Home brew AMSR2 extent & area calculation

[etienne]

Which is nice, but does this source have any data before of after these dates ?

As the topic title suggests, Wipneus provides area and extent data from the subsequent AMSR2 instrument. What are you and/or Etienne endeavouring to achieve?

Hello,

This graph is more or less what I tried to get :


http://kuroshio.eorc.jaxa.jp/JASMES/climate/data/graph/JASMES_CLIMATE_SIE_197811_000000_5DAVG_PS_9999_LINE_NHM_201.png

But not as a trend, as a calculated average of yearly values. It is not so easy to manage these values, so I think the trend will be ok for me. Still quite surprising that the trend is -173.321 km^2 per day. That's a lot. If the trend stays the same, there would be around 80 more years to go before an ice free artic summer, but this probably won't be the case because nothing else is linear.

In the graph I made with the 2002 to 2011 data, the trend on extend is about 350 km^2 per day, so there would be only 40 years left. I wonder how many years would be left if we had more recent data.

Best regards,

Etienne

[AmbiValent]

Question to Wipneus: Over at Shadow CT area, 2016 seems to have fallen to 4th place, while with AMSR2, 2016 area is still below 2012. Would it be still in first place, or not, or is there not enough data?

[Jim Hunt]

But not as a trend, as a calculated average of yearly values.

In which case perhaps what you after is the contents of?

ftp://sidads.colorado.edu/DATASETS/NOAA/G02135/north/daily/data/

As you suggest, I too have my doubts about the value of simply extrapolating extent numbers, but there you will find several decades of them to process as you wish.

[etienne]

ftp://sidads.colorado.edu/DATASETS/NOAA/G02135/north/daily/data/

Thank you very much. Here is the average extend I get. Average is always done on the 364 days before and including actual date (89-12-31 is the average for the year 1989).

I only used the easy data, which means years with 365 days of data available.


Some years have 366 days, this explains why dates are not always December 31st at the bottom of the graph.

I am surprised by the result, I have to think a little bit more about it. Maybe I am not enough aware of climatological events to be able to understand what is shown.

Best regards,

Etienne

Added 2016-06-26 : I think that what surprises me is that we talk everywhere of a record breaking year (CO2 concentration, average temperature on earth, maybe record breaking sea ice extend in september...), but if you look at this graph, you get the feeling that this year is below the trend (which means higher sea ice extend as expected). Do we have really low melting weather ? Is there an external ice supply ? Are we loosing volume without loosing extend ? or is this graph nonsense ?

Added 2016-06-27. I calculated the trend line. The slope is 170km^2 per day and we are 10733km^2 higher than the trend. The minimum is in June 2013, which means the average between June 2012 and June 2013 which is not a surprise.

Added 2016-06-28. I calculated the standard deviation : 0.2E6 km^2, or 200'000 km^2.

[oren]

The animation is of the Greenland Sea. Now even the pack ice in the north is going, leaving just a lot of fast ice which is also under attack.

The situation in the Greenland Sea looks totally unprecedented. Can anyone comment if there's been such a state of the ice before around this date? At least from the regional charts it seems not.

[Neven]

The situation in the Greenland Sea looks totally unprecedented. Can anyone comment if there's been such a state of the ice before around this date? At least from the regional charts it seems not.

If you look at the sea ice concentration comparison maps on the ASIG for June 25th, you'll see that this is quite unprecedented. In fact, it's pretty crazy what is going on on the Atlantic side of the Arctic. Only 2012 comes close, but not in the Greenland Sea.

[Rob Dekker]

Wipneus, it may be that you already addressed this issue before, but I have a question :

In your latest AMSR2 report above you state :

Extent: -65.5 (-562k vs 2015, -414k vs 2014, -934k vs 2013, -165k vs 2012)
Area: -82.4 (-441k vs 2015, -320k vs 2014, -839k vs 2013, -273k vs 2012)

This suggests that 2016 ice concentration is lower now than it was in 2012 at the same date (area difference is bigger than the extent difference).

But the concentration maps you provide on Neven's "regional" site :
https://sites.google.com/site/arcticseaicegraphs/regional
suggest that at this time, 2016 ice concentration is higher than in 2012.
Could you shed some light on that apparent contradiction ?

Maybe related : I understand that AMSR2 data was not available until at least late July 2012.
So which data source (extent and area) do you use for 2012 in the AMSR2 updates ?

[iceman]

The situation in the Greenland Sea looks totally unprecedented. Can anyone comment if there's been such a state of the ice before around this date? At least from the regional charts it seems not.

