Support the Arctic Sea Ice Forum and Blog

Author Topic: The 2019/2020 freezing season  (Read 173822 times)

Glen Koehler

  • Frazil ice
  • Posts: 350
    • View Profile
  • Liked: 337
  • Likes Given: 757
Re: The 2019/2020 freezing season
« Reply #800 on: January 28, 2020, 05:37:41 PM »
I've done my due diligence this morning (UTC), and there's still a total absence of any ice >= 4m thick to be seen on the most recent CS2/SMOS reanalysis. Chapter and verse over at:
http://GreatWhiteCon.info/2020/01/wheres-the-thickest-arctic-sea-ice-gone/

Thanks Jim.  The  Jan 11, 2020 vs Jan 11, 2019 contrast in the Great White Con images is much less dramatic than the Jan 19, 2020 vs 2019 images posted upthread.  Ditto, the PIOMAS difference maps posted by Oren are not as dramatic either.  I suspect there is an issue with color scaling in the Jan 19, 2020 CryoSat image upthread.  So I consider that issue resolved or at least unimportant.

   What is important is the lack of >4m thick ice and overall thinner and less consistent thickness shown in your January 2020 Great White Con maps.  Not only is the thickest ice gone, but the ASI as whole looks weaker and more variable.  In particular, the Atlantic side from the North Pole down to Greenland and Fram Starit looks more vulnerable than same dates in 2019 and 2018.

   The 2020 melt season seems to have potential for dramatic acceleration relative to the long term trend.  In simpler words: it looks like it could be on the edge of falling apart by September 2020.

The Walrus

  • Grease ice
  • Posts: 677
    • View Profile
  • Liked: 94
  • Likes Given: 42
Re: The 2019/2020 freezing season
« Reply #801 on: January 28, 2020, 08:27:21 PM »
Possibly.  However, the sea ice extent this year is greater than 8 of the past 10 years.  The only two years with greater extent were 2013 & 14, which also had the highest minima over the past decade.

Jim Hunt

  • Young ice
  • Posts: 4742
  • Stay Home, Save Lives
    • View Profile
    • The Arctic sea ice Great White Con
  • Liked: 507
  • Likes Given: 44
Re: The 2019/2020 freezing season
« Reply #802 on: January 28, 2020, 09:37:55 PM »
However, the sea ice extent this year is greater than 8 of the past 10 years.

So sea ice thickness is irrelevant to your potential prognostications? Here's another sea ice motion video for you:



Guess what happened next?
Reality is merely an illusion, albeit a very persistent one - Albert Einstein

kassy

  • Nilas ice
  • Posts: 2447
    • View Profile
  • Liked: 1168
  • Likes Given: 1016
Re: The 2019/2020 freezing season
« Reply #803 on: January 28, 2020, 10:09:23 PM »
It is amazing to see. Especially if you remember what the age/thickness charts looked like in the early noughties.

It would have been nice to have a month year counter pasted over some continent.
Þetta minnismerki er til vitnis um að við vitum hvað er að gerast og hvað þarf að gera. Aðeins þú veist hvort við gerðum eitthvað.

Jim Hunt

  • Young ice
  • Posts: 4742
  • Stay Home, Save Lives
    • View Profile
    • The Arctic sea ice Great White Con
  • Liked: 507
  • Likes Given: 44
Re: The 2019/2020 freezing season
« Reply #804 on: January 28, 2020, 10:15:57 PM »
It would have been nice to have a month year counter pasted over some continent.

Isn't the year/week counter at the bottom right sufficient?
Reality is merely an illusion, albeit a very persistent one - Albert Einstein

Aleph_Null

  • Frazil ice
  • Posts: 157
    • View Profile
  • Liked: 48
  • Likes Given: 253
Re: The 2019/2020 freezing season
« Reply #805 on: January 28, 2020, 10:22:40 PM »
Excellent global summary posted by Bob Henson yesterday:

"from January 1 to 27, the AO has averaged above +2.0. According to NOAA, only four Januarys since 1950 have seen a positive AO this strong, including 1993 (+3.495), 1989 (+3.106), 1957 (+2.062), and 2007 (+2.034). Even if the AO weakens in the next few days, this month could end up with the third highest January value in 71 years of recordkeeping."

https://www.wunderground.com/cat6/Polar-Vortex-Keeps-Cold-Bottled-How-Long-Will-It-Last?cm_ven=cat6-widget

The Walrus

  • Grease ice
  • Posts: 677
    • View Profile
  • Liked: 94
  • Likes Given: 42
Re: The 2019/2020 freezing season
« Reply #806 on: January 28, 2020, 10:48:39 PM »
However, the sea ice extent this year is greater than 8 of the past 10 years.

