2016 sea ice area and extent data

[Wipneus]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Thu 2016.4904  -93.3  7.164613 +115.8 11.689271   +22.6 18.853884
Fri 2016.4932  -70.3  7.094268 +117.1 11.806354   +46.7 18.900622
Sat 2016.4959  -44.8  7.049490  +90.7 11.897066   +45.9 18.946556
Sun 2016.4986 -103.7  6.945792  +22.2 11.919313   -81.5 18.865105

Declining most are ESS (-29k), CAB (-26k) and Laptev (-16k).

Shadow NSIDC extent is 9.3344 a drop of -111.9. By extent the leading regions are Hudson (-29k), Laptev (-19k) and Chukchi (-15k).

The attached delta map shows exactly where the changes where.

[Steven]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

Sun 2016.4986  ... 6.945792 

This is the 5th lowest value for this date.  It is about 410k above 2012,  168k above 2010,  132k above 2007,  and 92k above the 2011 value for this date.

[BornFromTheVoid]

NSIDC extent, both the daily and 5 day trailing average, is now above 2010 and 2012.

[Wipneus]

Strong declines, but some of it must be the first-of-the-month effect.

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:
day  CT-date       NH               SH                Global
Fri 2016.4932  -70.3  7.094268 +117.1 11.806354   +46.7 18.900622
Sat 2016.4959  -44.8  7.049490  +90.7 11.897066   +45.9 18.946556
Sun 2016.4986 -103.7  6.945783  +22.4 11.919443   -81.3 18.865226
Mon 2016.5014 -202.8  6.742958   -1.6 11.917858  -204.4 18.660816

I estimate that the FOTM effect may attribute some -35k to the total, so still a strong decline. The CAB and ESS are particular strong: -93k and -53k. The CAA increased +24k.

Shadow NSIDC extent is now 9.1786 dropping -155.8k. The FOTM effect may as high as -100k, so todays drop is certainly inflated.

Attached is the delta map. Notice the red coasts around Okhotsk, Bering, Scandinavia, Newfoundland? That is ice no longer since the first of July.
The huge drop in concentration in the CAB and some increase near the Canadian islands is noticable.
 

[Lord M Vader]

WOW, a double century break by monday! 

[Jim Pettit]

WOW, a double century break by monday! 

Didn't gain much, though; Day 0.5014 has been, historically, the day each year with the highest average daily area decrease. In fact, 10 out of the past 20 Day 5014s have seen a double century area loss, while the average decrease for the day over the past 30 years has been 184k km2. At any rate, both 2012 and 2014 lost more, so 2016 is now in 5th place for only the fifth time this year, and the first time since early February. The leaders: 2012, 2007, 2010, and, in fourth place, 2011.

[Tor Bejnar]

...
Didn't gain much, though; Day 0.5014 has been, historically, the day each year with the highest average daily area decrease. In fact, 10 out of the past 20 Day 5014s have seen a double century area loss, while the average decrease for the day over the past 30 years has been 184k km2. At any rate, both 2012 and 2014 lost more, so 2016 is now in 5th place for only the fifth time this year, and the first time since early February. The leaders: 2012, 2007, 2010, and, in fourth place, 2011.

Thanks, Jim.  Perspective is always helpful!

[magnamentis]

WOW, a double century break by monday! 

that are the data from yesterday, monday as far as i understood is just the day they will be officially published for 3 days back, would appreciate correction in case i see this wrong. however, could well be that:

here comes the cliff
i and others were expecting :-)

[BornFromTheVoid]

Update for the week to July 2nd

The current 5 day trailing average is on 9,312,000km2 while the 1 day extent is at 9,110,000km2.

(All the following data is based on a trailing 5 day average)
The daily anomaly (compared to 81-10) is at -1,529,000km2, an increase from -1,509,000km2 last week. The anomaly compared to the 07, 11 and 12 average is at -225,000km2, a decrease from -437,000km2 last week. We're currently 3rd lowest on record, down from lowest as last week.



The average daily change over the last 7 days was -77.9k/day, compared to the long term average of -74.1k/day, and the 07, 11 and 12 average of -108.4k/day.
The average long term change over the next week is -85.4k/day, with the 07, 11, and 12 average being -126.9k/day.



The extent drop so far this July is the 25th smallest on record. To achieve the largest monthly loss, a drop of at least 115.8k/day is required (requiring >116.8k/day with with single day values), while the smallest loss requires a drop of less than 63.7k/day (<61.0k/day with single day values) and an average loss requires a drop of 83.6k/day (~85.3k/day with single day values).



The extent loss in June was the 17th largest on record, while the average extent was the lowest on record.

[Wipneus]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Sat 2016.4959  -44.8  7.049490  +90.7 11.897066   +45.9 18.946556
Sun 2016.4986 -103.7  6.945783  +22.4 11.919443   -81.3 18.865226
Mon 2016.5014 -202.6  6.743183   -1.7 11.917756  -204.3 18.660939
Tue 2016.5041 -132.7  6.610508  -43.0 11.874745  -175.7 18.485253

Huge drop in the CAB, -79k; Hudson at distance -27k.

Shadow NSIDC extent  9.1101  dropping -68.5k.  Hudson fore runner -33k, Baffin -18k.

