Support the Arctic Sea Ice Forum and Blog

Author Topic: Home brew AMSR2 extent & area calculation  (Read 1559458 times)

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2250 on: July 05, 2016, 04:44:40 PM »
The surface melt extent dropped a bit. Most of on the Canadian side of the CAB.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2251 on: July 06, 2016, 06:37:31 PM »
Drops are not big enough to prevent that 2012 has taken the lead in extent. Still leading in area but the margins are very small.

Update 20160705.

Extent: -111.8 (-696k vs 2015, -228k vs 2014, -342k vs 2013, +17k vs 2012)
Area: -131.2 (-581k vs 2015, -445k vs 2014, -469k vs 2013, -16k vs 2012)
 
You will find the updated graphs in the top post

Extent declines were biggest in Baffin (-29k), Hudson (-27k) and Kara (-24k).

Area declined most in Hudson (-37k), CAB (-34k) and Kara (-18k).

The regional delta map is of the Kara and Barents regions. The ice edge can be seen withdrawing all along.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2252 on: July 06, 2016, 06:49:53 PM »
Animation of the Greenland Sea. A cloud of floes that seemingly move in random directions.

A-Team

  • Young ice
  • Posts: 2697
    • View Profile
  • Liked: 637
  • Likes Given: 33
Re: Home brew AMSR2 extent & area calculation
« Reply #2253 on: July 06, 2016, 06:52:34 PM »
AMSR2 at PolarView is showing quite a bit of action around the North Pole itself. There's not quite cloud-penetrating radar coverage ('film strips', first two images) of the immediate region by Sentinel 1A because of its orbit/swath (outer circle envelope at ~256 km) nor in the AMSR2 product (inner grayed-out circle ~95 km) but the region shows up on Modis visible though with some distortion at the pole (2x image, bottom).

The composite image is made by layering the non-cloudy July images with the 4th on top and cutting out its cloudy regions to reveal clear areas (if any) on earlier images. The last 30 days and more have been very cloudy.

The palette is buried at PolarView so a familiar bit of the Beaufort that utilizes the same color range as at the pole is provided as an inset (2nd image, upper left). The distance scale is also a little ambiguous on both WorldView and PolarView as to whether the end bars are part of the indicated distance or not.

The distances here are only approximate for other reasons as well. One degree of latitude anywhere on the globe is commonly said to be 111.3 km but that's not the case for WGS84 ellipsoid where it is 111.694 at the pole. (The datum and projection are stated for WV but not evident at PV). There's a nice discussion at  https://en.wikipedia.org/wiki/Latitude#Latitude_on_the_ellipsoid
« Last Edit: July 07, 2016, 03:20:44 AM by A-Team »

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2254 on: July 06, 2016, 06:56:38 PM »
and the surface melt extent, according to the ADS/Jaxa data.

seaicesailor

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2255 on: July 06, 2016, 07:05:56 PM »
Animation of the Greenland Sea. A cloud of floes that seemingly move in random directions.

The strength of the Spitsbergen current (western island of Svalbard) is generating a lot of turbulence when shearing against the Greenland current. This is very visible now that this last is so weakened, and the sparse ice acts as a marker!

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2256 on: July 07, 2016, 12:33:01 PM »
Is it going to be exciting after all? Double century area drop makes the lead to 2012 a bit bigger.

Update 20160706.

Extent: -92.6 (-700k vs 2015, -215k vs 2014, -401k vs 2013, +36k vs 2012)
Area: -220.4 (-667k vs 2015, -505k vs 2014, -641k vs 2013, -129k vs 2012)
 
You will find the updated graphs in the top post

The area drop is mostly caused by the big drop in the CAB (-87k), Chukchi is second (-38k) and Hudson continues its nose dive with a drop of -34k.

Extent declines are more evenly spread, Chukchi has the biggest drop: -18k.

Regional maps are the Arctic Basin, a big patch of concentration drop is visible in the CAB. At the same place nullschool shows rain following the fronts of a polar low.
 

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2257 on: July 07, 2016, 12:44:40 PM »
Animation is from ESS. In contrast with nearby Laptev, the ESS seems to have started on what looks like an early melt.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2258 on: July 07, 2016, 12:50:09 PM »
The surface meting extent (from ADS/Jaxa data) has increased a bit. I guess rain will affect this measure as melting does.

A-Team

  • Young ice
  • Posts: 2697
    • View Profile
  • Liked: 637
  • Likes Given: 33
Re: Home brew AMSR2 extent & area calculation
« Reply #2259 on: July 08, 2016, 04:51:27 PM »
Somebody somewhere is doing something with the Sentinel 1A images of the Arctic Ocean. Presumably its vastly better resolution and cloud-penetrating ability can be assimilated into improved sea ice concentration products such as AMSR2 3.125 km.

A lot of S1A imagery is captured each 24 hours (first image) though only half the Arctic is covered per week, in contrast to nearly complete daily coverage by AMSR2 on the GCOM-W satellite. PolarView (http://www.polarview.aq/arctic) provides 'Extra-Wide Swath Mode 400 km wide, 25 x 100 m spatial resolution' such as the S1A_EW_GRDM_1SDH_20160707 below.