So sea ice thickness is irrelevant to your potential prognostications?

No.  Just less relevant.  Neither yields a perfect correlation, but extent does better.

Glen Koehler

  • Frazil ice
  • Posts: 350
    • View Profile
  • Liked: 337
  • Likes Given: 757
Re: The 2019/2020 freezing season
« Reply #807 on: January 28, 2020, 11:55:55 PM »
No.  Just less relevant.  Neither yields a perfect correlation, but extent does better.
Correlation of what with what?
Volume is the bottom line measure of how much ice there is. 
Thickness indicates a key component of the Volume calculation (Thickness x Extent or Thickness x Area).
  Also, it provides qualitative information about the condition of that ice.  Older thicker ice is more resistant to melt than younger, thinner ice.

Extent and Area are more directly related to the albedo impact of ASI decline.

So each measure has value for addressing different issues / answering different questions.

Glen Koehler

  • Frazil ice
  • Posts: 350
    • View Profile
  • Liked: 337
  • Likes Given: 757
Re: The 2019/2020 freezing season
« Reply #808 on: January 29, 2020, 12:12:04 AM »
    Is there an index that relates ice thickness to melt resistance?
    It would be interesting to see a calculation of the melt resistance for the total ASI melt resistance for standardized month/day dates compared across years.
    That might be a better indicator than even Volume for how much change has occurred to the ASI.  For example -- A million km3 of ice in 2020 might be more vulnerable to melt (require less melting energy) than a million km3 of ASI in 2010.

    In addition to thickness, other variables like continuity, average floe size and variability, and salinity could go into it.  That all seems intractably difficult, esp. since values of those characteristics for the Arctic ice pack are probably not available. But a measure of the average thickness for the ASI ice pack combined with some reasonably accurate method to account for the relative melt resistance for sea ice of different thicknesses could be enlightening.  (Though an Arctic wide melt resistance value would need to account for volume within each thickness category separately, so average ASI thickness would not suffice).

    But first things first.  Is there some formula for translating the thickness of a single chunk of ASI ice into a relative or absolute melt resistance value?   I postulated such a thing by reversing the Thorndike ice growth rate curve for different ice thicknesses a few months ago and posted it on ASIF, but was informed by ASIF contributors who actually know about these things that a simple reversal of the energy in/out equation was not valid.  Thorndike defined an ice growth vs. thickness curve, so it seems like Thorndike or somebody must have done the opposite, define a ice melt vs. thickness curve (for a fixed melting energy input).  That would give us another handle for understanding the significance of the loss of thickest oldest ice, and the overall decline in average ASI thickness.
 
« Last Edit: January 29, 2020, 12:31:21 AM by Glen Koehler »

Niall Dollard

  • Grease ice
  • Posts: 757
    • View Profile
  • Liked: 292
  • Likes Given: 70
Re: The 2019/2020 freezing season
« Reply #809 on: January 29, 2020, 01:52:41 AM »
Possibly.  However, the sea ice extent this year is greater than 8 of the past 10 years.  The only two years with greater extent were 2013 & 14, which also had the highest minima over the past decade.

Looking at extent figures now (Jan 27th) is not much guarantee of a high minimum come September.

Example 1 : 2007 . The joint second lowest min on record. Yet extent on Jan 27th 2007 was higher than it is currently in 2020. A lot can happen.

Example 2 : 2012. The lowest year of all. Jaxa extent on Jan 27th 2012 was only 115 k less than extent is currently

Re the ice thickness  charts, I am not that confident in what they show. Some are better than others.

Last year over on the Nares Strait thread, we were saying that the ice north of Greenland in the Lincoln Sea must be thin as it kept fragmenting and there was no arch in the strait. Yet the 2019 chart above appears to show thicker ice above Greenland than this year.

Instead last December the arch did form. Was the ice than thicker this winter contrary to what the charts indicate.

Or is arch formation more to do with tides, currents and surface winds coming together at the right time rather than how thick the ice is ?   

Paul

  • Frazil ice
  • Posts: 216
    • View Profile
  • Liked: 44
  • Likes Given: 3
Re: The 2019/2020 freezing season
« Reply #810 on: January 29, 2020, 02:43:31 AM »
I think some are forgetting volume in January 2019 will probably be higher than 2020 because the volume during the latter parts of 2018 increased quite steady. It was during February 2019 the volume flatlined due to frequent warmth coming in from the Bering Sea and the spring of 2019 probably was not the best for sea ice as open water was observed very early in the Beaufort sea.