The attached delta map shows it all.

[abbottisgone]

The last graph is interesting if June has always been considered a 'stalling month' for ice loss!

<modified the comment to show what oren means; N.>

[oren]

The last graph is interesting if June has always been considered a 'stalling month' for ice loss!

Do you realize that you can comment without quoting the whole post?

[Wipneus]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Sun 2016.4986 -103.7  6.945783  +22.4 11.919443   -81.3 18.865226
Mon 2016.5014 -202.6  6.743183   -1.7 11.917756  -204.3 18.660939
Tue 2016.5041 -132.8  6.610378  -43.0 11.874765  -175.8 18.485143
Wed 2016.5068 -125.4  6.484964  -46.9 11.827910  -172.3 18.312874

Fourth century in a row, supported by CAB (-57k), Hudson (-38k) and CAA (-26k). ESS went up by +17k.

Shadow NSIDC extent is 9.0479  dropping -62.8k. Hudson (-32k) and Kara (-29k) lost most.

In the attached delta image the drop in concentration north of Ellesmere is noticeable.

[bbr2314]

I would think it is likely that this month sees the largest % decline of any July in history, possibly any month on record, though with August starting at an area baseline of 3-4MKM2 it could easily beat July's crash in terms of raw % lost.

With only 6.5MKM2 left, 30 more days of century drops leave us staring 3KM2 in the face...

[magnamentis]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Sun 2016.4986 -103.7  6.945783  +22.4 11.919443   -81.3 18.865226
Mon 2016.5014 -202.6  6.743183   -1.7 11.917756  -204.3 18.660939
Tue 2016.5041 -132.8  6.610378  -43.0 11.874765  -175.8 18.485143
Wed 2016.5068 -125.4  6.484964  -46.9 11.827910  -172.3 18.312874

Fourth century in a row, supported by CAB (-57k), Hudson (-38k) and CAA (-26k). ESS went up by +17k.

Shadow NSIDC extent is 9.0479  dropping -62.8k. Hudson (-32k) and Kara (-29k) lost most.

In the attached delta image the drop in concentration north of Ellesmere is noticeable.

hard to resist to refer to past postings haha... even worse than expected, even the 200k happened while temps above 80N are cooler than average. but the what are 0.2-0.3C less against the 20C less in january and almost through the entire winter. a perfect fit to general conditions, thinner, warmer, less solid and the likes.

@bbr2314 i see you smile big time :-)

[bbr2314]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Sun 2016.4986 -103.7  6.945783  +22.4 11.919443   -81.3 18.865226
Mon 2016.5014 -202.6  6.743183   -1.7 11.917756  -204.3 18.660939
Tue 2016.5041 -132.8  6.610378  -43.0 11.874765  -175.8 18.485143
Wed 2016.5068 -125.4  6.484964  -46.9 11.827910  -172.3 18.312874

Fourth century in a row, supported by CAB (-57k), Hudson (-38k) and CAA (-26k). ESS went up by +17k.

Shadow NSIDC extent is 9.0479  dropping -62.8k. Hudson (-32k) and Kara (-29k) lost most.

In the attached delta image the drop in concentration north of Ellesmere is noticeable.

hard to resist to refer to past postings haha... even worse than expected, even the 200k happened while temps above 80N are cooler than average. but the what are 0.2-0.3C less against the 20C less in january and almost through the entire winter. a perfect fit to general conditions, thinner, warmer, less solid and the likes.

@bbr2314 i see you smile big time :-)

I would be smiling if my ego were more important than the health of the planet...

At this point I think my assertion that this year will go ice-free is looking more and more likely... maybe we maintain a token million KM2 dispersed in old floes along the basin's periphery, but beyond the ice that's still 3-4M+ in thickness, it looks to me like everything else is going to go... note that on HYCOM/etc, the highest concentrations outside of what's pressed up against the CAA is along the ATL coast... right next to an impending scorch of SSTs surpassing 50F.

Equally important is the abysmal state of Greenland Sea ice, which has now allowed the fast ice N of Greenland to begin breaking up and joining what's being exported towards FRAM... another front pressing into the thickest ice there is. Perhaps what's more interesting/impactful over there besides the raw potential of sea ice loss is the water's encroachment into Greenland's glaciers, which are likely going to begin calving in an unprecedented fashion.

[magnamentis]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Sun 2016.4986 -103.7  6.945783  +22.4 11.919443   -81.3 18.865226
Mon 2016.5014 -202.6  6.743183   -1.7 11.917756  -204.3 18.660939
Tue 2016.5041 -132.8  6.610378  -43.0 11.874765  -175.8 18.485143
Wed 2016.5068 -125.4  6.484964  -46.9 11.827910  -172.3 18.312874

Fourth century in a row, supported by CAB (-57k), Hudson (-38k) and CAA (-26k). ESS went up by +17k.

Shadow NSIDC extent is 9.0479  dropping -62.8k. Hudson (-32k) and Kara (-29k) lost most.