It is quite feasible to track individual edge floes on the AMSR2 product such as Big Block in the Beaufort or the one off Svalbard (red arrow, 2nd image) and identify their S1A counterparts, which then can be analyzed at their much higher resolution to provide feedback to the AMSR2 processing algorithm. The gray shading on the AMSR2 image shows a region that can be compared on 06 July 16.

The 4th image is quite wide but still nowhere near the ultimate resolution S1A provides (next post). It needs a click to display properly.
« Last Edit: July 08, 2016, 06:18:45 PM by A-Team »

A-Team

  • Young ice
  • Posts: 2697
    • View Profile
  • Liked: 637
  • Likes Given: 33
Re: Home brew AMSR2 extent & area calculation
« Reply #2260 on: July 08, 2016, 05:04:09 PM »
Here is the edge of the melt zone at full S1A resolution on 07 Jul 15. The animation shows some options for measuring sea ice concentration by use of posterization, various levels of reductions in allowed grayscales (ie binning).

The 2x AMSR2 is quite a nice product. The palette is done right: a tinted grayscale ramp with distinct bars correspond to distinct colors in the map.That is illustrated in the second frame of the second animation below, in which all the pixels in the image corresponding to a certain color on the blue scale (bottom) are replaced by green.

The problem is, the human eye cannot readily distinguish all the color differences on this (or any other 256 color palette). For example, you cannot tell the difference between pixels representing 100% and 99% sea ice concentrations.

Frames 3 looks at values in the 96-100% range sea ice concentration and frame 4 colors them maximally non-harmoniously (the Glasbey palette). Frame 5 does this for the whole image. It's clear from this that the AMSR2 3.125 product is not binning and subtle differences are represented. Whether they correspond to what S1A sees is another matter.

The final animation blows up the AMSR2 region corresponding to the S1A image and recolors the high concentration portion of the palette, with the tapered outcome towards the edge making sense. (To retain the palette, rescaling needs to be done to an integral multiple without using any interpolation, ie not bicubic or linear etc.)
« Last Edit: July 08, 2016, 06:21:35 PM by A-Team »

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2261 on: July 08, 2016, 06:19:54 PM »
Moderate declines, but margins stay approximately the same.

Update 20160707.

Extent: -67.3 (-649k vs 2015, -205k vs 2014, -372k vs 2013, +29k vs 2012)
Area: -85.9 (-581k vs 2015, -477k vs 2014, -638k vs 2013, -162k vs 2012)
 
You will find the updated graphs in the top post

Regional extent dropped most in the Hudson region : -19k.

Regional area dropped most in the ESS and Hudson (-32k and -31k).

Regional delta map of the day is Hudson. It is going fast now, with ice in the Bay remaining on the south-west shores. 

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2262 on: July 08, 2016, 06:24:58 PM »
Animation of today is the Nares Strait compared with last year. Both are on the brink of flushing, but not yet. We know that last year it broke on the 12th of July.

Click on the picture to animate.

BTW, no surface melt report today. I will save a couple of day more data for the next update, unless of course something remarkable happens.

iceman

  • Frazil ice
  • Posts: 285
    • View Profile
  • Liked: 7
  • Likes Given: 19
Re: Home brew AMSR2 extent & area calculation
« Reply #2263 on: July 09, 2016, 12:35:23 PM »
   ....
Somebody somewhere is doing something with the Sentinel 1A images of the Arctic Ocean. Presumably its vastly better resolution and cloud-penetrating ability can be assimilated into improved sea ice concentration products such as AMSR2 3.125 km.
   ....

Nice work, thanks for showing us the capabilities of this source.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2264 on: July 09, 2016, 06:31:10 PM »
No change in the rankings. After today 2012 starts a little nose dive, dropping 520k extent in three days. Declines like today will be no match for that.

Update 20160708.

Extent: -83.8 (-647k vs 2015, -185k vs 2014, -312k vs 2013, +33k vs 2012)
Area: -62.8 (-511k vs 2015, -445k vs 2014, -555k vs 2013, -135k vs 2012)
 
You will find the updated graphs in the top post

ESS and Chukchi declined most in extent: -27k and -26k.

In the regional area arena, ESS dropped -32k, Hudson -26k and Laptev -16k. Greenland Sea increased by +28k.

Regional delta map is from ESS and Laptev. ESS is quite active and the ice cover is dropping earlier than in other years (in my limited data set).

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2265 on: July 09, 2016, 06:38:12 PM »
Animation of Kara where the sea is getting emptier more and more.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2266 on: July 10, 2016, 10:27:21 AM »
The decline speed-up did not come, so the area is now only a hair below 2012 (and extent is getting further above).

Update 20160709.

Extent: -82.7 (-658k vs 2015, -141k vs 2014, -267k vs 2013, +193k vs 2012)
Area: -61.0 (-511k vs 2015, -385k vs 2014, -489k vs 2013, -20k vs 2012)
 
You will find the updated graphs in the top post

Most of the extent decline is from Hudson (-33k). Chukchi and CAA are second (both -16k).

Area declined fastest in CAA (-24k), Geenland Sea (-17k) and Hudson (-17k) are second. The CAB increased by +25k.