What we can say, the ice is probably not as vulnable as it was in 2017 and its too early to say what the melt season may or may not produce.

Thomas Barlow

  • Frazil ice
  • Posts: 138
    • View Profile
  • Liked: 46
  • Likes Given: 24
Re: The 2019/2020 freezing season
« Reply #811 on: January 29, 2020, 04:35:20 AM »
Well, that's new. Haven't seen anything like this for a while.
According to this, sea-ice extent has about 10 lower years for this time of year. Even 2005 and 2006 are lower.
Still among the lowest group on record, so not celebrating too much.
http://nsidc.org/arcticseaicenews/charctic-interactive-sea-ice-graph/
« Last Edit: January 29, 2020, 10:49:41 PM by Thomas Barlow »

gandul

  • Grease ice
  • Posts: 703
    • View Profile
  • Liked: 164
  • Likes Given: 10
Re: The 2019/2020 freezing season
« Reply #812 on: January 29, 2020, 08:12:54 AM »
    Is there an index that relates ice thickness to melt resistance?
    It would be interesting to see a calculation of the melt resistance for the total ASI melt resistance for standardized month/day dates compared across years.
    That might be a better indicator than even Volume for how much change has occurred to the ASI.  For example -- A million km3 of ice in 2020 might be more vulnerable to melt (require less melting energy) than a million km3 of ASI in 2010.
Of course, it is not that it requires less energy to melt, but it enables faster heat transfer. Thinner ice permits direct solar radiation to reach water beneath ice. Thin ice also breaks down more easily into smaller chunks, eventually melting out by top, bottom, lateral melting and wave washing. It is also increasingly first year ice, which is easier to melt than multi year ice.
This reduction of melt resistance is another feedback contributing to the loss of Arctic ice.

Jim Hunt

  • Young ice
  • Posts: 4742
  • Stay Home, Save Lives
    • View Profile
    • The Arctic sea ice Great White Con
  • Liked: 507
  • Likes Given: 44
Re: The 2019/2020 freezing season
« Reply #813 on: January 29, 2020, 10:00:48 AM »
Last year over on the Nares Strait thread, we were saying that the ice north of Greenland in the Lincoln Sea must be thin as it kept fragmenting and there was no arch in the strait. Yet the 2019 chart above appears to show thicker ice above Greenland than this year.

And what happened to the "thicker ice above Greenland" of which you speak after mid January 2019? According to the theory which is mine:

Quote
Over the winter of 2018/19 ASCAT revealed that there was a relentless movement of multi-year ice towards both the North Atlantic and the Beaufort Sea. Perhaps a significant amount of the multi-year ice that survived the winter of 2018/19 has now simply melted away in warm water, to be replaced by much less robust first year ice in the area between the North Pole and the Siberian coast?

I'll allow that some melted out in Baffin Bay as well.

RIP Terry Jones:



John & Eric are still with us!


Reality is merely an illusion, albeit a very persistent one - Albert Einstein

Jim Hunt

  • Young ice
  • Posts: 4742
  • Stay Home, Save Lives
    • View Profile
    • The Arctic sea ice Great White Con
  • Liked: 507
  • Likes Given: 44
Re: The 2019/2020 freezing season
« Reply #814 on: January 29, 2020, 10:26:15 AM »
Is there some formula for translating the thickness of a single chunk of ASI ice into a relative or absolute melt resistance value?

Not that I'm aware of. Melting is a much messier process than freezing.

A "Great Arctic Cyclone" in August seems to have a significant effect on Arctic sea ice melt for example!

https://neven1.typepad.com/blog/2012/08/arctic-summer-storm-open-thread.html
Reality is merely an illusion, albeit a very persistent one - Albert Einstein

The Walrus

  • Grease ice
  • Posts: 677
    • View Profile
  • Liked: 94
  • Likes Given: 42
Re: The 2019/2020 freezing season
« Reply #815 on: January 29, 2020, 06:05:42 PM »
Possibly.  However, the sea ice extent this year is greater than 8 of the past 10 years.  The only two years with greater extent were 2013 & 14, which also had the highest minima over the past decade.

Looking at extent figures now (Jan 27th) is not much guarantee of a high minimum come September.

Example 1 : 2007 . The joint second lowest min on record. Yet extent on Jan 27th 2007 was higher than it is currently in 2020. A lot can happen.

Example 2 : 2012. The lowest year of all. Jaxa extent on Jan 27th 2012 was only 115 k less than extent is currently

Re the ice thickness  charts, I am not that confident in what they show. Some are better than others.