In the attached delta image the drop in concentration north of Ellesmere is noticeable.

hard to resist to refer to past postings haha... even worse than expected, even the 200k happened while temps above 80N are cooler than average. but the what are 0.2-0.3C less against the 20C less in january and almost through the entire winter. a perfect fit to general conditions, thinner, warmer, less solid and the likes.

@bbr2314 i see you smile big time :-)

I would be smiling if my ego were more important than the health of the planet...

At this point I think my assertion that this year will go ice-free is looking more and more likely... maybe we maintain a token million KM2 dispersed in old floes along the basin's periphery, but beyond the ice that's still 3-4M+ in thickness, it looks to me like everything else is going to go... note that on HYCOM/etc, the highest concentrations outside of what's pressed up against the CAA is along the ATL coast... right next to an impending scorch of SSTs surpassing 50F.

Equally important is the abysmal state of Greenland Sea ice, which has now allowed the fast ice N of Greenland to begin breaking up and joining what's being exported towards FRAM... another front pressing into the thickest ice there is. Perhaps what's more interesting/impactful over there besides the raw potential of sea ice loss is the water's encroachment into Greenland's glaciers, which are likely going to begin calving in an unprecedented fashion.

- point about ego is taken

- still disagree on "Ice Free"

- generally see things similar

as to ice free, we both always pointed to the amount of energy that is in the system and sooner or later, which is probably now, must have a significant impact. following that very same path and considering the amount of ice that is totally in place by now ( volume ) i think that the energy that it would take to melt all that out within 70-80 days is simply not there, IMO it's as matter of factly not there, as the energy it takes for this years events
was getting in place last year already, i could search to 2015 post where i mentioned that the amount of energy surplus in the system must be excessive while as we all know that did not come to effects last year.

however i hope it's clear that i'm not sure, it's a thesis, let's see what happens and how knows all the positive and negative feedbacks that could overturn things once of a sudden, at least short term.

[Wipneus]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Mon 2016.5014 -202.6  6.743183   -1.7 11.917756  -204.3 18.660939
Tue 2016.5041 -132.8  6.610378  -43.0 11.874765  -175.8 18.485143
Wed 2016.5068 -125.3  6.485084  -46.9 11.827910  -172.1 18.312994
Thu 2016.5096  +47.5  6.532555 +103.1 11.931030  +150.6 18.463585

Huge recovery of the CAB (+109k). CAA (-21k), Hudson (-18k) and ESS (-16k) are not enough to prevent an uptick in the total.

Shadow NSIDC is 8.9542  dropping -93.7k. Hudson (-23k) and Baffing (-17k) are the biggest decliners.

Attached delta map shows where exactly the changes were.

[magnamentis]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Mon 2016.5014 -202.6  6.743183   -1.7 11.917756  -204.3 18.660939
Tue 2016.5041 -132.8  6.610378  -43.0 11.874765  -175.8 18.485143
Wed 2016.5068 -125.3  6.485084  -46.9 11.827910  -172.1 18.312994
Thu 2016.5096  +47.5  6.532555 +103.1 11.931030  +150.6 18.463585

Huge recovery of the CAB (+109k). CAA (-21k), Hudson (-18k) and ESS (-16k) are not enough to prevent an uptick in the total.

temps in the CAB droped 2-4C which makes this "recovery" most probably a matter of overfrozen melt ponds.

from the below pic one can see that even though most of the time things look very similar across various sources this time the discrepancy is astounding.

link to the site with non-cropped images and indices.

[abbottisgone]

The last graph is interesting if June has always been considered a 'stalling month' for ice loss!

Do you realize that you can comment without quoting the whole post?

Do you realise your buddys chose extinction?

[abbottisgone]

I would think it is likely that this month sees the largest % decline of any July in history, possibly any month on record, though with August starting at an area baseline of 3-4MKM2 it could easily beat July's crash in terms of raw % lost.

With only 6.5MKM2 left, 30 more days of century drops leave us staring 3KM2 in the face...

Yeh, well,.. if you're saying 3K is global market panic I think it's already been written in triplicate!

[Neven]

abbottisgone is gone for 30 days.

[BornFromTheVoid]

Big drop on the NSIDC single day extent yesterday, down 154k. However, this was not enough to keep up with 2012, which we are now 166k behind.
We have dropped back below 2010 though, into 2nd lowest on record once again.

[Wipneus]

Double century for CT-area.

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Mon 2016.5014 -202.6  6.743183   -1.7 11.917756  -204.3 18.660939
Tue 2016.5041 -132.8  6.610378  -43.0 11.874765  -175.8 18.485143
Wed 2016.5068 -125.3  6.485084  -46.9 11.827910  -172.1 18.312994
Thu 2016.5096  +47.3  6.532421 +103.1 11.931036  +150.5 18.463457
Fri 2016.5123 -228.2  6.304220 +188.1 12.119158   -40.1 18.423378

Revenge of the CAB: -96k, and support from Laptev (-38k) and Kara (-29k).

Shadow NSIDC is 8.7998 a drop of -154.4k. Big drops in Hudson (-43k), Baffin (-37k), Kara (-26k) and Beaufort (-20k).

The attached delta map shows where the big concentration drops exactly were.

[Wipneus]

Version 2.0 of NSIDC Sea Ice Index has arrived, including details what the difference with previous version 1.2 is.