Regional delta map is Chukchi, the decline of the ice pack seems to be on schedule.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2267 on: July 10, 2016, 10:29:20 AM »
Animation is of the Hudson. Somebody seems to have put the light out over there.

A-Team

  • Young ice
  • Posts: 2697
    • View Profile
  • Liked: 637
  • Likes Given: 33
Re: Home brew AMSR2 extent & area calculation
« Reply #2268 on: July 10, 2016, 05:53:37 PM »
Quote
thanks for showing capabilities of S1A resource.
Somebody somewhere is writing a paper doing this more fully. Below is a case where the PolarView and AMSR2 3.125 products can be vetted against multiple visible scenes and a S1A active radar tile of the same date, July 9th. The animation shows successive zooms of this region for this much higher resolution imagery.

It appears that PolarView has grossly misread sea ice concentration in the small area checked. The ice pack has some old fractures but appears entirely solid despite some darker areas. The AMSR2 3.1, visible channels and S1A are in agreement. Of these, S1A is the least affected by intervening atmospheric conditions.

The 2nd image shows by upper color picking and replacement (split palette region) that AMSR2 3.1 actually is showing 100% ice pack in this region. Some of the fringing regions are slightly lower. It is only when a map product is correctly made (ie not some muddy compressed jpg at below-data resolution) that this technique can be used by product end users.

The larger area between the pole and CAA is quite problematic for both PolarView and AMSR2 3.125 as far as Aqua/Terra/Suomo are concerned. Here clouds move between orbital passes but ice does not as the time interval is too short (as can be quickly animated by clicking layers on and off in EarthExplorer), so clouds can be distinguished from ice by this method (or 367 infrared). However all sources are consistent with open floes to the 'northeast' of the north pole.

Looking through Beitsch 2014 paper, it appears their new 3.1k algorithm was vetted against clear sky Modis 250 km visible but not against Sentinel 1A (too new), RadarSat (coverage dates and access issues), Sentinel 2AB (too new), or Landsat-8 (too limited to coastal areas).

http://www.mdpi.com/2072-4292/6/5/3841/htm Beitsch  2014 free full text
http://neven1.typepad.com/blog/2013/06/new-map-on-the-block.html nice overview by Neven
http://www.polarview.aq/arctic
ftp://ftp-projects.zmaw.de/seaice/AMSR2/3.125km/
http://tinyurl.com/zvwwfh2 Modis Aqua
« Last Edit: July 10, 2016, 07:19:20 PM by A-Team »

A-Team

  • Young ice
  • Posts: 2697
    • View Profile
  • Liked: 637
  • Likes Given: 33
Re: Home brew AMSR2 extent & area calculation
« Reply #2269 on: July 10, 2016, 07:09:41 PM »
Quote
Somebody somewhere is writing a paper more fully assimilating SAR and AMSR2 3.1 ...
... if such exist, the way to locate them is search by full title of Beitsch 2014 for articles citing it. That turns up 3 articles using Radarsat2 rather than S1A:

Polynya impacts on water properties in a Northeast Greenland fjord
  Polynyas are an important yet poorly understood phenomena of high latitude
oceans. A storm event during late December 2013 with down-fjord winds of up to 25 m/s
forced the collapse of the landfast ice over the Young Sound (YS) fjord outlet in northeast ...

Lead detection in Arctic sea ice from CryoSat-2: quality assessment, lead area fraction and width distribution
  Leads cover only a small fraction of the Arctic sea ice but they have a dominant
effect on the turbulent exchange between the ocean and the atmosphere. A supervised
classification of CryoSat-2 measurements is performed by a comparison with visual ...

Improving Arctic sea ice edge forecasts by assimilating high horizontal resolution sea ice concentration data into the US Navy's ice forecast systems
  This study presents the improvement in ice edge error within the US Navy's
operational sea ice forecast systems gained by assimilating high horizontal resolution
satellite-derived ice concentration products. Since the late 1980's, the ice forecast systems ...

Loitering of the retreating sea ice edge in the Arctic Seas
  Each year, the arctic sea ice edge retreats from its winter maximum extent through
the Seasonal Ice Zone (SIZ) to its summer minimum extent. On some days, this retreat
happens at a rapid pace, while on other days, parts of the pan-arctic ice edge hardly move ...

Shelfbreak current over the Canadian Beaufort Sea continental slope: Wind‐driven events in January 2005
  The shelfbreak current over the Beaufort Sea continental slope is known to be one
of the most energetic features of the Beaufort Sea hydrography. In January 2005, three
oceanographic moorings deployed over the Canadian (eastern) Beaufort Sea continental ...

Long-term coastal-polynya dynamics in the southern Weddell Sea from MODIS thermal-infrared imagery
  Based upon thermal-infrared satellite imagery in combination with ERA-Interim
atmospheric reanalysis data, we derive long-term polynya characteristics such as polynya
area, thin-ice thickness distribution, and ice-production rates for a 13-year investigation ...

SMOS sea ice product: Operational application and validation in the Barents Sea marginal ice zone
  Brightness temperatures at 1.4 GHz (L-band) measured by the Soil Moisture and
Ocean Salinity (SMOS) Mission have been used to derive the thickness of sea ice. The
retrieval method is applicable only for relatively thin ice and not during the melting period. ...

