IJIS

[Espen]

IJIS:

11,262,361 km²（May 15, 2016）

[abbottisgone]

The world has been saved... happy face, happy face 

[werther]

Ah, in the line of forecast. Bering and Okhotsk almost depleted. Only Labrador Sea/Baffin Bay producing losses. Weather will have to produce a few 'tricks' to include Kara-, Chukchi Sea and the Svalbard region again.

[jdallen]

Ah, in the line of forecast. Bering and Okhotsk almost depleted. Only Labrador Sea/Baffin Bay producing losses. Weather will have to produce a few 'tricks' to include Kara-, Chukchi Sea and the Svalbard region again.

I think those tricks may be around, werther.

Looking at the latest Worldview maps, ice all along the Atlantic edge is in terrible shape.  As well as having the edge pushed back, it looks like warmer surface water is being blown in all across it.  The ice in the Greenland sea has retreated hundreds of miles north and is disintegrating rapidly.  I've included a small detail from along the Greenland coast about midway between Denmark strait and the Fram.

The edge looks like that from near Iceland all the way around to the western Kara.

[DavidR]

The world has been saved... happy face, happy face 

That  was just  drawing breath, as the extent of the record passed the 600K Km^2 mark  and 4 std devs from the IJIS mean.

The previous record low for this day was 11867281,  604920 above todays mark,
The record prior to this year for std deviation was 2.41 std deviations,  for any day of the year.

[Shared Humanity]

Ah, in the line of forecast. Bering and Okhotsk almost depleted. Only Labrador Sea/Baffin Bay producing losses. Weather will have to produce a few 'tricks' to include Kara-, Chukchi Sea and the Svalbard region again.

I think those tricks may be around, werther.

Looking at the latest Worldview maps, ice all along the Atlantic edge is in terrible shape.  As well as having the edge pushed back, it looks like warmer surface water is being blown in all across it.  The ice in the Greenland sea has retreated hundreds of miles north and is disintegrating rapidly.  I've included a small detail from along the Greenland coast about midway between Denmark strait and the Fram.

The edge looks like that from near Iceland all the way around to the western Kara.

What is really needed to drop SIA dramatically is for huge amounts of ice to be pushed south through the Fram so that Greenland SIA and SIE is maintained or actually begins to grow even in the face of those warm waters which are melting the ice rapidly. If we see Greenland Sea ice stabilize or grow slowly, we will know we are screwed.

[Espen]

IJIS:

11,197,909 km²（May 16, 2016）

[werther]

In three days under eleven? Possible. A 12 day lead...

[Pmt111500]

That  was just  drawing breath, as the extent of the record passed the 600K Km^2 mark  and 4 std devs from the IJIS mean.

The previous record low for this day was 11867281,  604920 above todays mark,
The record prior to this year for std deviation was 2.41 std deviations,  for any day of the year.

+-4 sigma is pretty far out, let alone -4 sigma. It could be nice to see the whole dataset calculated as sigma-values against a common baseline and another plot calculated without the knowledge of future values (as the data comes in). The latter one should likely be calculated against a short (say 10-year) individual baseline for each data point. Not sure if this would be useful in identifying future tipping points elsewhere on the planet, and likely something like this has been tried previously, but that doesn't mean these wouldn't be nice to see. Maybe I should power up the computer and calculate these myself. The first one should look pretty much like the anomaly-graph of course but the other one might be interesting. Without any priors, the first ten years (as the ten year baseline develops(this could of course be discarded alltogether) should see a wave-pattern but what happens after this?

[DavidR]

The attached graph  shows the IJIS daily values for each fortnight over the next two months.

There are three things I find interesting.

1. The expected increasing decline in the trend lines.
2. The fact that the years that start above the trend line tend to stay there and vice versa
3. The further you are from the trend line early the further you get.

2012 and 2015 were the only years that significantly varied from the pattern.  I  suspect the cause of the 2015 variation is evident in the DMI temperature graph  for May.

If  I  add 2016 to this graph it drags the May 16th trend line down to  be parallel with June 1st and sits on the trend line for June 1st.

[DavidR]

That  was just  drawing breath, as the extent of the record passed the 600K Km^2 mark  and 4 std devs from the IJIS mean.

The previous record low for this day was 11867281,  604920 above todays mark,
The record prior to this year for std deviation was 2.41 std deviations,  for any day of the year.