 

NOAA@NSIDC is pleased to announce the release of Sea Ice Index Version 2. Improvements include using the most recently available version of the Goddard Space Flight Center (GSFC) input sea ice concentration data and revising three computations in the Sea Ice Index processing code. The new GSFC data have undergone additional manual quality control procedures at the GSFC that go farther to remove spurious ice. The GSFC data are available as their own data set from the NSIDC DAAC as Sea Ice Concentrations from Nimbus-7 SMMR and DMSP SSM/I-SSMIS Passive Microwave Data [http://nsidc.org/nsidc-0051]. The other three updates adjust calculations present in the old Sea Ice Index processing code that were in need of improvement. Because of these updates, minor changes in some of the ice extent and area numbers can be seen with the average change in the trends over all 12 months being on the order of 0.05 percent/decade (approximately 500 sq. km/year).

For a complete description of these changes and their effects, please see:
http://nsidc.org/data/docs/noaa/g02135_seaice_index/#jul-2016

Access to the data and documentation is provided on the Sea Ice Index web site at: https://nsidc.org/data/seaice_index/

In practice there seem to be no big changes for the daily values. The biggest changes are in the past (as in 20th century). I have checked my (shadow) calculations for 2016 with those of NSIDC, and found perfect match (differences between +/- 500 km2, because NSIDC rounds at thousands of km2).

That may not be true for NSIDC's monthly averages (area and extent). That may have some influence on the September average value used by Neven's pol and SIPN's Sea Ice Outlook.
Since I don't calculate monthly values, I cannot comment on the actual changes this has.

[Buddy]

Shadow NSIDC is 8.7998 a drop of -154.4k. Big drops in Hudson (-43k), Baffin (-37k), Kara (-26k) and Beaufort (-20k).

Certainly looks like we are in "the meat" of the melt season.  A LOT of "crappy ice" could give us a month of sharp daily drops.  The central Russian coast now has a good "foothold" of open water and that should help.  Still expect SUB 2012 extent come mid September...

[Jim Pettit]

ADS-NIPR (IJIS) Extent:

8,401,287 km2 (06 July)

Down 5,541,220 km2 (39.74%) from 2016 maximum of 13,942,507 km2 on 29 February.

5,223,832 km2 above record minimum extent of 3,177,455 km2 (16 September 2012).

Down 85,785 km2 (-1.01%) from previous day.

Down 632,442 km2  (-7.%) over past seven days (daily average: -90,349 km2).

Down 572,421 km2  (-4.68%) for July (daily average: -95,404 km2).

1,032,096 km2 below 2000s average for this date.

212,659 km2 below 2010s average for this date.

695,050 km2 below 2015 value for this date.

21,712 km2 below 2012 value for this date.

1st lowest year-to-date (01 January - 06 July) average.

1st lowest July to-date average.

1st lowest value for the date.

127 days this year (67.91% year-to-date) have recorded the lowest daily extent.

33 days (17.65%) have recorded the second lowest.

14 days (7.49%) have recorded the third lowest.

174 days in total (93.05%) have been among the lowest three on record.

CT Area:

6,304,220 km2 (07 July [Day 0.5123])

Down 6,617,138 km2 (51.21%) from 2016 maximum of 12,921,358 km2 on 29 March [Day 0.2384].

4,070,211 km2 above record minimum area of 2,234,010 km2 (14 September 2012).

Down 228,201 km2 (-3.49%) from previous day.

Down 790,048 km2 (-11.21%) over past seven days (daily average: -112,864 km2).

Down 790,048 km2 (-7.09%) for July (daily average: -112,864 km2).

814,801 km2 below 2000s average for this date.

205,664 km2 below 2010s average for this date.

511,039 km2 below 2015 value for this date.

244,289 km2 above 2012 value for this date.

3rd lowest value for the date.

* - NOTE: due to the prolonged absence of official CT sea ice area data, I've incorporated Wipneus' area numbers as calculated from NSIDC data. The official numbers will be inserted if/when available. In the meantime, thanks, Wipneus!

[Wipneus]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Wed 2016.5068 -125.3  6.485084  -43.0 11.874765  -168.3 18.359849
Thu 2016.5096  +47.3  6.532421  -46.9 11.827910    +0.5 18.360331
Fri 2016.5123 -228.3  6.304103 +103.1 11.931036  -125.2 18.235139
Sat 2016.5151 -149.5  6.154631 +188.1 12.119158   +38.7 18.273789

Biggest supporters are the CAB (-54k), ESS (-38k) and Baffin (-21k).

Shadow NSIDC extent is now 8.7799  a drop of -19.9k. Biggest drop is Baffin (-29k), Beaufort increased by +16k.

The attached delta map shows concentration drops in he central Arctic.

[Steven]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

Sat 2016.5151  ...  6.154631

Now in 3rd place, behind 2012 and 2010.  Here are the CT-area values for this date (day 0.5151):

2012      6.04  million km²
2010      6.15
2016     6.15
2007      6.19
2011      6.23
2015      6.65
2006      6.71
2014      6.73
2008      6.74
2013      6.77
...

