Support the Arctic Sea Ice Forum and Blog

Author Topic: Jakobshavn Isbræ / Ilulissat Isfjord / West Greenland  (Read 575564 times)

solartim27

  • ASIF Citizen
  • Posts: 438
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #650 on: June 02, 2015, 06:52:03 AM »
Wind compacted the melange against the calving front??  Looking at Modis, there hasn't been any fresh snow in the area.
FNORD

Wipneus

  • ASIF Governor
  • Posts: 3641
    • View Profile
    • Arctische Pinguin
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #651 on: June 02, 2015, 08:05:06 AM »
I have some difficulties interpreting the Sentinel images from June 1, maybe Wipneus can help?
I know there were some calving activities lately.

I am not sure what I can do. There is for 1 June only a EW med-res image available, just good enough to be not entirely useless. Wait for a IW mode scan, with a higher resolution and much less speckle noise. These are taken much more regularly of JacobsHavn compared to e.g. Zachariae (waiting for months now).

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #652 on: June 02, 2015, 04:38:56 PM »
The thing to remember is that neither Landsat-8 and Sentinel-1 is single purpose. Polar scenes by their nature have very poor intrinsic contrast (off-white on white). Further, some users need the actual instrumental luminance values (eg  band 10 Landsat thermal for ground temperature). Together this means the products provided are not optimally adjusted as downloaded -- a jungle scene has different issues from an ice stream.

The first round of contrast enhancement is usually histogram normalization. That is, the image is not making full use of  dynamic range (eg 20-190 instead of 0-255 on an 8-bit). That's easily fixed by adding a constant and multiplicative rescaling (brightness and contrast knobs).

That is often followed by a second order adjustment of mid-tones (gamma correction). This can greatly improve the image visually but in general loses the original 'scientific' content. Contrarily, functions like log or sq root can have a logical rationale but they seldom produce a favorable visual outcome. (Gimp has a plugin that allows *any* function to be implemented as its power series expansion.)

The next level of contrast enhancement is histogram equalization. This uses the cumulative distribution function (see wikipedia) of pixel intensities to flatten the histogram (ie expand contrast for well-represented pixel values). This is quick and very effective on both Landsat and Sentinel but using the global histogram again means discarding neighborhood information.

Thus neither method is optimal for detecting a developing fracture (Espen's goal in #645) since neighboring pixel correlation is important to that.

It's better here to start with adaptive histogram equalization (CLAHE, see wikipedia). That is provided in Fiji/ImageJ2 at the bottom of the 'Process' menu; see http://fiji.sc/wiki/index.php/Enhance_Local_Contrast_%28CLAHE%29  The image below shows its effect on Landsat's low-budget preview image that Espen likes because of its small file size.

It is imperative to work within the 16-bit world of Landsat and Sentinel for initial contrast adjustments and co-registration rotations. Otherwise you will get unoccupied bars in the histogram.That cannot be currently done within Gimp. However it is easy to do a quick in-and-out within ImageJ and return to your usual image processor with an 8-bit after the benefits of 16-bit have been exhausted..

For a linear feature with predictable orientation (~parallel to the calving front), an edge detector would be the natural follow-on. These are typically directional derivatives or convolutions.

However nothing you can do will make a silk purse out of a sow's ear (Sentinel EW). Ultimately there's no substitute for a really high resolution image, one commensurate with the task at hand.

I'll post a 0.5 m resolution DigiGlobe image of growing fracture tips over at the Petermann forum. Even prior to polar-appropriate contrast enhancement, it's a huge improvement over the 15 m band 8 of Landsat, even if you've applied  CLAHE or retinex in Wipneus's pan-sharpening protocol.
« Last Edit: June 02, 2015, 08:32:45 PM by A-Team »

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #653 on: June 02, 2015, 06:10:18 PM »
I took a semi-quantitative look at the big calving event over May 26-28 first reported by Espen in #643. According to the 15 m resolution Landsat band 8, 13317 pixels @ 225 sq m per pixel calved off, so a little under 3 sq km (green tint below), roughly a third of that necessary to reach the 26 Sep 14 maximal retreat (red line).

This has to be slightly corrected for Jakobshavn forward motion of ~26 m per day by the 28th (measured 5 km upstream) which causes the front on the 28th to be too far forward in terms of image overlap, resulting in an under-estimate of area calved. We don’t know if the calving was a single event; if so, that would have been recorded as a westward berg velocity surge on the north side web cam.

Taking the overall thickness of ice at the calving front at 1400 m, this pencils out to 4.5 cu km of non-floating ice. When it melts, that results in 2.78 microns of smoothed-out sea level rise per cubic km, so 0.010 mm. Move along people, nothing to see here (?)