+-4 sigma is pretty far out, let alone -4 sigma. It could be nice to see the whole dataset calculated as sigma-values against a common baseline and another plot calculated without the knowledge of future values (as the data comes in). The latter one should likely be calculated against a short (say 10-year) individual baseline for each data point. Not sure if this would be useful in identifying future tipping points elsewhere on the planet, and likely something like this has been tried previously, but that doesn't mean these wouldn't be nice to see. Maybe I should power up the computer and calculate these myself. The first one should look pretty much like the anomaly-graph of course but the other one might be interesting. Without any priors, the first ten years (as the ten year baseline develops(this could of course be discarded alltogether) should see a wave-pattern but what happens after this?

Are you thinking  of something  like this which  gives 2016,  and the Maximum and the Minimum extent for each day measured in std deviations.

[abbottisgone]

The world has been saved... happy face, happy face 

That  was just  drawing breath, as the extent of the record passed the 600K Km^2 mark  and 4 std devs from the IJIS mean.

The previous record low for this day was 11867281,  604920 above todays mark,
The record prior to this year for std deviation was 2.41 std deviations,  for any day of the year.

They are some scary stats 

[Peter Ellis]

1. The expected increasing decline in the trend lines.
2. The fact that the years that start above the trend line tend to stay there and vice versa
3. The further you are from the trend line early the further you get.

I don't believe your (3).  Have you tried (e.g.) plotting the residuals from the May 16th trend against the residuals from the July 16th trend?

[DavidR]

3. The further you are from the trend line early the further you get.

I don't believe your (3).  Have you tried (e.g.) plotting the residuals from the May 16th trend against the residuals from the July 16th trend?

No I admit to  just eyeballing it.  However I  might do  so when I  have time.

[OldLeatherneck]

The attached graph  shows the IJIS daily values for each fortnight over the next two months.

There are three things I find interesting.

1. The expected increasing decline in the trend lines.
2. The fact that the years that start above the trend line tend to stay there and vice versa
3. The further you are from the trend line early the further you get.

2012 and 2015 were the only years that significantly varied from the pattern.  I  suspect the cause of the 2015 variation is evident in the DMI temperature graph  for May.

If  I  add 2016 to this graph it drags the May 16th trend line down to  be parallel with June 1st and sits on the trend line for June 1st.

GOOD work on that chart!  Everyone has a different way of plotting and showing data.  Each method provides additional pieces to a complex puzzle.  Which then leads to more questions.

[Pmt111500]

That  was just  drawing breath, as the extent of the record passed the 600K Km^2 mark  and 4 std devs from the IJIS mean.

The previous record low for this day was 11867281,  604920 above todays mark,
The record prior to this year for std deviation was 2.41 std deviations,  for any day of the year.

+-4 sigma is pretty far out, let alone -4 sigma. It could be nice to see the whole dataset calculated as sigma-values against a common baseline and another plot calculated without the knowledge of future values (as the data comes in). The latter one should likely be calculated against a short (say 10-year) individual baseline for each data point. Not sure if this would be useful in identifying future tipping points elsewhere on the planet, and likely something like this has been tried previously, but that doesn't mean these wouldn't be nice to see. Maybe I should power up the computer and calculate these myself. The first one should look pretty much like the anomaly-graph of course but the other one might be interesting. Without any priors, the first ten years (as the ten year baseline develops(this could of course be discarded alltogether) should see a wave-pattern but what happens after this?

Are you thinking  of something  like this which  gives 2016,  and the Maximum and the Minimum extent for each day measured in std deviations.

Thanks for this. This is not quite quite what I meant and I was thinking of Cryosphere Today data and this is IJIS thread but that is quite enough.  . So striking anomaly I'm tempted to steal the image and post it on my fb-feed.

[DavidR]

Thanks for this. This is not quite quite what I meant and I was thinking of Cryosphere Today data and this is IJIS thread but that is quite enough.  . So striking anomaly I'm tempted to steal the image and post it on my fb-feed.

Knock  your self out!

[Pmt111500]

Thanks, done, with the explanation that negative sign is omitted, the image could be clearer upside down and max (green) should be on the other side of zero... 

[Espen]

IJIS:

11,144,719 km²（May 17, 2016）

[BornFromTheVoid]

After yesterdays drop, if we followed the extent loss from now til minimum of any of the last 11 years, we'd be at least 2nd lowest on record.