[bbr2314]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Wed 2016.5068 -125.3  6.485084  -43.0 11.874765  -168.3 18.359849
Thu 2016.5096  +47.3  6.532421  -46.9 11.827910    +0.5 18.360331
Fri 2016.5123 -228.3  6.304103 +103.1 11.931036  -125.2 18.235139
Sat 2016.5151 -149.5  6.154631 +188.1 12.119158   +38.7 18.273789

Biggest supporters are the CAB (-54k), ESS (-38k) and Baffin (-21k).

Shadow NSIDC extent is now 8.7799  a drop of -19.9k. Biggest drop is Baffin (-29k), Beaufort increased by +16k.

The attached delta map shows concentration drops in he central Arctic.

The hurricane-esque gyre that now dominates the entire basin is plainly obvious to me here but perhaps I am seeing things...

[plinius]

The hurricane-esque gyre that now dominates the entire basin is plainly obvious to me here but perhaps I am seeing things...

Correct, this is a common human perception issue to identify structures that are not really there. Great for preventing being eaten by a tiger, but also good for seeing ghosts, gods, conspiracies, or a hurricane in your water glass ;-).
http://www7320.nrlssc.navy.mil/hycomARC/navo/arcticicespddrfnowcast.gif
the centre of that drift curl does not coincide with your perceived hurricane (apart from that - never forget that clouds, rain bands, etc. imprint on the sea ice area, so you always have arcs there that are not real).

[magnamentis]

Correct, this is a common human perception issue to identify structures that are not really there. Great for preventing being eaten by a tiger, but also good for seeing ghosts, gods, conspiracies, or a hurricane in your water glass ;-)

you must have very good friends here that you get away with that condescending tone the entire day and without a forced sabatical. you discourage people which i think is the opposite of the purpose here, even if you were 100% right but at least half of your reasoning was plain wrong.

[Neven]

Magnamentis, I agree with you in principle, but bbr2314 has been dishing out himself too lately. Let's not turn this into a snowball. It's the Internet.

---

With CT's string of century breaks, my crude compactness maps are getting interesting. For instance, 2016 just dipped below 2012 on the CAMAS graph (Cryosphere Today area divided by MASIE):

[Rob Dekker]

Neven and Wipneus,
I have been following the ice concentration numbers for a while now, and I noticed that the higher resolution observations tend to show a record, while the lower resolution numbers do not.

For example, from the "Home brewed AMSR2" thread, Wipneus reported again that 2016 is in the lead for ice concentration (with only 2015 being almost equal in concentration, but 700k behind in extent) :

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)

The ice concentration map from Wipneus tends to show the same thing :
https://sites.google.com/site/arctischepinguin/home/amsr2/grf/amsr2-compact-compare.png

High resolution (AMSR2 3.125 km resolution) is in the lead for 2016, while lower resolution (Bootstrap AMSR2 10 km, or NASA team SSIMS 25 km resolution) is lagging behind, and not showing 2016 in the lead (yet ).

I've been thinking about that difference, and it seems to me that if high resolution shows lower ice concentration than lower resolution, considering that ice concentration in the main pack should not matter at which resolution it is measured, that maybe the ice edge is smaller (less fragmented) this year than in other years. But considering the fragmented ice edge in areas like the Beaufort, that explanation is not satisfying.

[edit] Come to think of it, if the ice edge is highly fragmented at the smaller than 25 km (or even 10 km) resolution, then highest resolution (3.125 km) will show more high-resolution pixels with low ice concentration than low-resolution pixels with low ice concentration, which will lower the high-resolution ice concentration and explain the observations.
Is that what is going on ?

Wondering if you have any rational explanation for this apparent difference in ice concentration from  the different resolution observations.

[Peter Ellis]

Rob:  Fundamentally, it means that areas that are of intermediate concentration in a low resolution source are being split into a mix of high concentration pixels and open water when you look at a higher resolution. Since melt ponds are of the order of metres across, they will not be resolved even by the highest resolution source - however openings and leads may be wide enough to be resolved at higher resolutions.

Ergo, I suspect what this means is that this year there is more open water within the pack and less melt ponding than in previous years.  That's consistent with the weather conditions which have favoured divergence over direct solar input.

[Rob Dekker]

Ergo, I suspect what this means is that this year there is more open water within the pack and less melt ponding than in previous years.  That's consistent with the weather conditions which have favoured divergence over direct solar input.

Thanks Peter, but that "this year there is more open water within the pack and less melt ponding than in previous years" does not explain the different observations from different resolution methods. After all, if open water this year is small (<3 km), it would gain the same ice concentration for high and low resolution methods. And if it is large (>25 km) it would gain more area of low ice concentration for low resolution methods, and thus we should see lower concentration for low resolution methods.

[edit] Maybe only a fragmentation in between (3 and 25 km) could explain the observations ?

[Bill Fothergill]

@Rob & Peter

I think I can see how this could happen - or I could simply be misunderstanding the issue. (It comes with the growing number of candles on the birthday cake - the local fire department now insist on being notified in advance!)

The thought experiment goes like this...

Imagine that "System A" has resolution down to 10 sq kms, whereas "System B" only has 40 sq km resolution. Now imagine a 40 sq km block that System B measures as having 14% concentration. Using the normal 15% threshold, this block would be counted as empty for both area and extent.