There are lots of good reasons why identification numbers should be provided in forum posts even if they seem too geeky. In this instance, the 2LC80100112015146LGN00 of 26 May 15 is path 10, row 11 just like the 2014 maximal retreat image LC80100112014271LGN00, whereas LC80080112015148LGN00 has a very different viewing geometry, namely path 8, row 11.

This means the images cannot be co-registered without a 5.53º rotation of the 28 May 15. There being no truly fixed features on the ice sheet, the last rock outcrop in the SW corner, together with the tip of the ice-fall separating north and south arms, are the best choices for subsequent registration. A third triangulation point to the east is needed — but not available — to assess residual geometrical distortion.

This particular event mostly tracked pre-existing fractures that can be seen in the 26 May 15 image. It’s not so clear where the next round of calving will occur — that’s a game better played at Petermann. Sentinel EW images (see post #645 and #653) are a few bricks short of a wall in terms of adequate crevasse resolution -- the 2nd image shows various unconvincing enhancements.

Espen

  • ASIF Governor
  • Posts: 3078
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #654 on: June 02, 2015, 07:02:08 PM »
A-Team

Nice to see you back in in bar ;)
Have a ice day!

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #655 on: June 02, 2015, 07:57:34 PM »
Espen, what are you doing still online? I thought your expedition would be camped in front of the Jakobshavn calving front by now enjoying frozen MRE* and fresh kelp. Maybe 2016 for me. That was a great idea ... certainly this last event would have been an unforgettable experience.

Unless the weather got bad:

http://www.theguardian.com/uk-news/2013/aug/19/charity-trekker-froze-death-greenland-storm

It seems that the film crew in 'Chasing Ice' spent 3 weeks on the outermost rocks waiting for their big event. If so, it might make more sense to helicopter in (or out) so less time is spent trekking up the fjord with 25 kg packs.

I wonder though if permits can really be obtained -- there is a lot of scientific gear fixed in place and important ongoing research. Those cameras are meant for taking tourist selfies? The best viewing might be from the last rock outcrop ... but not until late July and all the 'view' campsites could already be taken.

*http://www.amazon.com/Meals-Ready-Genuine-Military-Surplus/dp/B005I5ML36
« Last Edit: June 02, 2015, 08:27:43 PM by A-Team »

Rubikscube

  • ASIF Citizen
  • Posts: 247
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #656 on: June 02, 2015, 11:20:19 PM »
Speaking of campsites just got me wondering; how long will it take before the first rock peak through the ice sheet on the opposite side (east/northeast) of the main channel? I guess it has been of greater interest to measure the ice thickness in the main trench rather than the surrounding high ground, but is it possible to get good idea of how thin the ice is there and how long it will take before its gone?

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #657 on: June 03, 2015, 08:18:11 PM »
is it possible to get good idea of how thin the ice is there and how long it will take before its gone?

The thickness ice-to-bedrock for a particular year could be read off a Cresis radar overflight that crossed your area of interest. If there was data for multiple years, that would provide a trend.

However the ice to just north of the main Jakobshavn channel is far from stagnant so it wouldn't be the 'same' ice. The whole ice sheet is sliding downhill so the ice there now will be replaced with ice from the northeast.

Espen ran into that in #630 comparing a late season 28 Sep 2014 to an early season 12 May 14. As the main channel receded past the tongue of the northern icefall, its buttressing was removed and it spilled into the newly opened water.

I looked around for recent ice sheet velocities of this region, finding a Jan 2015 image in a nice Sentinel overview article at  http://tinyurl.com/ps7brk8 and another in a May 25 open source article on accurate balance velocites, http://www.geosci-model-dev.net/8/1275/2015/gmd-8-1275-2015.pdf

Look for the 'big bend' in Jakobshavn to orient yourself. Then note how much of its surroundings to the north are flowing into the system. I posted some speculation earlier about iceshed capture in the future. It would be fairly tricky to demonstrate this over a limited time interval, even ten years, because of differences in satellite sensors.
« Last Edit: June 04, 2015, 03:19:15 PM by A-Team »

Rubikscube

  • ASIF Citizen
  • Posts: 247
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #658 on: June 04, 2015, 12:57:03 PM »
Thanks A-Team. Wonderful as usuall.

nukefix

  • ASIF Citizen
  • Posts: 407
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #659 on: June 04, 2015, 07:41:31 PM »
I'd like to re-iterate that log-intensity is often the preferred way of viewing SAR data. This is how I do it in the Sentinel-1 toolbox:

1. Open/import S-1 product
2. Right-click on the Intensity (virtual) band and select "Linear to/from dB" to generate a new virtual band containing log-intensity.
3. Open the new log-intensity-band
4. Play around with histogram in Colour Manipulation toolwindow.

The image-files themselves contain amplitude so if you want to use other tools to manipulate the image you should do the log_intensity = log10(amplitude*amplitude) conversion as a first step.

BTW I hear that there are plans for covering the Greenland margin with S-1 in IW-mode every 12 days - this would be fantastic obviously!