[Jim Pettit]

IJIS:

11,144,719 km²（May 17, 2016）

If IJIS extent were to flatline for the next two weeks and not drop by so much as another single square kilometer, 2016 would *still* be in second place (behind 2015) at the end of the May. Or stated another way, 2016 is now running *more* than two weeks ahead of the normal.

Ouch...

[OldLeatherneck]

IJIS:

11,144,719 km²（May 17, 2016）

If IJIS extent were to flatline for the next two weeks and not drop by so much as another single square kilometer, 2016 would *still* be in second place (behind 2015) at the end of the May. Or stated another way, 2016 is now running *more* than two weeks ahead of the normal.

Ouch...

What is even more scary to me is the rate we have been losing ice so far this year.  The IJIS loss from March 20th until now is slightly more than 25% above the 2003-2015 average.  If 2016 were to have continuous 25% above average losses from March 20th through September 10th, the IJIS Extent would be 2,146,554 Km².

Given that we have had just a dramatic head start this year, we don't need to maintain that rate of loss to still have a chance for a new record.  If losses were to be 10% above average from today  through September 10th, the IJIS Extent would be 3,170,533 Km². Average losses from today through September 10th would result would be 3,900,311 Km².

I beginning feel that 2016 will end the year at least in second place and fearing that it could be much lower than that.

[DoomInTheUK]

Two weeks ahead in June/July would be bad enough, but two weeks in mid May leaves a scary amount of open water just waiting for that peak insolation.

[Meirion]

Just run the CCI forecast for the next week http://cci-reanalyzer.org/Forecasts/#NH-SAT & tell me it doesn't look like another 500K gone if forecast holds. 10,650,000 by May 25th anyone? 10.5 by end of May now looks extremely likely.

[Buddy]

Extent would be 2,146,554 Km2.

I think that is LIKELLY.  The 2 mill"ish" area.....2 - 2.5 I would give an 85% chance of that happening.  Two mill is definitely NOT out of the question this year.

[Espen]

IJIS:

11,104,253 km²（May 18, 2016）

[gregcharles]

Extent would be 2,146,554 Km2.

I think that is LIKELLY.  The 2 mill"ish" area.....2 - 2.5 I would give an 85% chance of that happening.  Two mill is definitely NOT out of the question this year.

Really ... 85%? What figure would you call an even money bet then? How about 2.25? I'm as pessimistic as the next guy, but I'd be willing to take the over on that, with stakes of, say, a six pack. Oregon vs. San Diego ... Deschutes, maybe, or Breakside, against Ballast Point or Stone. It would ease the pain of the imminent collapse of our ecosystem.

[werther]

Not in three days. Like I thought, Bering and Okhotsk depleted. Waiting to see whether the Kara Sea will produce some firmer losses.
Anyway, tempo will get a bit less extreme.

[abbottisgone]

It's certainly looking healthier.

[PhilDPortsmouth]

Hi Espen,
Thank you for your daily post. I think you used to give the figure and then the amount of the drop. This was really useful as at the moment you have to go back to previous day - and this can be a long way back. Is it an easy thing to do?
Many thanks
Phil

[6roucho]

It's certainly looking healthier.

It is?

[abbottisgone]

It's certainly looking healthier.

It is?

You don't think it could be worse, perhaps? 

[6roucho]

It's certainly looking healthier.

It is?

You don't think it could be worse, perhaps? 

It seems to be more of the same. The trend remains the same from day to day within the error margins of our measuring systems. We're slowly but surely dropping below previous records on most variables. It will take a series of slow days (or some spectacular losses) to meaningfully change that.

[oren]

Hi Espen,
Thank you for your daily post. I think you used to give the figure and then the amount of the drop. This was really useful as at the moment you have to go back to previous day - and this can be a long way back. Is it an easy thing to do?
Many thanks
Phil

If people here would be kind enough to move commentary to the melting season thread, all would be ok.

[DavidR]

Hi Espen,
Thank you for your daily post. I think you used to give the figure and then the amount of the drop. This was really useful as at the moment you have to go back to previous day - and this can be a long way back. Is it an easy thing to do?
Many thanks
Phil

Phil,
If you want more detail you  can get  it here:
https://ads.nipr.ac.jp/vishop/data/graph/plot_extent_n_v2.csv

[Buddy]

Really ... 85%? What figure would you call an even money bet then? How about 2.25? I'm as pessimistic as the next guy, but I'd be willing to take the over on that, with stakes of, say, a six pack. Oregon vs. San Diego ... Deschutes, maybe, or Breakside, against Ballast Point or Stone. It would ease the pain of the imminent collapse of our ecosystem.