However, if System A looks at the same block, it might see three 10 sq km blocks, each with 15% concentration and a fourth 10 sq km block with 11%. This would therefore be register as 30 sq kms for extent and 4.5 sq kms for area.

If this situation gets repeated over a large section of the ice, then the higher-resolution system would be clocking up a significant contribution from the low-concentration peripheral regions. Should this be manifest over a sufficiently large region, it would have the effect of dragging the overall concentration downwards.


(NB It is not too difficult a mental exercise to change the conditions of the thought experiment such that the overall effect would be to increase the overall concentration, but that would be by eliminating smaller polynas within zones of already high concentration.)

[Richard Rathbone]

Some (and maybe most) is going to be down to the algorithm difference, not the resolution difference.

[Wipneus]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Thu 2016.5096  +47.3  6.532421 +103.1 11.931036  +150.5 18.463457
Fri 2016.5123 -228.3  6.304103 +188.0 12.119058   -40.3 18.423161
Sat 2016.5151 -149.5  6.154631 +327.0 12.446055  +177.5 18.600686
Sun 2016.5178 -117.7  6.036904 +173.2 12.619298   +55.5 18.656202

CAB area dropped -67k, Hudson follows with -31k. Laptev increased by +22k.

Shadow NSIDC extent is 8.6858  dropping -94.1k. Hudson continues its nose dive drops -40k.  Chukchi is second with -20k.

Delta map attached for illustration.

[Wipneus]

Neven and Wipneus,
I have been following the ice concentration numbers for a while now, and I noticed that the higher resolution observations tend to show a record, while the lower resolution numbers do not.

For example, from the "Home brewed AMSR2" thread, Wipneus reported again that 2016 is in the lead for ice concentration (with only 2015 being almost equal in concentration, but 700k behind in extent) :

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)

The ice concentration map from Wipneus tends to show the same thing :
https://sites.google.com/site/arctischepinguin/home/amsr2/grf/amsr2-compact-compare.png

High resolution (AMSR2 3.125 km resolution) is in the lead for 2016, while lower resolution (Bootstrap AMSR2 10 km, or NASA team SSIMS 25 km resolution) is lagging behind, and not showing 2016 in the lead (yet ).

I've been thinking about that difference, and it seems to me that if high resolution shows lower ice concentration than lower resolution, considering that ice concentration in the main pack should not matter at which resolution it is measured, that maybe the ice edge is smaller (less fragmented) this year than in other years. But considering the fragmented ice edge in areas like the Beaufort, that explanation is not satisfying.

[edit] Come to think of it, if the ice edge is highly fragmented at the smaller than 25 km (or even 10 km) resolution, then highest resolution (3.125 km) will show more high-resolution pixels with low ice concentration than low-resolution pixels with low ice concentration, which will lower the high-resolution ice concentration and explain the observations.
Is that what is going on ?

Wondering if you have any rational explanation for this apparent difference in ice concentration from  the different resolution observations.

Rob, first I am not aware about significant differences. Maybe because I am not paying attention enough. By extent 2016 3.125km UH extent is nearly the same as 2012 12.5km UH extent (fraction of a day difference). Same if I look at the ADS/Jaxa graph.

Resolution has some impact on extent, but not that much: I am comfortable with using the 12.5 km for 2012 and comparing it with 3.125 km calculations without any adjustment.
Regional extent is another matter, if you look at the Beaufort extent graph see how the lines for 10km Jaxa and 3.125 Uni Hamburg have separated because of the many leads and gaps that the latter can resolve but the first not. On total arctic extent the relative difefrence is much smaller.

And indeed I think that other factors, like the sea ice concentration algorithm are just as important. The 'NASA Team' algorithm used by NSIDC has the much lower concentration visible in the compactness graph.

Do not look at the CT-area numbers. I have summarized the questionable nature several times. Just the fact that 2016 has some more ice in lower latitudes (Hudson and Baffin) and less at higher (Beaufort, Barents) is enough to skew the rankings (low latitude ice is exaggerated, high latitudes underestimated.

[Rob Dekker]

Bill, Richard, Wipneus, Thanks, guys !

Bill, if I read your examples of System A and B correctly, then it shows what I suspected : If ice fragmentation is between 3 k and 25 km, then only can higher resolution method show a lower ice concentration than a lower resolution method.
But you are correct that the fragmentation has to be close to the 15% cut-off.
After all, is system A would have fragmentation variations around, say, 25% or 50%, then both high and low resolution methods would obtain the same “ice-concentration”…

Richard

Some (and maybe most) is going to be down to the algorithm difference, not the resolution difference.

Wipneus also commented that, but does either one of you know of such an algorithmic difference that obtains different ice concentration from the same source satellite data ?