If you look at the sea ice concentration comparison maps on the ASIG for June 25th, you'll see that this is quite unprecedented. In fact, it's pretty crazy what is going on on the Atlantic side of the Arctic. Only 2012 comes close, but not in the Greenland Sea.

My impression is that Fram Strait export has been relatively low in recent months.  If so, what looks bad in Greenland Sea provides a measure of protection for CAB.

[Wipneus]

Update 20160625.

Extent: -67.2 (-577k vs 2015, -394k vs 2014, -878k vs 2013, -98k vs 2012)
Area: -94.4 (-472k vs 2015, -361k vs 2014, -752k vs 2013, -239k vs 2012)
 
You will find the updated graphs in the top post

Hudson is showing yet another day with sustained declines in both extent (-35k) and area (-39k).
Area is dropping faster than area caused by the drops in CAB (-20k) and ESS (-22k).

The regional delta map is from Hudson. Looking at the regional graphs, if the current year behaves anything like 2012,2013 or 2014 the remaining ice will disappear very fast in the coming week or two. As 2015 shows, it may take a lot longer as well.

[Wipneus]

The animation is of the East Siberian Sea. The fast ice "behind" the New Siberian Islands is breaking up.

[Wipneus]

And the surface melt situation which has not changed much.

[AmbiValent]

The maps from 23rd and 24th seem to be mixed up.

[Wipneus]

Wipneus, it may be that you already addressed this issue before, but I have a question :

In your latest AMSR2 report above you state :

Extent: -65.5 (-562k vs 2015, -414k vs 2014, -934k vs 2013, -165k vs 2012)
Area: -82.4 (-441k vs 2015, -320k vs 2014, -839k vs 2013, -273k vs 2012)

This suggests that 2016 ice concentration is lower now than it was in 2012 at the same date (area difference is bigger than the extent difference).

The numbers are smaller yes.

But the concentration maps you provide on Neven's "regional" site :
https://sites.google.com/site/arcticseaicegraphs/regional
suggest that at this time, 2016 ice concentration is higher than in 2012.
Could you shed some light on that apparent contradiction ?

I fail to see this, some regions 2016 is higher other regions is lower. The sum of those 14 regions should be the same as the numbers above.

Maybe related : I understand that AMSR2 data was not available until at least late July 2012.
So which data source (extent and area) do you use for 2012 in the AMSR2 updates ?

The Uni Hamburg AMSR2 data source begins only in 2016. From the same source there is this SSMIS data, available from 2012 to October 2014. Since during 2013 and 2014 the extent and area calculated from SSMIS track the results from AMSR2 very well, I use this for 2012 extent and area.

[Wipneus]

The maps from 23rd and 24th seem to be mixed up.

They are 24 and 25. Thanks, corrected the labels.

[Wipneus]

Question to Wipneus: Over at Shadow CT area, 2016 seems to have fallen to 4th place, while with AMSR2, 2016 area is still below 2012. Would it be still in first place, or not, or is there not enough data?

That may be so. Note that I use data that only spans 2012 and later. I cannot say anything about 2007 and 2010 from these data.

Further CT-area is a peculiar measure. In the first place since it is based on NSIDC sea ice concentration, it will underestimate concentration melting ice more than some other data (Uni Hamburg, Jaxa) do. That may be a factor.

In the second place CT's calculations have some questionable properties:
- it includes lake ice. During the summer any alke ice is false, but not masked away or something;
- it does not calculates area of each grid cell correctly. The result is that ice on low latitudes is over estimated and at high latitudes (>70^o) underestimated;
- All ice is counted, there is no cut-off of 15% or something.
- I have never seen CT correct its numbers after NSIDC made corrections or enhancements; 

[magnamentis]

My impression is that Fram Strait export has been relatively low in recent months.  If so, what looks bad in Greenland Sea provides a measure of protection for CAB.

IMO CAB looks bad enough to melt on spot hence that protection only protects vulnerable ice if at all, while i see your point and it's a valid point to be considered as a possibility. the factual "provides" is a bit bold, "could provide" would be a bit more suitable.

[Richard Rathbone]

Question to Wipneus: Over at Shadow CT area, 2016 seems to have fallen to 4th place, while with AMSR2, 2016 area is still below 2012. Would it be still in first place, or not, or is there not enough data?

Further CT-area is a peculiar measure. In the first place since it is based on NSIDC sea ice concentration, it will underestimate concentration melting ice more than some other data (Uni Hamburg, Jaxa) do. That may be a factor.

I think Wipneus's ranking of 1st now is the more reliable one, but I take the CT ranking as an indicator of relative lack of momentum for this reason. By counting a lot of water covered ice as water now, the error in CT provides a leading indicator.