1)  Yes.  85%.  Really.
2)  An even money bet....likely around 2.25 or 2.35
3)  If you want to wager a 6 pack of Deschutes beer (yum)....I would be MORE THAN HAPPY to bet you that the ice extent will be equal to or below 2.5 million square kilometers.  I don't bet on "even money bets".....that is crap shoot.

This isn't about optimism or pessimism......this is about physics...AND....the likelihood of those physics playing out over the FUTURE 4 months.

A)  We have had 12 months in a ROW of record setting monthly temperature data.  And the last few HAVEN'T EVEN BEEN CLOSE.  Off the charts sort of stuff.   Yes...SOME of it due to El Nino...but not all.
B)  Ocean waters CONTINUE to absorb heat....but as it absorbs heat AND CO2....it loses its ability to store MORE heat over time.
C)  Greenland ice sheet has already had two "melting events" this year.
D)  Ice has broken away from the shoreline in some places VERY EARLY in the year.
E)  Beaufort is in trouble.  In 2012 it had a "big fat tail" of ice extent that lasted through the melting season (southern Arctic Ocean/northern Beaufort Sea.  It is likely TOAST this year.
F)  The 2016 melt season is now tracking parallel with the 2012 season.....only it has a 2 - 4 week head start.  The longer it maintains that "head start"....the more time that dark ocean will have to absorb heat and melt ice when the real fun begins in June and July.  I expect the "plunge" that occurred in mid June during the 2012 season....to start early this year.  Perhaps late May or early June.

I'm not a pessimist.  I'm objective.  I see the numbers and the basic physics....and I connect the dots.  It's not what I WANT to happen.  It's what the numbers/graphs/basic physics are telling me is LIKELY to happen.

A Duck walks into a bar and tells the bartender:  "Bartender....give me a beer, and put it on my "bill".  Oh....and make it a sic pack of Deschutes."......  

[Buddy]

Two other things I forgot to add....

1)  Temperatures in the Arctic shot term remain "hot".  The Arctic is running about 3-4 degrees warmer than normal over the last couple of weeks....and forecast to remain that way.

2)  The temps in the Arctic have been "off the charts" this year.  If you look at the DMI data for temperatures above 80 degrees north latitude this year.....and compare them with 2012....you'll see a BIG DIFFERNCE.  The DMI temp STILL hasn't crossed under the "average temperature"....its remained warmer than normal.  If you look at about the 15th graphic down the page on the following link.....you will find the 2016 - 2012 DMI graphs......and you'll notice 2016 vs the other years....especially 2012.

https://www.blogger.com/blogger.g?blogID=3668502585335462792#editor/target=post;postID=3138607357704998522;onPublishedMenu=allposts;onClosedMenu=allposts;postNum=11;src=link

[OldLeatherneck]

Is There Really a "May Pause" in the Rate of Ice Loss??

I don't know if others have noticed this phenomena before, but it appears to me the rate of extent loss seems to peak in late April and remains on a plateau through early June. Below are a series of charts that I put together trying to prove or disprove whether or not there is a plateau.  I will briefly describe each chart.

Chart 1:
It was when I was calculating the the average daily losses for the 10 day period I use for future loss projections. 



Chart 2:
It appears that average daily losses are ~50K Km² for the period  spanning late April through early June.



Chart 3:
Here is where things started getting more complicated.  I knew if I was going to compare the daily losses for each year from 2003 through 2015, I would need to do some smoothing since using using day-by-day values would be too noisy.  I did 3-day, 5-day and 7-day rolling averages.  The first plot below is the 7-day rolling average of daily extent losses from March 15 through September for each 15th for years 2003-2015.  A plateau is clearly discernible even in this noisy plot.  The second plot is nothing more than the averages of the data in the first graph, where the plateau is very evident.


SUMMARY
Having spent much of my life analyzing performance data from shipboard radars, I'm  comfortable in stating that there is a pause in the daily rate of extent loss during May.  However, I've never been a gifted statistician or computer programmer.  Therefore I welcome any comments that may contradict or validate my assumptions.

Has this pause been noted before? If not, should we determine why??

[crandles]

I would suggest a first thought might be, if there was no land the area of ice might be roughly circular most of the time. With lots of circumference to attack there might not be a pause. Whereas with land hemming in the ice such that there is little that is easily attacked at the time of the pause, a pause might result.