Meanwhile, ice concentration for the NASA team algorithm is rapidly adjusting to the other (higher resolution) methods, in showing that 2016 ice concentration really is quite low, and possibly approaching record level :

https://14adebb0-a-62cb3a1a-s-sites.googlegroups.com/site/arctischepinguin/home/amsr2/grf/amsr2-compact-compare.png?attachauth=ANoY7cqRjXxWOpxudwqjcHAY9Xw1gDHGSly6aQOUmaaGVokfZ9VUQaZFKft6alqr8SFzrRJ2NTn9XmdMUOy8fWQiCJaWja95bHtTMpPiInPdQfrd3VCpUTLkYpCulz2P5O7oQyX7eYTCcXfdM36NdyfCouWcktL6g8iVD3khxL-eE7CSgrsPl6JatjFJfVvICIaDS3fCyy0bx45mN8rO_Q9iLWoIdyL6GaYSzDDU0ciPyYk6g10bcvSiFT5ZuC8jzefiMQJc0a66&attredirects=0

[Andreas T]

.....

Some (and maybe most) is going to be down to the algorithm difference, not the resolution difference.

Wipneus also commented that, but does either one of you know of such an algorithmic difference that obtains different ice concentration from the same source satellite data ?

......

have a look at https://www.researchgate.net/profile/Natalia_Ivanova5/publication/272794543_Satellite_passive_microwave_measurements_of_sea_ice_concentration_an_optimal_algorithm_and_challenges/links/54eeeedd0cf2e55866f3b64f.pdf
this is freely accessible, but very technical (not very accessible in that sense). But it gives an impression what algorithms do and how they can give different results by dealing with weather and surface conditions differently.
 It is also useful to be aware that the seaice concentration plots that we look at are the result of processes which have to interpret the brightness temperature data coming in (have a look at AMSR2 data on worldview) which do not give by themselves straightforward ice/ no ice information.

[Wipneus]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Fri 2016.5123 -228.3  6.304103 +188.0 12.119058   -40.3 18.423161
Sat 2016.5151 -149.5  6.154631 +327.0 12.446055  +177.5 18.600686
Sun 2016.5178 -118.1  6.036550 +173.2 12.619207   +55.1 18.655757
Mon 2016.5205  -56.2  5.980352 +232.0 12.851168  +175.8 18.831520


None of the regions really stands out, Beaufort has the biggest loss at -18k.

Shadow NSIDC extent is 8.5593 a drop of -126.4k. Here Beaufort drops big way with -67k. Greenland Sea is second at a distant -23k.

The attached delta map illustrates these numbers. It reveals the CAB having many changes in local concentration that mostly cancel, so no big change there.

[Rob Dekker]

.....

Some (and maybe most) is going to be down to the algorithm difference, not the resolution difference.

Wipneus also commented that, but does either one of you know of such an algorithmic difference that obtains different ice concentration from the same source satellite data ?

have a look at https://www.researchgate.net/profile/Natalia_Ivanova5/publication/272794543_Satellite_passive_microwave_measurements_of_sea_ice_concentration_an_optimal_algorithm_and_challenges/links/54eeeedd0cf2e55866f3b64f.pdf
this is freely accessible, but very technical (not very accessible in that sense). But it gives an impression what algorithms do and how they can give different results by dealing with weather and surface conditions differently.
 It is also useful to be aware that the seaice concentration plots that we look at are the result of processes which have to interpret the brightness temperature data coming in (have a look at AMSR2 data on worldview) which do not give by themselves straightforward ice/ no ice information.

Thank you Andreas.
I always thought that the significant difference in methods from Wipneus concentration graphs :

https://14adebb0-a-62cb3a1a-s-sites.googlegroups.com/site/arctischepinguin/home/amsr2/grf/amsr2-compact-compare.png?attachauth=ANoY7cp_hktB3zd5pIuHTBDHQkFMea-QccNN1rwFNqyS49CbJah5Umem5XV5ZIGh1ViTiIxoayh6a7uz3VMEbwPzGG3SFaEHh6dyfNx_7Ajr2ur_LMkL8fZ5WBAlTgTPgd9Pd9RPwTAl9XjsOJFbzMweyDi0ATaDeci2aMkG5QLoz-QPQay6r6GwLJb6i1VoIyQzf79S9T2WmL5yrayIyxCp5UjNrKXgck0c_BEDMF5PgYSu7go55UFk2qmXma1U1h-bc7SxlGnS&attredirects=1

between the NASA Team algorithm on SSMIS 25 km and Bootstrap on AMSR2 10 km (as well as Uni Hamburg ASI algorithm on AMSR2 3.125 km) was caused by the resolution difference of these three methods.

But after reading the paper you referenced, and some quick "back-of-the-envelope" calculations on how much "resolution" can affect "ice concentration", I'm not so sure any more.

[Wipneus]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Sat 2016.5151 -149.5  6.154631 +327.0 12.446055  +177.5 18.600686
Sun 2016.5178 -118.1  6.036550 +173.2 12.619207   +55.1 18.655757
Mon 2016.5205  -56.1  5.980457 +232.2 12.851426  +176.1 18.831883
Tue 2016.5233 -156.3  5.824159 +140.1 12.991487   -16.2 18.815646

The big decline is lead by CAB (-58k), ESS (-35k) and Hudson (-25k).

Shadow NSIDC extent is now 8.5125 a drop of -46.9k. Hudson drops -52k, Beaufort 'recovers' +34k.

The attached delta map shows the details.