Prediction of Arctic Sea Ice for Ship Routing: Forecast Experiment and Ship Cruise
  Results are presented from a sea ice forecasting experiment combined with a field
campaign undertaken in the Barents Sea in March 2014. The simulations were performed
with the regional coupled atmosphere-sea ice-ocean model HAMMER, recently ...

Error assessment of satellite-derived lead fraction in the Arctic
  Leads within consolidated sea ice control heat exchange between the ocean and
the atmosphere during winter, thus constituting an important climate parameter. These
narrow elongated features occur when sea ice is fracturing under the action of wind and ...

Assessment of error in satellite derived lead fraction in Arctic
  Response to the comments of reviewer #2 to ... Assessment of error in satellite derived lead
fraction in Arctic by N. Ivanova, P. Rampal, and S. Bouillon ... In the following we provide detailed
answers (marked by “A”) to your comments (“R”) and describe what changes are done in ...

Assimilation of ice and water observations from SAR imagery to improve estimates of sea ice concentration
KA Scott et al (not yet cited itself)
http://www.tellusa.net/index.php/tellusa/article/view/27218 free full text
  In this paper, the assimilation of binary observations calculated from synthetic
aperture radar (SAR) images of sea ice is investigated. Ice and water observations are
obtained from a set of [wide-swath 50m Radarsat2] SAR images by thresholding ice and water probabilities calculated ...


[[Oddly this cites a different Beitsch 2014, namely "Beitsch, A., Kaleschke, L. and Kern, S. 2014. Investigating high-resolution AMSR2 sea ice concentrations during the February 2012 fracture event in the Beaufort Sea. Remote Sens. 6(5), 3841–3856. free full text at http://tinyurl.com/hbscow7

Improved Sea Ice Concentration Estimation Through Fusing Classified SAR Imagery and AMSR-E Data
L Wang ... KA Scott
  A method to automatically combine binary ice/water information from synthetic
aperture radar (SAR) sea ice images with the Advanced Microwave Scanning Radiometer-
EOS (AMSR-E) daily ice concentration product is proposed for the purpose of generating ...

Sea Ice Concentration Estimation During Melt From Dual-Pol SAR Scenes Using Deep Convolutional Neural Networks
L Wang  ... KA Scott
High-resolution ice concentration maps are of great interest for ship navigation and ice hazard forecasting. In this case study, a convolutional neural network (CNN) has been used to estimate ice concentration using synthetic aperture radar (SAR) scenes captured during the melt season. These dual-polarization RADARSAT-2 satellite images are used as input, and the ice concentration is the direct output from the CNN. With no feature extraction or segmentation ...


Atmospheric Correction of Brightness Temperatures for Sea Ice Concentration Retrieval using 89 GHz Algorithms
  Sea ice, as a major component of cryosphere, has a significant impact on the heat transfer
and salt fluxes in polar regions and hence drives global climate changes. 5%-8% of the
ocean is covered by sea ice (Comiso et al., 2003). Because of its high albedo, sea ice ...

Sea and Freshwater Ice Concentration from VIIRS on Suomi NPP and the Future JPSS Satellites
  Information on ice is important for shipping, weather forecasting, and climate monitoring.
Historically, ice cover has been detected and ice concentration has been measured using
relatively low-resolution space-based passive microwave data. This study presents an ...

Sea ice leads in the Arctic Ocean: Model assessment, interannual variability and trends
  Sea ice leads in the Arctic are important features that give rise to strong localized
atmospheric heating; they provide the opportunity for vigorous biological primary production;
and predicting leads may be of relevance for Arctic shipping. It is commonly believed that ...
« Last Edit: July 10, 2016, 11:00:02 PM by A-Team »

Michael Hauber

  • Grease ice
  • Posts: 900
    • View Profile
  • Liked: 77
  • Likes Given: 14
Re: Home brew AMSR2 extent & area calculation
« Reply #2270 on: July 11, 2016, 08:02:51 AM »
Do I read the legends correctly on the first page (my colour eyesight is not the best)?  That extent according to SSMI 25k is significantly higher than all other years, and area according to SSMI 25k is significantly lower?  Is this reasonable, or is it due to the recent sensor issues impacting CT and NSIDC indices?  These figures don't seem to be comparable to other recent years which are all on much smaller resolutions, so is there any value to including these measures?
Climate change:  Prepare for the worst, hope for the best, expect the middle.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2271 on: July 11, 2016, 10:16:38 AM »
Do I read the legends correctly on the first page (my colour eyesight is not the best)?  That extent according to SSMI 25k is significantly higher than all other years, and area according to SSMI 25k is significantly lower? 

You read that correctly

Quote
Is this reasonable, or is it due to the recent sensor issues impacting CT and NSIDC indices?

Nothing to do with the sensor issues. The source data (gridded sea ice concentration) was calculated with a different algorithm, from different microwave bands from different satellites.

[/quote] These figures don't seem to be comparable to other recent years which are all on much smaller resolutions, so is there any value to including these measures?
[/quote]

The purpose of this thread is to study sea ice extent and area calculated from the Uni Hamburg AMSR2 sea ice concentration product. Comparing with results calculated in (as much as possible ) the same way from other products is valuable to me. Not to include every year and product is just to keep the graph readable.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2272 on: July 11, 2016, 02:17:26 PM »
Now in second position after 2012, both in extent and area.