[Peter Ellis]

Remember that they've had to transition to a new satellite. NSIDC have calibrated their algorithm for extent measurement, but I don't think the area measurements have been fully calibrated. This means we have no idea whether this year's CT-Area data is more susceptible to confusion by melt ponds, or less susceptible.

[Rob Dekker]

Peter, NSIDC is quite forthcoming with their statements that they calibrated all their ice data :

As of June 14, 2016, NSIDC has completed the transition to the Defense Meteorological Satellite Program (DMSP) F-18 satellite for sea ice data. Sea Ice Index updates have also resumed.

and

The daily Sea Ice Index provides a quick look at Arctic-wide changes in sea ice. It provides consistently processed daily ice extent and concentration images and data since 1979.

http://nsidc.org/data/seaice_index/

And they resumed updates of their ice concentration maps, the latest one here



So I do not see any reason to assume that NSIDC did not calibrate their 'area' (concentration) numbers.

Regarding CT - area, and what it means, Wipneus did give a reasonable overview of why it is 'peculiar'.

[Wipneus]

Rob is correct. The "re-calibration" is about the parameters of the algorithm used to calculate sea ice concentration. Extent and area are numbers derived from sea ice concentration data (and use no additional re-calibrated parameters).

The outcome of the said re-calibration was, if I understand correctly, that the same parameters can be used for f18 as for f17. No adjustment was needed after all. I did not see any update from the preliminary sea ice concentration data. The same conclusion was apparently made by OSI/SAF that switched over from f17 to f18 within two days.

[Wipneus]

Medium declines, 2016 still ahead of 2012-2015.

Update 20160626.

Extent: -43.5 (-592k vs 2015, -338k vs 2014, -746k vs 2013, -85k vs 2012)
Area: -49.4 (-484k vs 2015, -318k vs 2014, -676k vs 2013, -226k vs 2012)
 
You will find the updated graphs in the top post

Regionally extent declines are unremarkable, with Hudson continuing to be on a fast track: -26k.

Area in the Laptev made a big bump (+42k). Adjacent ESS (-28k) and Kara (-20k) kept declining though.

Regional maps are of Chukchi and East Siberian Seas. Melt is continuing near the shorelines. The "pack" is mostly still high concentration, except where last weeks polar low passed.

[Wipneus]

The animation is of the Canadian Archipelago. The ice in the Amundsen Gulf is mobilizing.

[Wipneus]

The surface melting extent did not change much. The difference between the Pacific and Atlantic sides gets even more pronounced.

[abbottisgone]

The situation in the Greenland Sea looks totally unprecedented. Can anyone comment if there's been such a state of the ice before around this date? At least from the regional charts it seems not.

If you look at the sea ice concentration comparison maps on the ASIG for June 25th, you'll see that this is quite unprecedented. In fact, it's pretty crazy what is going on on the Atlantic side of the Arctic. Only 2012 comes close, but not in the Greenland Sea.

The multi-year sea ice seems to be disappearing from the coasts of Greenland and CAA.. is that what we're all seeing?

 

[Shared Humanity]

Wipneus....

Your animation of the CAA clearly shows, for the first time this melt season IMHO, widespread deterioration of the large floes that are entering the open water in the Beaufort.

[Feeltheburn]

Which data is more reliable?  Or do people just pick the set they like best?  I was looking around and found this graph.  Is it unreliable or am I reading it wrong?  It doesn't seem to show massive melting in 2016 for the time being like everything is saying here.  I am looking at lots of different graphs, and they all seem to tell different stories.  Some indicate that ice is falling off a cliff, while others like this one seem to tell a different reality.  How do we know which graph is the "real" truth and the one we should rely on?

Thanks.

[Rob Dekker]

Which data is more reliable?  Or do people just pick the set they like best?  I was looking around and found this graph.  Is it unreliable or am I reading it wrong?  It doesn't seem to show massive melting in 2016 for the time being like everything is saying here.  I am looking at lots of different graphs, and they all seem to tell different stories.  Some indicate that ice is falling off a cliff, while others like this one seem to tell a different reality.  How do we know which graph is the "real" truth and the one we should rely on?

Thanks.

Ah. Found your graph :
http://ocean.dmi.dk/arctic/icethickness/thk.uk.php
DMI. How did Neven say it ? The gift that keeps on giving ?

Feeltheburn, that graph is about ice volume and from a model (HYCOM-CICE) that does not have a very good track record for being accurate on ice thickness, and neither does it claim to be.
If you want to look at ice volume, I suggest you look at PIOMAS. Neven gives monthly updates on its developments. The latest of which
http://neven1.typepad.com/blog/2016/06/piomas-june-2016.html
shows that at the end of May, 2016 is running at an all time record low ice volume.