Shared Humanity

  • ASIF Upper Class
  • Posts: 2315
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #660 on: June 05, 2015, 03:46:41 PM »
is it possible to get good idea of how thin the ice is there and how long it will take before its gone?

The thickness ice-to-bedrock for a particular year could be read off a Cresis radar overflight that crossed your area of interest. If there was data for multiple years, that would provide a trend.

However the ice to just north of the main Jakobshavn channel is far from stagnant so it wouldn't be the 'same' ice. The whole ice sheet is sliding downhill so the ice there now will be replaced with ice from the northeast.

Espen ran into that in #630 comparing a late season 28 Sep 2014 to an early season 12 May 14. As the main channel receded past the tongue of the northern icefall, its buttressing was removed and it spilled into the newly opened water.

I looked around for recent ice sheet velocities of this region, finding a Jan 2015 image in a nice Sentinel overview article at  http://tinyurl.com/ps7brk8 and another in a May 25 open source article on accurate balance velocites, http://www.geosci-model-dev.net/8/1275/2015/gmd-8-1275-2015.pdf

Look for the 'big bend' in Jakobshavn to orient yourself. Then note how much of its surroundings to the north are flowing into the system. I posted some speculation earlier about iceshed capture in the future. It would be fairly tricky to demonstrate this over a limited time interval, even ten years, because of differences in satellite sensors.

If you study the most accurate topography maps of Greenland, a large area of the ice sheet that feeds these two active walls that are flowing into the fjord is grounded below sea level. In fact this area of the sheet is far more vulnerable than the active wall that is to the north.

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #661 on: June 05, 2015, 03:55:56 PM »
Nukefix, useful instructions ... but can you modify the post above to include the best image you can make of the calving front? It takes but seconds to save a screenshot and hit the attachment button below.

At this point, the case has not been made for using Sentinel during between April and October. There is a very clear Landsat for 04 Jun 15 at EarthExplorer, LC80090112015155LGN00.

So we can make side-by-side comparison with your Sentinel image, I am attaching the 15 m Landsat bumped out to 10 m. Note too the 3 new papers out saying Greenland velocity measurements can be not only automated to a few seconds but also made as accurate if not better than with insar.

nukefix

  • ASIF Citizen
  • Posts: 407
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #662 on: June 08, 2015, 12:57:43 AM »
I can make an image of the terminus when I'm back at the office. SAR is still weather and and lighting-independent and can be used to monitor speeds during the polar night, which will never work with optical. S-1 IW cannot measure the speeds close to the terminus due to lack of sufficiently high resolution, but it can produce ice-sheet wide maps. If Landsat or Sentinel-2 can be used to patch holes in S-1 - based velocity estimates that would be just great, so both methods are needed!

Espen

  • ASIF Governor
  • Posts: 3078
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #663 on: June 11, 2015, 10:00:42 PM »
Jakobshavn update:

Both the southern and northern branch of Jakobshavn is now widening and sligthly retreating:
Have a ice day!

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #664 on: June 12, 2015, 12:35:49 AM »
The whole place is on the move. Looks quite different overall from last year. Yet weather has been colder longer. Maybe less snowpack to begin with?

Very clear 15 m LC80100112015162LGN00_B8 below

Shared Humanity

  • ASIF Upper Class
  • Posts: 2315
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #665 on: June 12, 2015, 01:52:44 PM »
The  crevassing  and thinning is so  apparent in these detailed images.

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #666 on: June 12, 2015, 03:41:58 PM »
Here are previous clear days with the same Landsat-8 geometry (path,row = 10,11). The animation seems to run without further clicking. It compares day 162 (11 Jun 15) with day 159 (08 Jun 14) of last year. I've added the 21 Jun 13 even though that's ten days forward --  Landsat-8 entered into its 3rd year of availability a while back.

Based on fixed rocks in the lower left and the peninsula separating north and south branches, the three images align correctly -- but have a lot of changes elsewhere.

The 2014 calving front is the one showing greatest retreat for the date. This might be because the ice stream is advancing faster this year relative to rate of calving in the 'new abnormal' of stationary cold weather.

LC80 10011 2015 162 LGN00 most recent
LC80 10011 2015 146 LGN00
LC80 10011 2015 114 LGN00
LC80 10011 2015 050 LGN00

LC80 10011 2014 303 LGN00
LC80 10011 2014 271 LGN00 max retreat
LC80 10011 2014 271 LGN00
LC80 10011 2014 239 LGN00
LC80 10011 2014 223 LGN00
LC80 10011 2014 207 LGN00
LC80 10011 2014 191 LGN00
LC80 10011 2014 159 LGN00 closest to most recent 2015

LC80 10011 2013 172 LGN00 21-JUN-13 clear
LC80 10011 2013 156 LGN00 05-JUN-13 cloudy
LC80 10011 2013 140 LGN01 20-MAY-13 clear
LC80 10011 2013 124 LGN01 04-MAY-13 cloudy
LC80 10011 2013 108 LGN01 18-APR-13 calving front clear
« Last Edit: June 12, 2015, 10:51:18 PM by A-Team »

oren

  • ASIF Upper Class
  • Posts: 1590
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #667 on: June 12, 2015, 03:49:34 PM »
So JH actually managed to gain ground in the past 12 months.

