (-36k).

Regional CAB area is making a very early drop as has been noticed elsewhere. You could pose the question how significant that is. Borders of the CAB are sometimes logical following geographic features (CAA, Barents, Greenland Sea), some places there is some  arbitrary choice in it (borders with Beaufort, Chukchi, ESS and Laptev).

An alternative would be to look not just at the CAB but at the Arctic Basin, including CAB, Beaufort, Chukchi, ESS and Laptev. That graph is attached below. It is instructive to compare the three data sets used. The UH AMSR2 has 2016 as the lowest, NSIDC has 2016 somewhere in the middle and Jaxa AMSR2 puts it as the highest.

What is the truth? Normally I would regard Jaxa AMSR2 as the most accurate of the three methods. The rankings change from day to day, perhaps the cyclonic disturbances still have to settle. We see how it develops.

Oh, about AB extent the methods agree: extent is among the highest for the date.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4086
    • View Profile
    • Arctische Pinguin
  • Liked: 830
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2394 on: July 21, 2016, 07:00:24 PM »
Animation is of the Greenlands Sea, where the ice is having a hard time. The fast ice in NE Greenland can been seen dropping a little notch at the top. I expect bigger losses soon.

A-Team

  • Young ice
  • Posts: 2556
    • View Profile
  • Liked: 432
  • Likes Given: 29
Re: Home brew AMSR2 extent & area calculation
« Reply #2395 on: July 21, 2016, 09:53:22 PM »
Quote
I expect bigger losses soon
Considering the scale of 2.4 km per meridional pixels, retreat along the broad polar front (yellow and blue band) has been substantial over the last five days.
« Last Edit: July 21, 2016, 09:58:40 PM by A-Team »

jdallen

  • Young ice
  • Posts: 3194
    • View Profile
  • Liked: 402
  • Likes Given: 200
Re: Home brew AMSR2 extent & area calculation
« Reply #2396 on: July 21, 2016, 10:01:16 PM »
Quote
I expect bigger losses soon
Considering the scale of 2.4 km per meridional pixels, retreat along the broad polar front (yellow and blue band) has been substantial over the last five days.
<holds up thumb and squints>  30-50KM retreat along a 1500KM or so front maybe?
This space for Rent.

Andreas T

  • Nilas ice
  • Posts: 1143
    • View Profile
  • Liked: 23
  • Likes Given: 4
Re: Home brew AMSR2 extent & area calculation
« Reply #2397 on: July 21, 2016, 11:01:38 PM »
TERRA confirms the weakening of the landfast ice on NE Greenland coast. Apart from the notch there is also a hole opening where the ice is thin.
http://go.nasa.gov/2acdEvS

A-Team

  • Young ice
  • Posts: 2556
    • View Profile
  • Liked: 432
  • Likes Given: 29
Re: Home brew AMSR2 extent & area calculation
« Reply #2398 on: July 21, 2016, 11:20:49 PM »
Quote
ice-free North Pole later this season?
few dozen km retreat over five days?
The 21st was fairly clear around the pole, as seen in the respective 367 cloud cover frames of the animation which has suomi, aqua, and terra imaging in succession. Moderate ice movement is detectable even at the short time intervals involved. The scale of the scene is 0.5 km per pixel making dimensions 350 x 350 = 122500 km2. http://go.nasa.gov/2acfRYf to replicate.

The WorldView projection EPSG3413 is not equal area so counting retreat pixels to average retreat over the entire front is problematic. That front spans ~90º (Zachariae to Severnaya Zemlya) or a quarter of the length of the 80th parallel which is 6981/4 = 1745 km so 17,450 km2 per 10 km of retreat.
« Last Edit: July 21, 2016, 11:34:21 PM by A-Team »

ktonine

  • Frazil ice
  • Posts: 363
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2399 on: July 21, 2016, 11:59:28 PM »
The animation below looks at UH AMSR2 3k over 16-20 Jul 2016 to see if the Laptev and other areas showing 100% sea ice concentration are really in that condition or rather consistently affected by heavy cloud cover artifacts.

The animation reflects the fact that the arctic has been drawn and quartered.  There are 4 obvious areas of high concentration each in its own quadrant. 

The ice north of 85 degrees will likely reach its lowest concentration sometime during the 2nd week of August -- so there's still another two to three weeks before we start seeing refreeze overtake melt (as far as extent and area are concerned) in the core around the north pole. The low concentration swaths that have criss-crossed the central arctic will get even weaker --unless we see a large compaction event.