Meanwhile, why did you go there (to volume) at all ?
Why not simply stay with the topic of the thread here, and quote the numbers of AMSR2 ice extent and area that Wipneus is reporting here on a daily basis ? Or any other extent and area numbers ?
Or is all of the extent and area data available still not providing enough support for your point ? And PIOMAS not yet showing it either ?

Also, I did not see anyone on this thread claim that ice is "falling of a cliff".
Why did you claim so ?

[Rob Dekker]

And may I add that your claim of "massive melting in 2016 for the time being like everything is saying here." is just indicative of your own perception.
There was no such thing in reality. At least not in June.

Not on this thread, nor on any other on this forum or the ASIB itself.
What made you say all this, Feeltheburn ?

[edit] That said, the sea ice situation in 2016 is quite serious.
If only because area and extent and snow cover and volume are still running at record levels.
That simply does not bode well for the remainder of the melting season, and if you are a Bernie guy as your avatar pretends to be, you would not be joking about that, nor trying to find far-out arguments to deny it.

[bbr2314]

And may I add that your claim of "massive melting in 2016 for the time being like everything is saying here." is just indicative of your own perception.
There was no such thing in reality. At least not in June.

Not on this thread, nor on any other on this forum or the ASIB itself.
What made you say all this, Feeltheburn ?

[edit] That said, the sea ice situation in 2016 is quite serious.
If only because area and extent and snow cover and volume are still running at record levels.
That simply does not bode well for the remainder of the melting season, and if you are a Bernie guy as your avatar pretends to be, you would not be joking about that, nor trying to find far-out arguments to deny it.

What? There has been melt everywhere! Just because the extent and area numbers haven't plummeted as quickly as other years (despite JAXA still being #1) does not mean there hasn't been melt. Do you people look at satellites? Jeez.

One month change shows a pretty catastrophic failure of the pack's integrity. The extent/area numbers hide what is happening deep inside the pack. No other year compares.

[Rob Dekker]

What? There has been melt everywhere! Just because the extent and area numbers haven't plummeted as quickly as other years (despite JAXA still being #1) does not mean there hasn't been melt. Do you people look at satellites? Jeez.

One month change shows a pretty catastrophic failure of the pack's integrity. The extent/area numbers hide what is happening deep inside the pack. No other year compares.

bbr, I don't dispute your assertions. Bottom melt has been happening all through the Arctic, and the Beaufort specifically. And I've been saying all along that the ice that is pushed into the Beaufort is melting out and pulling the area of the CAB it came from to shreds.

I'm just disputing the claim by Feeltheburn that "everything" here suggests "massive melting", because it doesn't and it didn't. Extent and area stalled in early June, and this thread specifically reports about that on a daily basis.

Yet here we are, end of June, still record low extent and area, and with a CAB that is showing serious amounts of leads, which, combined with already weakened state after a record warm winter, spells trouble for the remainder of the season.

That's why I took offense of Feeltheburn's statements.

[seaicesailor]

Which data is more reliable?

Do your research

[oren]

There was no such thing in reality. At least not in June.

What? There has been melt everywhere!

Rob meant there was no such claim being made, not that there was no such melting... 
Chill off all. The ice will decide for itself at the end.

[Wipneus]

The declines in extent and area are enough to keep the leading position. 2012 and 2014 are slowly closing the gap though.

Update 20160627.

Extent: -79.1 (-602k vs 2015, -287k vs 2014, -726k vs 2013, -73k vs 2012)
Area: -100.2 (-528k vs 2015, -263k vs 2014, -664k vs 2013, -204k vs 2012)
 
You will find the updated graphs in the top post

Extent declined most in the Hudson (-38k). Greenland Sea (-23k) and Baffin (-16k) make up the rest.

The same regions are declining in area (-42k, -42k and -15k). Beaufort increased by +18k.

The delta map is from Baffin. The ice in the south, near Labrador, is going fast away.

[Wipneus]

Another animation of the Beaufort shows more dancing ice floes.

[Wipneus]

The surface melt extent declined a bit. At the same time more melt extent can be seen on the Atlantic side, north of Svalbard.

[slow wing]

Is the Big Block of the Beaufort finally turning dark?


Question about that Beaufort gif: which of these three are more true:
 
  a) when compared to all previous years in the satellite record, the state of the ice as shown is a real mess for around the summer solstice; or

  b) it's the enhanced contrast Wipneus uses that makes it look worse than it is; or

  c) other years have been like this but the graphics available to us hasn't been as good as this year so we haven't appreciated that.