[jdallen]

<snippage>
SUMMARY
Having spent much of my life analyzing performance data from shipboard radars, I'm  comfortable in stating that there is a pause in the daily rate of extent loss during May.  However, I've never been a gifted statistician or computer programmer.  Therefore I welcome any comments that may contradict or validate my assumptions.

Has this pause been noted before? If not, should we determine why??

The data you present pretty much supports your conclusion.

Shooting from the hip, my impression is that the pause is primarily mechanical rather than climatological.  I believe it reflects a number of factors:

- Diminishing ice earlier in the season in the peripheral seas
- Lag at higher latitudes as they accumulate necessary heat required to start reducing extent

In short, once the lower latitude ice has dropped out, it takes a while for remaining ice which is better buffered from ocean heat by land masses to reach a point where melt starts in earnest. 

Pretty much I see it as the "flip side" of the refreeze, once major portions of the CAB and peripheral seas have iced up.  The areas in question reach a balance between heat loss/gain and ice coverage.

[sedziobs]

Has this pause been noted before? If not, should we determine why??

Thanks for the clear presentation.  As for the reason, I think it can be broken down by regions (as jdallen also points out).  The following observations are taken from Wipneus' regional extent graphs:

In March and April, Okhotsk and Baffin (and Bering in recent years) begin to build the melt rate.  In May, other regions have not started to melt significantly, so the rate stays the same.  By mid-June, Hudson and Kara begin to melt rapidly, along with a considerable acceleration in Baffin and steady melt in Chukchi.  Other regions start to melt in earnest in July.

If we are to remain weeks ahead of other years, then I think the typical June melt regions need to get started quickly.  I think Kara and Chukchi are candidates, as they have already seen warm temperatures and positive anomalies persist through the GFS forecast.  Hudson hasn't seen the warmth of other regions, and the forecast is mixed aside from a warm surge on May 23-24.  So I think 2016 extent will also see a "May Pause" before Hudson starts its plunge.

[Nick_Naylor]

The May slowdown appears to be an extent thing - it's not apparent in area or volume.

I also suspect it's more correlated to the level of extent than it is to the time of year,  so it looks to be getting earlier over the years.

My first impression is that once open water begins to appear in the periphery of any given body of water, fragmentation begins and each cell of an extent analysis begins falling from 100% to 15% coverage while the ice begins moving and filling in the gaps.

[magnamentis]

all that and i think the locaction has a significant impact. i mean peripheral ice is at much lower latitudes than the almost enclosed environment of the arctic basin and then on the periphery, where we have open ocean, there are more waves and more undisturbed wind patterns, not to forget the huge currents that transport a lot of heat into the peripherals while the central arctic is hardly reached by those.

last but not least the thickness and temps of the ice is much lower in peripheral areas.

[OldLeatherneck]

Thanks to everyone for their valuable insights into the multiple reasons that the rate of SIE  can and will quite often reach a stable level this time  of year.  Everyone was polite, stayed on topic and didn't ruffle any feathers!!

We can all learn so much more when the atmosphere of communication is collegial.

[theoldinsane]

Thanks to everyone for their valuable insights into the multiple reasons that the rate of SIE  can and will quite often reach a stable level this time  of year.  Everyone was polite, stayed on topic and didn't ruffle any feathers!!

We can all learn so much more when the atmosphere of communication is collegial.

+1

[Espen]

IJIS:

11,041,226 km²（May 19, 2016）

[BornFromTheVoid]

[OldLeatherneck]

This will be my last posting of charts until I return from Europe in the middle of June.  However, I may be jumping in with comments from time to time.

The first chart indicates that 2016 should still be well below 2016 by the 15th of June.  However, for it to stay near 1M below 2012 will need  to maintain daily losses between 60-70K.

The second chart shows that 2016 has consistently had losses about 25% below the long term averages. If 2016 continues losses at this rate, it will be flirting with 2012's record low.

[Buddy]

The first chart indicates that 2016 should still be well below 2016 by the 15th of June.  However, for it to stay near 1M below 2016

You mean...."still be well below 2012"

And...."stay near 1M below 2012"

PS...have a GREAT trip and thanks for all your charts...much appreciated.

[OldLeatherneck]

The first chart indicates that 2016 should still be well below 2016 by the 15th of June.  However, for it to stay near 1M below 2016

You mean...."still be well below 2012"

And...."stay near 1M below 2012"

Corrected!!...thanks for noticing!!

PS...have a GREAT trip and thanks for all your charts...much appreciated.