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #668 on: June 12, 2015, 06:44:33 PM »
So JH actually managed to gain ground in the past 12 months.

There is quite a bit of variability within a given season, within a given year and from year to year. Although the latter shows a strong multi-decadal trend of calving front retreat, that does not mean that every year sets a new record.

The position of the calving front is a tug-of-war between competing processes -- glacial advance and propensity to calve -- so is not necessarily a good proxy for contribution to sea level rise (volume of grounded ice plus intra-glacial meltwater discharged).

Below is a comparison of the 4th and 11th of June 2015. As an RGB composite, color interference banding shows up as 2-3 areas of rapid motion (bottom center).

Here matching path,row images were not available but no systemic geometric distortion occurred. The 9,11 path,row has the calving front area almost directly below the satellite and would be most favorable for precise velocity measurements. However timing intervals can be problematic if the next orbit encounters clouds. Severely crevassed areas are not favorable either because mere widening can be mistaken for movement.

It would not be easy to determine whether/how much the ice fall on the north shore of the south bank has accelerated. However as the calving front retreats, that does remove buttressing holding back peripheral areas so faster movement there is the expectation. There are limits to any speed-up however, depending on the gravitational driving forces, basal resistance, temperature-dependent internal deformability and so forth.
« Last Edit: June 12, 2015, 10:55:55 PM by A-Team »

nukefix

  • ASIF Citizen
  • Posts: 407
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #669 on: June 16, 2015, 05:59:12 PM »
Here's a full-res zoom to S-1 IW of the southern trunk of Jakobshaven, from 8.6.2015 (ellipsoid geocoded to Polar Stereographic, 10m pixel-size, ~20m resolution).

nukefix

  • ASIF Citizen
  • Posts: 407
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #670 on: June 16, 2015, 06:24:05 PM »
Here's the required operations in the S-1 Toolbox to do this:

1. Open the S-1 image (GRD IW-product downloaded from scihub for example)
2. Click on intensity-band to visualize (converting from linear to dB beforehand (right-click on band and select Linear to/from dB) is advisable)
3. Zoom on the area of interest for subsetting and right-click and select Spatial subset from view
4. Run the attached graph to produce calibrated ellipsoid-geocoded to 10m pixel-size (bicubic) Polar Stereographic projected image that is converted to dB (log-intensity)
5. Click on Sigma-zero in dB band in the generated product to visualize
6. Save the generated product (for example by right-clicking on product and selecting Save Product as
7. Play with the histogram under "colour-management", 100% stretch works quite well already

How to use the attached graph-file:

1. Save the graph-file converting the .txt - ending to .xml (for example in C:\temp)
2. In S1TBX, open graph builder by selecting Graphs/Graph Builder
3. In Graph Builder, click Load and load the graph-file saved in step 1
4. Select your input product (saved in step 6 in the list above) in the tabs in Graph Builder
5. Select the output product name/path as appropriate
6. Run the graph by clicking Process in the Graph Builder
7. Click on Sigma-zero in dB band in the generated product to visualize
8. Play with histogram as in the list above
« Last Edit: June 16, 2015, 06:47:58 PM by nukefix »

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #671 on: June 17, 2015, 02:21:12 PM »
Very helpful. The S-1 toolbox is definitely the way to go. Without the toolkit, I was still able to download and open one of these GDH images for Petermann without it but the outcome was sub-optimal.

http://forum.arctic-sea-ice.net/index.php/topic,53.msg54273.html#msg54273

Do you have a step-by-step protocol for producing interferometric images? I was under the impression that Sentinel-1B (2016 launch) was necessary for this but they somehow produced the Napa earthquake subsidence fringes just from 1-A. (This is not another Radarsat emulation of Sentinel).

nukefix

  • ASIF Citizen
  • Posts: 407
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #672 on: June 17, 2015, 06:25:43 PM »
Do you have a step-by-step protocol for producing interferometric images? I was under the impression that Sentinel-1B (2016 launch) was necessary for this but they somehow produced the Napa earthquake subsidence fringes just from 1-A. (This is not another Radarsat emulation of Sentinel).
For InSAR just a repeated orbit is needed, with S-1A that happens every 12 days and with the 2nd unit there's a chance for InSAR every 6 days (provided that the data is acquired).