Update 20160710.

Extent: -67.6 (-597k vs 2015, -100k vs 2014, -235k vs 2013, +301k vs 2012)
Area: -51.5 (-404k vs 2015, -392k vs 2014, -404k vs 2013, +59k vs 2012)
 
You will find the updated graphs in the top post

Most of the extent loss was in the Hudson, -39k.

Regional area shows a mixed picture: Laptev and CAB increased the area (+33k and +28k). ESS, Chukchi and Hudson lost area (-31k, -22k and -22k).

Since Hudson is the region where the extent is still declining fast, here is how it looks like as a delta map.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2273 on: July 11, 2016, 02:22:13 PM »
Animation of the Canadian Archipelago. The fast ice in the channels is breaking in sevarl places. Same in the Nares Strait.

A-Team

  • Young ice
  • Posts: 2697
    • View Profile
  • Liked: 637
  • Likes Given: 33
Re: Home brew AMSR2 extent & area calculation
« Reply #2274 on: July 11, 2016, 04:02:32 PM »
A very rare clear day between the North Pole and Franz Josef Land on July 10th. The AMSR2 3.125 km product agrees very well with the visual scene (which has had cloud removal processing in ImageJ). The Terra Modis image shows the state of the ice at much higher resolution than AMSR2 ... it is well worth a click.

« Last Edit: July 11, 2016, 06:00:07 PM by A-Team »

Tor Bejnar

  • Young ice
  • Posts: 3637
    • View Profile
  • Liked: 633
  • Likes Given: 407
Re: Home brew AMSR2 extent & area calculation
« Reply #2275 on: July 11, 2016, 04:35:34 PM »
I particularly like the firey edge.  I presume it shows the melting and dispersal of floes that get into the 'hot' Atlantic water.

This screen shot is from A-Team's post and includes the 60ºE meridian and a bit of Franz Josef Land.
Arctic ice is healthy for children and other living things.

A-Team

  • Young ice
  • Posts: 2697
    • View Profile
  • Liked: 637
  • Likes Given: 33
Re: Home brew AMSR2 extent & area calculation
« Reply #2276 on: July 11, 2016, 07:19:12 PM »
Quote
firey edge... presume it shows the melting and dispersal of floes that get into the 'hot' Atlantic water.
Indeed, an intriguing feature. There was a good suggestion earlier of the edge being caught in cross-currents, the Transpolar Drift meets the North Atlantic Current. The motion (1st image) seems to show a central eddy but mostly (as wipneus notes above) somewhat indecisive motion. July 09-11 were the only clear dates recently, though the AMSR2 3.125 shows floe motion reliably (see above) for longer periods.

The streamers could potentially be wind-blown foam, stringers of ices from disintegrating floes, condensed vapor plumes, or clouds formed as  warm water encountersr cold. The 367 infrared images for the three dates show they are not related to prevailing higher level clouds. Wayne might know what is going on, as might S1A.

The big picture, as people have been observing for months, is ice carried by a classic Transpolar Drift being obliterated along a broad front as it meets the anomalously poleward warm North Atlantic waters, with the magnitude of the effect seemingly pushing Fram export into near-irrelevancy.

In the last animation, viewed over and over, you can see what must be weather systems sweeping across the ice but also surprisingly rapid movement of bulk ice. This means two things: the AMSR2 3km algorithm is not always able to get down to the surface ice (ie remove all traces of atmospheric artifacts) and a rolling multi-day average won't remove cloud effects unless it is motion-compensated. Ice motion can even be seen in Aqua/Terra/Suomo even though they are passing overhead in quick succession. It helps though that weather systems often move through much more rapidly than the ice pack is seen moving.
« Last Edit: July 12, 2016, 03:38:02 AM by A-Team »

Tor Bejnar

  • Young ice
  • Posts: 3637
    • View Profile
  • Liked: 633
  • Likes Given: 407
Re: Home brew AMSR2 extent & area calculation
« Reply #2277 on: July 11, 2016, 07:29:51 PM »
Thanks, A-Team, for listing multiple possibilities. I think "wind-blown foam" beats my "dispersal of floes".  I do hope someone can say more on this!
Arctic ice is healthy for children and other living things.

A-Team

  • Young ice
  • Posts: 2697
    • View Profile
  • Liked: 637
  • Likes Given: 33
Re: Home brew AMSR2 extent & area calculation
« Reply #2278 on: July 11, 2016, 09:35:17 PM »
Quote
multiple possibilities."wind-blown foam"
I might have checked nullschool or similar for wind direction on that date though the time of day is a bit dodgy on strip image assemblages. (Good project for someone: dump the GFS and feed nullschool the more reliable ECMWF; code is open source.)

Currents ... are the old schoolbook maps still applicable in 2016 given implied new encroachments? Note 10 July 16 is not that different from the same date in 2013. (AMSR2 3km is not available for 2012 or earlier.)

S1A radar ... yes but. It's not so easy to interpret in this instance.