[Neven]

I'd go for a).

[Mikko]

I agree with Neven. It's a). In 2015 we saw this kind of floe system at Beaufort but it just didn't melt. Earlier there hasn't been anything much like this. Now the heavy floes both go dark and disintegrate.  At Chukchi or even ESS we don't now see these heavy floes in the system like at all. There is crazy little hard ice remaining. And it's still June. Maybe we won't see significant extent record this year but ...

[seaicesailor]

In 2015, it did melt. Of course, depends on the extension considered, but in practice ...

[jdallen]

I agree with Neven. It's a). In 2015 we saw this kind of floe system at Beaufort but it just didn't melt. Earlier there hasn't been anything much like this. Now the heavy floes both go dark and disintegrate.  At Chukchi or even ESS we don't now see these heavy floes in the system like at all. There is crazy little hard ice remaining. And it's still June. Maybe we won't see significant extent record this year but ...

Watch  for weather driving those floes south towards the coast.  Seawater temps there are just absurd - over 2C as you get to the coast, 0-2C as far as a couple hundred KM from shore.  We get a good blow that pushes ice that way, that heat will tear into that ice at more than 10CM/day; it's the flip side of the ice "killing ground" that's appeared north of Svalbard.

[Darvince]

Per OSPO NOAA, the water temperatures in the open water of the Beaufort are even higher than shown on that, with some areas as high as 8C due to influx from the Mackenzie River:

http://www.ospo.noaa.gov/data/sst/contour/global.c.gif

[Wipneus]

Slower than usual declines make 2016 loose the lead over 2012 in extent, the lead in area is nearly halved.

Update 20160628.

Extent: -63.7 (-562k vs 2015, -184k vs 2014, -685k vs 2013, +60k vs 2012)
Area: -12.1 (-477k vs 2015, -169k vs 2014, -558k vs 2013, -115k vs 2012)
 
You will find the updated graphs in the top post

The small decline in area is partly explained by increases in CAA and Hudson regions (+27k and +26k). The ESS dropped -20k in area, other regions has smaller drops.

By extent Baffin dropped -18k, other regions has smaller drops.

So signs of melting are back on the Eurasian side of the arctic. The regional delta map of today is from the ESS.

[Wipneus]

And the animation is from Kara that is declining steadily, within the range set by recent years.

[Michael Hauber]

Is the Big Block of the Beaufort finally turning dark?


Question about that Beaufort gif: which of these three are more true:
 
  a) when compared to all previous years in the satellite record, the state of the ice as shown is a real mess for around the summer solstice; or

  b) it's the enhanced contrast Wipneus uses that makes it look worse than it is; or

  c) other years have been like this but the graphics available to us hasn't been as good as this year so we haven't appreciated that.

c) can't be ruled out.  Compare 2008 to 2016 on this day via ADS  The chart for area on the first page of this thread shows that 2016 started well ahead of all other recent years (since 2012) in Beaufort, but other years have been catching up since late May.  2012 briefly caught up a couple weeks ago, has fallen behind, and current trends suggest a strong chance of 2012 catching up again quite soon.  Although the last I looked at forecasts I thought that Beaufort might be about to speed up again which would skip ahead of where 2012 was.

[Wipneus]

And the surface melt situation: no net change.

[Wipneus]

With a slower than usual extent drop 2016 has lost most of the lead to 2014. Area sees a century, keeping 2016 ahead in this "race".

Update 20160629.

Extent: -47.5 (-556k vs 2015, -35k vs 2014, -642k vs 2013, +75k vs 2012)
Area: -120.8 (-589k vs 2015, -171k vs 2014, -622k vs 2013, -177k vs 2012)
 
You will find the updated graphs in the top post

Regional extent declined most in Baffin and Beaufort, both -16k.

Regional area in the Hudson recovered from yesterday's uptick by -40k. Baffin, Beaufort, Chukchi and CAB declined about -23k. Laptev area increased +29k.

The regional delta map is of the Beaufort that saw a big loss in sea ice concentration.

[Wipneus]

The animation is of the Laptev Sea. Area and extent are now definitively higher for this day than in any other year in my collection. It is working on it though.

[Wipneus]

And the surface melting situation, which did not change much.

[Darvince]

Does the sea ice melt extent take into account heavier melt or does it regard all amounts of melt as the same?

[Wipneus]

Does the sea ice melt extent take into account heavier melt or does it regard all amounts of melt as the same?

The latter, in the same spirit as ice extent is calculated (where everything above 15% is counted as 100% ice).