S-1 Interferometry with the IW-mode is somewhat involved due to the novel SAR mode but the toolbox is able to do it, tutorials cover the required steps:

http://step.esa.int/main/doc/tutorials/sentinel-1-toolbox-tutorials/

A hindrance with the IW-mode is that movement of the ice (in the along-track direction) generates phase-jumps that make interpretation of the interferogram more difficult. In the worst case the interferogram needs to treated as a mosaic of small (20km * 80km) interferograms.

edit: however InSAR is not needed for velocity mapping, simple feature- or speckle-tracking works well for 12-day repeats already and 6 days promises better coverage of the more difficult areas (snowfall) 
« Last Edit: June 17, 2015, 06:40:03 PM by nukefix »

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #673 on: June 18, 2015, 03:21:14 PM »
movement of the ice (in the along-track direction)

So the case of Jakobshavn which moves east to west (and slightly down) and a satellite moving almost north to south is quite unfavorable (for error term) because there is almost no component of motion in the along-track direction? Whereas Petermann would be more favorable as it moves south to north and NEGIS is intermediate as it moves diagonally though Zachariae again is west-east.

The purpose of Sentinel 1-B for Greenland would then not be velocity but rather simultaneous 'stereo pairs' that allow construction of precise digital elevation maps (and their change over time)?

nukefix

  • ASIF Citizen
  • Posts: 407
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #674 on: June 18, 2015, 03:35:28 PM »
movement of the ice (in the along-track direction)
So the case of Jakobshavn which moves east to west (and slightly down) and a satellite moving almost north to south is quite unfavorable (for error term) because there is almost no component of motion in the along-track direction? Whereas Petermann would be more favorable as it moves south to north and NEGIS is intermediate as it moves diagonally though Zachariae again is west-east.
Strictly speaking the phase-jump is not an error-term but real signal related to along-track motion (it is caused by the difference in look-angle between individual radar "bursts"). Let's see in the coming years what the InSAR community can wring out of this data.

The purpose of Sentinel 1-B for Greenland would then not be velocity but rather simultaneous 'stereo pairs' that allow construction of precise digital elevation maps (and their change over time)?
Radargrammery (stereo pairs) will not work for many reasons including resolution and the repeat-period. However, S-1A&B will be great for ice velocity measurements using coherent and incoherent feature-tracking so there's no need to do interferometry as such. InSAR can be used for more precise measurements (in the across-track direction) in the future provided that the experts figure out how to separate phase caused by along-track ice movement from phase coming from the DEM.

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #675 on: June 18, 2015, 11:29:38 PM »
Let's see in the coming years what the InSAR community can wring out of this data....  in the future provided that the experts figure out how to separate phase caused by along-track ice movement from phase coming from the DEM.
Seems like they should have figured this out way back in the grant proposal for the satellite series, rather than later on the fly. Maybe. They already had the RadarSat-2 simulation of Sentinel to work with.

It troubles me that these experts cannot simply record a script as they go along processing images. Then people who are not experts could just change the pointers to the data files and run the script on ESA computers without becoming experts themselves. I'll wager five such scripts would cover 95% of routine client data usage.

For example, that's what I see NASA offering on Landsat-7. You see a preview scene you want on earthexplorer, just put it in your cart and their computer will process it to Level 2 during an idle moment and email you a link to the product in a day or so.

In other words, nobody ever requests Level 2 on 99% of these older scenes, especially the cloudy ones, so it isn't worth the CPU time to go back to process scenes that are never going to be used when they are drowning in incoming scenes of real-time interest.

How likely is it that a highly trained photo analyst will want to also become a highly trained radar instrument designer or highly trained satellite data re-processor? Not very, different skill sets require different aptitudes.

nukefix

  • ASIF Citizen
  • Posts: 407
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #676 on: June 19, 2015, 09:50:05 AM »
I think there are several plans in Europe for offering processing capability to scientists and indeed the ability to run scripts is what is offered.

Interpreting and processing SAR images should be in the skill-set of remote sensing specialists, in the end it is not that difficult even if there's a steep learning curve in the beginning. Luckily there are several online courses that explain things from basics to PhD level stuff.
« Last Edit: June 21, 2015, 11:46:03 PM by nukefix »

Rubikscube

  • ASIF Citizen
  • Posts: 247
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #677 on: June 21, 2015, 03:24:07 PM »
The 2014 calving front is the one showing greatest retreat for the date. This might be because the ice stream is advancing faster this year relative to rate of calving in the 'new abnormal' of stationary cold weather.

It seems quite logical that a glacier like JH, which behaves so differently from month to month (unlike other glaciers like Zach, which can keep retreating at a steady phase in the midst of winter), is also more likely to respond to 5-6 weeks of abnormal cold by significantly slowing down its rate of calving. Either way it seems the "new abnormal" is finally about to end, ECMWF puts most of Greenland on 10 days of continuous roasting, starting from today.