There might be better color in Landsat or S2A; there might be coverage since features are fairly close to land.

seaicesailor

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2279 on: July 11, 2016, 10:05:44 PM »
A very rare clear day between the North Pole and Franz Josef Land on July 10th. The AMSR2 3.125 km product agrees very well with the visual scene (which has had cloud removal processing in ImageJ). The Terra Modis image shows the state of the ice at much higher resolution than AMSR2 ... it is well worth a click.

Greatness! Thank you A-Team
Seems a lot of melting is happening not as evident as whenn it is in Greenland Sea. I was not sure if the front was just stagnant.

Andreas T

  • Nilas ice
  • Posts: 1143
    • View Profile
  • Liked: 17
  • Likes Given: 4
Re: Home brew AMSR2 extent & area calculation
« Reply #2280 on: July 11, 2016, 10:48:04 PM »
Thanks, A-Team, for listing multiple possibilities. I think "wind-blown foam" beats my "dispersal of floes".  I do hope someone can say more on this!
if the competition is for least plausible explanation "wind blown foam" beats all comers :D
these patterns are often found near the ice edge when wind is blowing away from the ice, but why should wind blown foam only occur in the vicinity of ice? And why does it not occur when the wind is blowing toward the ice?
I thought the red tint in 367 band false colour images is always ice but looking at the great barrier reef and some pacific atolls the red shows up (faintly) there as well not sure whether its surf or water over white sand. Learned something new for the "interpretation" thread.
But more likely is that disintegrating floes in the final stages are in such small pieces well below the resolution of the satellite image that they form this amorphous smear of lightened pixels. These pieces are swept into agglomerations by wind and wave action which form the distinctive streamers.
Look at the Obuoy9 "movie" for late August 2015 (lat/lon position can be found in image)http://obuoy.datatransport.org/monitor#buoy9/movie and compare with worldview images of northeast Greenland coast http://go.nasa.gov/29CKw29 the patterns are different because the wind conditions don't match, but I think similar ice fragments are what we are seeing here.
« Last Edit: July 11, 2016, 10:53:33 PM by Andreas T »

P-maker

  • Frazil ice
  • Posts: 293
    • View Profile
  • Liked: 52
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2281 on: July 11, 2016, 11:02:51 PM »
I believe the correct term for this fiery stuff is 'brash ice', which normally consists of broken and melting pieces of ice not more than two meters across.

The funny thing is that brash ice is normally the result of fully disintegrating icebergs. This year North  of Svalbard, the multi-year sea ice in the transpolar drift breaks apart instantly, when it hits the warm Atlantic waters.

Tor Bejnar

  • Young ice
  • Posts: 3637
    • View Profile
  • Liked: 633
  • Likes Given: 407
Re: Home brew AMSR2 extent & area calculation
« Reply #2282 on: July 11, 2016, 11:13:49 PM »
Here is a close up of brash ice from here:


Seems reasonable.
Arctic ice is healthy for children and other living things.

magnamentis

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2283 on: July 12, 2016, 01:38:07 AM »
Here is a close up of brash ice from
Seems reasonable.

Et Voilà :-)

Thanks for this which is exactly it  8)

Sleepy

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2284 on: July 12, 2016, 07:09:09 AM »
Good project for someone: dump the GFS and feed nullschool the more reliable ECMWF; code is open source.
That would be absolutely wonderful, but this is ECMWF you're talking about. I have an account at ECMWF, but that isn't enough to download the open source version of IFS without further access. That has happened several times in the past when I tried to access other stuff. It feels more like when I applied for a visa to russia in the 90's, ending up paying someone to sit at the embassy for days, until those behind the desk got bored and granted the visa.
Quote
The OpenIFS programme at ECMWF provides academic and research institutions with an easy-to-use version of the ECMWF IFS (Integrated Forecasting System) (OpenIFS model), the single column model (SCM) and the offline-surface model (OSM). The OpenIFS model provides the full forecast capability of IFS, supporting software and documentation but without the data assimilation system. OpenIFS has a support team at ECMWF for technical assistance but limited resources for detailed scientific assistance.
Quote
The OpenIFS models are free but require an OpenIFS license. Once a license is granted, download access is provided to the source code and accompanying files and tools.
I'm part of their funding as a tax payer and would love to see the same ease of access into their products as the Americans provide.

Rob Dekker

  • Nilas ice
  • Posts: 2386
    • View Profile
  • Liked: 118
  • Likes Given: 119
Re: Home brew AMSR2 extent & area calculation
« Reply #2285 on: July 12, 2016, 08:34:07 AM »
I'm really quite new to weather models, so I have a request :

Can somebody explain these terms that sound like Chinese to me (like GFS and ECMWF and EURO) on the Glossary thread :
http://forum.arctic-sea-ice.net/index.php/topic,153.0.html
And maybe give links to where they are posted, and what their strengths and weaknesses are ?

Much appreciated !
This is our planet. This is our time.
Let's not waste either.