[Wipneus]

I always thought that the significant difference in methods from Wipneus concentration graphs :

https://14adebb0-a-62cb3a1a-s-sites.googlegroups.com/site/arctischepinguin/home/amsr2/grf/amsr2-compact-compare.png?attachauth=ANoY7cp_hktB3zd5pIuHTBDHQkFMea-QccNN1rwFNqyS49CbJah5Umem5XV5ZIGh1ViTiIxoayh6a7uz3VMEbwPzGG3SFaEHh6dyfNx_7Ajr2ur_LMkL8fZ5WBAlTgTPgd9Pd9RPwTAl9XjsOJFbzMweyDi0ATaDeci2aMkG5QLoz-QPQay6r6GwLJb6i1VoIyQzf79S9T2WmL5yrayIyxCp5UjNrKXgck0c_BEDMF5PgYSu7go55UFk2qmXma1U1h-bc7SxlGnS&attredirects=1

between the NASA Team algorithm on SSMIS 25 km and Bootstrap on AMSR2 10 km (as well as Uni Hamburg ASI algorithm on AMSR2 3.125 km) was caused by the resolution difference of these three methods.

But after reading the paper you referenced, and some quick "back-of-the-envelope" calculations on how much "resolution" can affect "ice concentration", I'm not so sure any more.

That is true. The drop in "compactness" is mostly caused by the effects of liquid water: from dry ice/snow to wet ice/snow to melt ponds. NSIDC sea ice concentration, calculated by the NASA Team algorithm, is more sensitive than the other two methods used in that graph.

The effects of resolving leads and open water are there and similar, but not that big.

[magnamentis]

day  CT-date       NH               SH                Global
Sat 2016.5151 -149.5  6.154631 +327.0 12.446055  +177.5 18.600686
Sun 2016.5178 -118.1  6.036550 +173.2 12.619207   +55.1 18.655757
Mon 2016.5205  -56.1  5.980457 +232.2 12.851426  +176.1 18.831883
Tue 2016.5233 -156.3  5.824159 +140.1 12.991487   -16.2 18.815646
[/tt]

that makes accounts for 120k per day over the last 4 days, so much to lack of momentum. i think that most people don't pay enough attention that 20% ice coverage still counts full to extent, hence in this phase
area is the better indicator of what's going on. extent can theoretically drop from 100 to zero in almost no time once the last cm of ice cover is reached. that's was an example to make my point not sayin' that the ice will go to zero extent withing a few days, hence "theoretically possible" not likely to happen of course.

[Tor Bejnar]

I do get tired of reading absolutest statements like "area is the better indicator of what's going on."

Area is a great indicator of some things, and a so-so indicator of other things. 

If you want a clue about what the current heat gain in the Arctic is, area is an indicator, along with cloud cover and time of year and latitude (and probably some others), you'll want to track.

If you want to know how much of the Arctic is covered in ice, you will want to pick a number somewhere between "area" and "extent".  All those melt ponds in July are counted as being water and not ice, when, in fact, the sea is completely covered with ice under every single melt pond.  The "area" numbers are a great indicator of actual ice area only after the establishment of freezing surface temperatures in ~late September and October and before the onset of new melt ponds in ~May.

If one's concern is shipping, extent maps are probably the best indicator of where it is safe to float.  For anyone wanting to know if they need a parka before stepping out of their heated ship's cabin, a thermometer's reading is probably the best indicator of "what's going on."

[BornFromTheVoid]

Update for the week to July 9th

The current 5 day trailing average is on 8,667,000km2 while the 1 day extent is at 8,512,000km2.

(All the following data is based on a trailing 5 day average)
The daily anomaly (compared to 81-10) is at -1,576,000km2, an increase from -1,529,000km2 last week. The anomaly compared to the 07, 11 and 12 average is at +19,000km2, an decrease from -225,000km2 last week. We're currently 3rd lowest on record, the same as last week.



The average daily change over the last 7 days was -92.0k/day, compared to the long term average of -85.4k/day, and the 07, 11 and 12 average of -126.9k/day.
The average long term change over the next week is -82.9k/day, with the 07, 11, and 12 average being -111.3k/day.



The extent drop so far this July is the 14th largest on record. To achieve the largest monthly loss, a drop of at least 123.3k/day is required (requiring >127.8k/day with with single day values), while the smallest loss requires a drop of less than 54.6k/day (<52.4k/day with single day values) and an average loss requires a drop of 84.5k/day (~85.2k/day with single day values).

[Wipneus]

Here is the shadow CT-area report based on calibrated F18 NSIDC sea ice concentration data:

day  CT-date       NH               SH                Global
Sun 2016.5178 -118.1  6.036550 +173.2 12.619207   +55.1 18.655757
Mon 2016.5205  -56.1  5.980457 +232.2 12.851426  +176.1 18.831883
Tue 2016.5233 -156.1  5.824326 +140.1 12.991531   -16.0 18.815857
Wed 2016.5260 -108.8  5.715556  +14.6 13.006105   -94.2 18.721661

The century is supported by the CAB (-58k) and Hudson (-31k).

Shadow NSIDC extent is 8.4181 dropping  -94.4k. Hudson is star: -76k. Chukchi increased +24k, Beaufort dropped -20k.

Attached delta map shows what is going on.