Espen

  • ASIF Governor
  • Posts: 3078
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #678 on: June 27, 2015, 10:35:03 PM »
And here is a real big one for the season:
Have a ice day!

Espen

  • ASIF Governor
  • Posts: 3078
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #679 on: June 27, 2015, 10:56:41 PM »
But still far from the record retreat of September 28 2014:
Have a ice day!

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #680 on: June 28, 2015, 02:45:43 AM »
It's been quite a while since the last cloud-free day. Odd darker regions are apparent in the central calving front -- possibly an indentation of the calving front will develop?
« Last Edit: June 28, 2015, 02:50:56 AM by A-Team »

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #681 on: July 01, 2015, 01:05:37 AM »
The image below shows, at 7.5 m resolution, net calving front recession and channel widening in the NE that took place at Jakobshavn between 28 May 15 and 30 Jun 15. (Those dates are 32 days apart so have matching Landsat path,row and can be precisely co-registered on fixed rocks.)

This glacier is moving right along now, with quite measurable pixel shifts even over the last two days. However geometric distortion and lack of fixed fiducial points to the east mean too much relative error. I'll look for a pair of clear days at a 16 day interval....
« Last Edit: July 01, 2015, 01:10:52 AM by A-Team »

Espen

  • ASIF Governor
  • Posts: 3078
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #682 on: July 01, 2015, 06:02:57 AM »
In the latest Modis image, it looks like a big piece took off at the northern shore of the southern branch.
« Last Edit: July 01, 2015, 06:09:55 AM by Espen »
Have a ice day!

Shared Humanity

  • ASIF Upper Class
  • Posts: 2315
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #683 on: July 01, 2015, 10:53:29 PM »
Grounded below sea level as it is, I still think the northern face of the southern branch is going to surprise us over the next couple of years.

solartim27

  • ASIF Citizen
  • Posts: 438
    • View Profile
FNORD

Espen

  • ASIF Governor
  • Posts: 3078
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #685 on: July 05, 2015, 08:41:05 PM »
Pretty impressive calving seen between July 2 and July 4 2015 at Jakobshavn:
Have a ice day!

Espen

  • ASIF Governor
  • Posts: 3078
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #686 on: July 06, 2015, 10:04:23 PM »
Here we go again, another calving to note, (and very near the maximum retreat):

« Last Edit: July 06, 2015, 10:25:31 PM by Espen »
Have a ice day!

Espen

  • ASIF Governor
  • Posts: 3078
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #687 on: July 06, 2015, 10:56:47 PM »
Just an update:
« Last Edit: July 06, 2015, 11:05:14 PM by Espen »
Have a ice day!

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #688 on: July 07, 2015, 12:23:41 AM »
Amazing for this date. In the broader picture, look at all those exposed ribs in the NE corner. Below, a different take on color than EarthExplorer provides in its preview imagery (30 m 731) using adaptively contrast-filtered bands 531 instead on the LC80090112015187LGN00 download package.

Note JI is putting out so much debris that the North Branch material can hardly exit. Indeed the whole fjord is more or less plugged with melange all the way out to the grounding moraine at the mouth ... seems unusual for this late date. However it does not appear solid enough to be furnishing much buttressing to the calving front.

Arrows in the 2nd image, the same Landsat but band 8 bumped to 7.5 m, show possible weaknesses that could go in the next rounds of calving. Note the two troughs bounding an unstable peninsula in the upper left (double arrows).
« Last Edit: July 07, 2015, 01:46:42 AM by A-Team »

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #689 on: July 09, 2015, 01:52:21 AM »
Here is the DMI preview resolution for today's Sentinel -- it does a nice job of capturing the massive plume out in Disco Bay. And it sure looks like major action at the calving front (over-enlarged in 3rd image) -- there's nothing for 8 July 15 on Landsat yet and may never be.

DMI seems to get an early look at Sentinel imagery. While this is a fantastic service, they could improve it by providing image accession numbers -- when I downloaded the latest image of the same date at ESA itself, they don't seem to match the dramatic 3rd image below. I combined horizontal hh and mixed hv polarizations into RGB color using ImageJ2 ... it does not work wonders but at least has the higher native resolution and some feature emphasis.

The key feature in the 3rd image is the dark wedge projecting SE off the dark line of the calving front. The north branch also has an interesting white streak coming down the central crevassed hillside. Hopefully the clear weather will hold until Landsat swings its lazy ass by on July 13th (a path 10, row 11).
« Last Edit: July 09, 2015, 04:05:12 AM by A-Team »

Espen

  • ASIF Governor
  • Posts: 3078
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #690 on: July 09, 2015, 06:11:53 AM »
I am not sure but I do believe we are beyond the September 27 2014 maximum retreat point:
Have a ice day!