Jim Hunt

  • Young ice
  • Posts: 4744
  • Stay Home, Save Lives
    • View Profile
    • The Arctic sea ice Great White Con
  • Liked: 507
  • Likes Given: 44
Re: Home brew AMSR2 extent & area calculation
« Reply #2286 on: July 12, 2016, 09:04:01 AM »
Can somebody explain these terms that sound like Chinese to me (like GFS and ECMWF and EURO) on the Glossary thread

Here you go:

http://forum.arctic-sea-ice.net/index.php/topic,153.msg83208.html#msg83208
Reality is merely an illusion, albeit a very persistent one - Albert Einstein

Rob Dekker

  • Nilas ice
  • Posts: 2386
    • View Profile
  • Liked: 118
  • Likes Given: 119
Re: Home brew AMSR2 extent & area calculation
« Reply #2287 on: July 12, 2016, 09:16:28 AM »
Can somebody explain these terms that sound like Chinese to me (like GFS and ECMWF and EURO) on the Glossary thread

Here you go:

http://forum.arctic-sea-ice.net/index.php/topic,153.msg83208.html#msg83208

Thanks Jim !
Do you also have the links to where these forecasts are made ? And what is this EURO that people talk about ? (To me EURO is still a currency  ;))
[edit] Ah! I just see that you already mentioned that EURO is the same as ECMWF. Thanks, and my apologies.
This is our planet. This is our time.
Let's not waste either.

Sleepy

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2288 on: July 12, 2016, 09:30:10 AM »
It's the ECMWF (EURO).
There's a good but rather long overview in the user guide to ECMWF forecast
products written by Anders Persson (retired now) but still updated at ECMWF here: http://www.ecmwf.int/sites/default/files/User_Guide_V1.2_20151123.pdf and one easy way to follow the forecasts is through Wettercentrale here: http://old.wetterzentrale.de/topkarten/fsecmeur.html and click on the N-hem.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2289 on: July 12, 2016, 03:48:27 PM »
Centuries in area and extent. Stays in second place after 2012.

Update 20160711.

Extent: -122.3 (-573k vs 2015, -128k vs 2014, -270k vs 2013, +272k vs 2012)
Area: -140.4 (-447k vs 2015, -474k vs 2014, -381k vs 2013, +75k vs 2012)
 
You will find the updated graphs in the top post

Regionally Chukchi has the biggest decline in extent (-33k), followed  Hudson (-27k).

Regional area dropped most in the CAB (-57k) followed by Baffin (-23k) and CAA (-20k).

Regional delta map is Chukchi, a large chunk of ice west of Wrangel Island has disappeared.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2290 on: July 12, 2016, 03:53:51 PM »
The animation is of the Beaufort. The gyre is always fun to watch, but the increase of open open water in a region between Chukchi and the pole is interesting too.

Tor Bejnar

  • Young ice
  • Posts: 3637
    • View Profile
  • Liked: 633
  • Likes Given: 407
Re: Home brew AMSR2 extent & area calculation
« Reply #2291 on: July 12, 2016, 05:19:07 PM »
Wipneus,
Can you help me with some interpretation? 
Attached is a screen shot of "Big Block" from the images just posted.  Given the surface is all purple (with a few spots of lavender), does this mean there are basically no melt ponds on it?  If this is true, isn't that surprising?  Is the 'rainbow' colored edge largely an artifact, given that different pixels cover different proportions of ice floe and open water?
Arctic ice is healthy for children and other living things.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2292 on: July 12, 2016, 05:52:07 PM »
Wipneus,
Can you help me with some interpretation? 
Attached is a screen shot of "Big Block" from the images just posted.  Given the surface is all purple (with a few spots of lavender), does this mean there are basically no melt ponds on it? 

I would say, it means they are not detected.

Quote
If this is true, isn't that surprising?

I think, correct me if I am wrong, that there is cloud, possibly thick cloud cover. This has two effects: first that the measure apparent concentration is higher and second that the effective resolution may be somewhat degraded. You can look at the earlier frames in the animation where the 'big block' (and other floes nearby) is both darker and sharper
Quote

Is the 'rainbow' colored edge largely an artifact, given that different pixels cover different proportions of ice floe and open water?

Yes, plus the loss of effective resolution that I mentioned above. Movement during the day would be another reason.

Tor Bejnar

  • Young ice
  • Posts: 3637
    • View Profile
  • Liked: 633
  • Likes Given: 407
Re: Home brew AMSR2 extent & area calculation
« Reply #2293 on: July 12, 2016, 06:07:18 PM »
Thanks much!
Arctic ice is healthy for children and other living things.

magnamentis

  • Guest
Re: Home brew AMSR2 extent & area calculation
« Reply #2294 on: July 12, 2016, 06:30:59 PM »
remember all the "poof" theory, which will be next and BTW the poor ice condition between ESS and the pole that was not shown in AMSR yesterday is clearly there today, just for those who remember the discussion and
still on course to second lowest without and lowest with GAC.

A-Team

  • Young ice
  • Posts: 2697
    • View Profile
  • Liked: 637
  • Likes Given: 33
Re: Home brew AMSR2 extent & area calculation
« Reply #2295 on: July 12, 2016, 07:06:28 PM »
The reasons for looking at Big Block are its recognizability, persistence, release from the ice pack in mid-February, the existence of a direct overflight transect with many thickness data points, and its size relative to satellite resolutions.