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #691 on: July 09, 2015, 08:18:39 PM »
Sentinel suspense builds ... I have not been paying attention to its orbital pattern nor to frequency of the useful GRDH saves. Here is Terra from today, terribly over-enlarged to be sure but presenting curious features, presumably associated with new calving. If an ice stream record falls in the glacier but nobody was there to see it, did it make any noise?

Today on June 9th, two new Sentinel images came in. DMI and Polar View offer convenient versions of these, perhaps at the price of resolution loss (third image) as I wait on the Sentinel hub downloads. The IW scene is 3 GB and may not be indexed by the quick look services. Oddly it is dated 16 Feb 15 despite being given the top listing.

S1A_EW_GRDM_1SDH_20150709T203809_20150709T203909_006739_009079_DD28
Date : 2015-07-09T20:38:09.717Z, Instrument : SAR-C, Mode : EW, Satellite : Sentinel-1, Size : 409 MB

S1A_IW_SLC__1SSH_20150216T100008_20150216T100035_004647_005BB8_E22B
Date : 2015-02-16T10:00:08.133Z, Instrument : SAR-C, Mode : IW, Satellite : Sentinel-1, Size : 3 GB
« Last Edit: July 10, 2015, 04:17:36 PM by A-Team »

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #692 on: July 10, 2015, 05:45:26 PM »
Seems like there are some advantages from using files from the Sentinel hub itself, notably the ability to manipulate within 16-bits (enlargement, rotation and local contrast), and separate HH and HV polarizations which can also be combined into RGB (blue taken as all 0 or black).

The great downside is that north is not up in Sentinel images nor is a rotational angle provided nor is this angle the same between images. This contrasts with DMI and Landsat. For example the 10 June 15 Sentinel needs an 18.9º rotation.

The HV is brighter in this instance. It seems to better distinguish calved from recently calved from freshly calved. I don't know what the physical basis for this is, maybe calved ice becomes less rough over time. Neither polarization combo is proposing about-to-calf.

The Jakobshavn calving front seems to have moved further into record retreat; that can be measured relative to the last fixed point, the rock island in lower left. If this is July 9th and maximal retreat in 2014 occurred in late Sept, this could get interesting. There are three issues: position relative to basal topography sills and troughs, loss of side buttressing if the calving front should retreat past the 'big bend', and greatly reduced obstruction to side tributary inflow.

The comparison to the Sept 27th maximal retreat of last year is somewhat problematic due to geometric effects in two different instruments, plus there has been more melt-out around our favorite reference rock late season adding to the alignment/rescaling uncertainty. Still, after moving Espen's calving front line over to this year's image shows there has been a modest level record retreat.

The June 10th has come in ... this one is rotated CW to better match true north and DMI. This has the effect of rotating the calving line around the reference rock.

S1A_EW_GRDM_1SDH_20150710T095917_20150710T100017_006747_0090B2_65F1
« Last Edit: July 10, 2015, 11:57:30 PM by A-Team »

Anne

  • ASIF Middle Class
  • Posts: 522
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #693 on: July 10, 2015, 10:58:04 PM »
Any thoughts on this video of calving from 20 June? Colorado Bob on Robert Scribbler suggests it's not a normal calving but forced by the sweet bluewater from behind.

Colorado Bob usually seems to know what he's talking about.

Whatever, it's pretty impressive.

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #694 on: July 10, 2015, 11:37:25 PM »
Anne, could you please link to the specific page where Robert Scribler quotes Colorado Bob with these words? 

Anne

  • ASIF Middle Class
  • Posts: 522
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #695 on: July 11, 2015, 02:00:48 AM »
Hi, A Team - sorry, my bad. It wasn't a comment of R Scribbler's so he didn't specifically endorse it, but a comment on one of his endless comment streams. I'm trying to find it again. But in any case, I was interested in how an informed person (not me) would interpret the video.

Anne

  • ASIF Middle Class
  • Posts: 522
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #696 on: July 11, 2015, 02:19:06 AM »
OK - just shows how false memory can be. The original post in the RS comments wasn't by CB but by Andy in San Diego, and it was CB's comments that struck me:
https://robertscribbler.wordpress.com/2015/07/09/unprecedented-fire-season-has-burned-11-million-acres-so-far-for-alaska-and-canada/#comment-44188
Apologies if I should have posted this in Stupid Questions instead.

Shared Humanity

  • ASIF Upper Class
  • Posts: 2315
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #697 on: July 11, 2015, 02:44:25 AM »
Seems like there are some advantages from using files from the Sentinel hub itself, notably the ability to manipulate within 16-bits (enlargement, rotation and local contrast), and separate HH and HV polarizations which can also be combined into RGB (blue taken as all 0 or black).

The great downside is that north is not up in Sentinel images nor is a rotational angle provided nor is this angle the same between images. This contrasts with DMI and Landsat. For example the 10 June 15 Sentinel needs an 18.9º rotation.

The HV is brighter in this instance. It seems to better distinguish calved from recently calved from freshly calved. I don't know what the physical basis for this is, maybe calved ice becomes less rough over time. Neither polarization combo is proposing about-to-calf.

The Jakobshavn calving front seems to have moved further into record retreat; that can be measured relative to the last fixed point, the rock island in lower left. If this is July 9th and maximal retreat in 2014 occurred in late Sept, this could get interesting. There are three issues: position relative to basal topography sills and troughs, loss of side buttressing if the calving front should retreat past the 'big bend', and greatly reduced obstruction to side tributary inflow.

The comparison to the Sept 27th maximal retreat of last year is somewhat problematic due to geometric effects in two different instruments, plus there has been more melt-out around our favorite reference rock late season adding to the alignment/rescaling uncertainty. Still, after moving Espen's calving front line over to this year's image shows there has been a modest level record retreat.

The June 10th has come in ... this one is rotated CW to better match true north and DMI. This has the effect of rotating the calving line around the reference rock.

S1A_EW_GRDM_1SDH_20150710T095917_20150710T100017_006747_0090B2_65F1

if you watch the video of the calving, particularly the last 4 minutes, I think you are seeing the  formation of that curious feature on the north bank of the main channel. During the video,the camera person shifts away from the massive ice berg and focuses in on what appears to be the north bank sliding into the fjord. I have felt for sometime that the north bank, which is grounded below sea level is seeing significant bottom melt as water from the fjord undercuts it. That initial calving off the main calving front was massive. I believe the force of the water that was displaced wreaked havoc on a portion of the north bank that has already lifted (no longer grounded but more of an ice shelf.)

Of course, I don't know what I am talking about.

A-Team

  • ASIF Upper Class
  • Posts: 1915
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #698 on: July 11, 2015, 07:57:20 AM »
It's a remarkable video of a major drawn-out calving event. It's worth noting though that at a 40 m per day advance, 1500 m thickness and 5 km width, this glacier is calving a cubic km of ice every 3 days or so.

After some cliff fall, the really large piece broke off to bedrock ~1400 meters below the sea. It rolled back and came out feet first which is a typical but non-intuitive calving mode. (In Chasing Ice, the berg tipped forward). In both clips, we see very dark blue, almost black ice surfacing, some of it perhaps with embedded till.

None of the blue marked in the image below is liquid water -- all of it is immobile ice. You can see a smallish volume of whitish-gray water surging over a divide on the giant calved piece at 5:37. The giant berg is quite a distance from the calving front at this point.

It has not proven possible to get instrumentation on an autonomous submersible or a mooring anywhere near the bedrock/ice interface to measure meltwater discharge. It would be quite dangerous to even bucket-sample surface water from a helicopter in the midst of an event like this.

The lowered water pressure suddenly created by the feet-first rotation to surface and its inrushing replacement have been treated in various modeling papers (and the recent paywalled seismic study), as have relative buoyancy properties of emerging basal fresh meltwater. Most glaciers in Greenland have distributed continuous discharge early in season that matures to more efficient tunnel collection systems later.

As Shared H points out, there is a lot worth looking at in this video depending on your interests. Something massive beyond human experience moving so rapidly will indeed cause all manner of unfamiliar havoc.

Jason Box's drone study was up at Store Glacier. I recall this looked at the role of meltwater in hydrofracturing crevasses and so inducing calving. JC Ryan et al UAV photogrammetry and structure from motion to assess calving dynamics at Store Glacier, a large outlet draining the Greenland ice sheet, The Cryosphere, 9, 1-11, doi:10.5194/tc-9-1-2015, 2015. All of Box's papers, including 3 new ones in review, can be found at https://sites.google.com/site/jboxgreenland/publications.

J Todd and P Christoffersen studied Store in 2014, remarking:

Water pressure is essential for basal crevasse penetration, but it may also be significant in surface crevasses. The process of hydrofracturing by water in surface crevasses is believed to have been a critical factor in the collapse of the Larsen B Ice Shelf. However, while water in surface crevasses may be important, it is extremely difficult to quantify. The relationship between surface melt rate and crevasse water depth depends on the distribution, shape and depth of crevasses, and melting and refreezing on crevasse walls, as well as potential drainage of water from crevasses into englacial, subglacial or proglacial water bodies. As such, it is currently impossible to estimate even an order of magnitude for crevasse water depth at Store in summer. However, outside the 3-month summer melt sea-
son, surface crevasses must be assumed to be dry. http://www.the-cryosphere.net/8/2353/2014/tc-8-2353-2014.pdf
« Last Edit: July 11, 2015, 08:06:51 AM by A-Team »

Anne

  • ASIF Middle Class
  • Posts: 522
    • View Profile
Re: Jakobshavn Isbræ / Sermeq Kujalleq / Ilulissat Icefjord
« Reply #699 on: July 11, 2015, 11:31:11 AM »
Thank you, A-Team.  :)