Thus if Modis has 250 m resolution and floe size is 100x100 km, that is 400x400 pixels = decent sized image with rather few edge pixel issues. For the AMSR2 putative resolution of 3.125 km, that pencils out to 32x32 = 1024 pixels, not even a postage stamp. Many would have mixed character, hanging out over some water, and even those truly interior could still be subject to passing cloud artifacts.

As discussed above, the AMSR2 3km palette allows each pixel overlying Big Block to be amplified to a convenient size and its palette value determined. This might not be a good as going back to the raw netCDF but it's good enough. Over a few consecutive days of clear weather, say 08-11 July, change in the distribution of these 1024 values is probably attributable to cloud artifacts (rather than floe microwave emission changes) which might then be somewhat removable. This is not feasible for smaller floes.

There has been massive confusion on the 2016 melt season forum in regards to interpretation of infrared and microwave imagery; some posters there have little interest in the underlying physics, preferring their own take to a learning curve. 

We also have occasional coverage of Big Block by much higher resolution S1A, S2A and Landsat-8. It may also have passed over or near various upward-looking fixed moorings and unmanned surface buoys on other floes. However Big Block has never been visited by researchers and there are no plans to do so as far as we know.

Some satellites carry infrared channels that simultaneously detect some forms of intervening atmosphere effects. However it's a bit of a mystery to me exactly how images are assembled in WorldView from multiple overlapping passes of these satellites, as they change during the course of the day, especially the blue color correction.

JayW has posted some very interesting animations from a site that seems to capture and serve images from each orbital pass over offshore Alaska.  Presumably WorldView layers are made out of these. That gives a much better sense of how fast cloud cover can pass over. AndreasT has made a start on Ceres but no one is looking at Calypso or CloudSat. I would say otherwise that forum binary 'clear' or 'cloudy' is a primitive approach to Arctic Ocean insolation, heat balance and image interpretation.

Terra, Aqua and Suomo pass overhead at slightly different times which helps as clouds move faster than this nearly stationary floe. However the visible shifts are enough at Nares Strait and other Beauford floes that three images from the same date can be animated though the exact time stamps are a bit murky.

The satellite formation called the A Train contains six satellites as of 14 July 2014:

OCO-2, lead spacecraft
GCOM-W1 SHIZUKU follows OCO-2 by 11 minutes
Aqua runs 4 minutes behind GCOM-W1, carries AMSR-E, Ceres and Modis
CloudSat 30 seconds behind Aqua
CALIPSO 15 seconds behind CloudSat
Aura has an offset orbit, lags Aqua by 15 minutes
Terra carries Modis and ASTER, seems not to be in the A-train.
« Last Edit: July 12, 2016, 08:29:16 PM by A-Team »

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2296 on: July 13, 2016, 09:52:47 AM »
Neck to neck race continues, area is now equal (too close to call) with 2012.

Update 20160712.

Extent: -123.2 (-635k vs 2015, -144k vs 2014, -239k vs 2013, +175k vs 2012)
Area: -54.3 (-424k vs 2015, -431k vs 2014, -357k vs 2013, -4k vs 2012)
 
You will find the updated graphs in the top post

Regional extent deline in the Hudson keeps continuing, today the largest (-58k) followed by CAA: -24k.

Chukchi had the largest area drop (-33k), followed by CAA (-29k) and Hudson (-25k). Greenland Sea and ESS had upticks of +27k and +25k.

Regional delta map is of the Barents section. The ice edge is retreating. In the middle of the Barents Sea some ice is popping up and off. My calculation regards this as false ice and does not include it in area and extent calculation. If persistent, it will be added the next day.

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2297 on: July 13, 2016, 10:00:45 AM »
Side by side animation of Laptev sea ice concentration 2014 and 2016. In 2014 ice cover Laptev declined faster than all years in my limited database. 

Click to start.

Andreas T

  • Nilas ice
  • Posts: 1143
    • View Profile
  • Liked: 17
  • Likes Given: 4
Re: Home brew AMSR2 extent & area calculation
« Reply #2298 on: July 13, 2016, 10:06:15 AM »
...

Regionally Chukchi has the biggest decline in extent (-33k), followed  Hudson (-27k).

Regional area dropped most in the CAB (-57k) followed by Baffin (-23k) and CAA (-20k).

Regional delta map is Chukchi, a large chunk of ice west of Wrangel Island has disappeared.
I don't doubt that this ice is in the process of melting at probably at high rate of volume loss and I don't have any information to estimate how much longer it will last but it is worth pointing out that despite falling below the 15% threshold in Wipneus' plot and apparently disappearing from the UB UH plot altogether there is a fair bit of that ice still visible on worldview http://go.nasa.gov/29OTNTc The grainy appearance of this rubble is probably an indication that it consists of the disintegrating parts of thick ice, which supports the earlier assessment of PIOMASS of that region
« Last Edit: July 13, 2016, 10:38:28 AM by Andreas T »

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4110
    • View Profile
    • Arctische Pinguin
  • Liked: 821
  • Likes Given: 0
Re: Home brew AMSR2 extent & area calculation
« Reply #2299 on: July 13, 2016, 10:16:11 AM »
Andreas, I completely agree. Lots of that ice has "reappeared" (SIC is over the 15% threshold) today. This (hoovering around 15% )may keep happening for some time, days or weeks depending on melt conditions and ice thickness.