Last year over on the Nares Strait thread, we were saying that the ice north of Greenland in the Lincoln Sea must be thin as it kept fragmenting and there was no arch in the strait. Yet the 2019 chart above appears to show thicker ice above Greenland than this year.

Instead last December the arch did form. Was the ice than thicker this winter contrary to what the charts indicate.

Or is arch formation more to do with tides, currents and surface winds coming together at the right time rather than how thick the ice is ?   

Actually, 2012 is the only real anomaly.  The average January extent in 2016 and 2019 were not much lower than 2007, and had similar minima.  The correlation is not great, but it is better than others.

Niall Dollard

  • Grease ice
  • Posts: 757
    • View Profile
  • Liked: 292
  • Likes Given: 70
Re: The 2019/2020 freezing season
« Reply #816 on: January 30, 2020, 08:48:24 PM »
XY Plot of NSIDC Average January extent versus the following September's min from year 2006 onwards.

TBH I think it is a bit of a meaningless plot because it takes no account of ice thickness and there are so many other factors that can affect ice between January and September.

Over this time period a january average of circa 13.7 was followed by a range of Sept Mins of anywhere between 3.39 and 5.05 million km2 

kassy

  • Nilas ice
  • Posts: 2447
    • View Profile
  • Liked: 1168
  • Likes Given: 1016
Re: The 2019/2020 freezing season
« Reply #817 on: January 30, 2020, 09:55:18 PM »
DUUDDD+(2012)UD-DDUED

Down up or equal.

Ds win.

I think some year those factors that saved the ice before might align the wrong way.

Also see the last dance of thick ice ase posted above by Jim Hunt.

A little more then a month and then we are back to the melting season.  :)
Þetta minnismerki er til vitnis um að við vitum hvað er að gerast og hvað þarf að gera. Aðeins þú veist hvort við gerðum eitthvað.

Glen Koehler

  • Frazil ice
  • Posts: 350
    • View Profile
  • Liked: 337
  • Likes Given: 757
Re: The 2019/2020 freezing season
« Reply #818 on: January 30, 2020, 10:35:00 PM »
XY Plot of NSIDC Average January extent versus the following September's min from year 2006 onwards.
TBH I think it is a bit of a meaningless plot

True that!  You'd be hard pressed to come up with a better example of random noise vs. a causal correlation.  Also, X drives Y so the graph should have January Extent on the X axis and following Sept. min. on the Y axis.

Nothing to see here folks, keep moving...

dnem

  • Grease ice
  • Posts: 589
    • View Profile
  • Liked: 260
  • Likes Given: 181
Re: The 2019/2020 freezing season
« Reply #819 on: January 30, 2020, 10:49:17 PM »
That hardly makes it meaningless!  It says that January extent is NOT predictive of September extent.  That's useful, no?

Is this a reasonable summary?: Strong positive AO and orderly Polar Vortex have made for "CAWCy" weather this winter. The cold arctic/warm continents has led to decent sea ice extent gains but also low snowfall across much of the Northern hemisphere.  This potentially sets the stage for an early warmup and early melt as the sun returns to the north.

I would also add that because of background warming, now, when we get a fairly "normal" winter without large breakdowns of the PV, it is now too warm to get normal snows across most of the NH.  Additionally, even with polar air bottled up at the pole it's not really THAT cold and extent gains are not extreme.

Looks like BBR might need to wait another year for his ice age.

Glen Koehler

  • Frazil ice
  • Posts: 350
    • View Profile
  • Liked: 337
  • Likes Given: 757
Re: The 2019/2020 freezing season
« Reply #820 on: January 30, 2020, 11:22:25 PM »
Agreed, I was conflating "statistical insignificance" with meaningless.  Yes, there is value in knowing that something doesn't work. 

Niall Dollard

  • Grease ice
  • Posts: 757
    • View Profile
  • Liked: 292
  • Likes Given: 70
Re: The 2019/2020 freezing season
« Reply #821 on: January 31, 2020, 12:07:48 AM »
Agreed, I was conflating "statistical insignificance" with meaningless.  Yes, there is value in knowing that something doesn't work.

Yes same here. That was my intention.

There should be no bravado about the current state of the Arctic.

The Walrus

  • Grease ice
  • Posts: 677
    • View Profile
  • Liked: 94
  • Likes Given: 42
Re: The 2019/2020 freezing season
« Reply #822 on: January 31, 2020, 02:45:19 PM »
XY Plot of NSIDC Average January extent versus the following September's min from year 2006 onwards.

TBH I think it is a bit of a meaningless plot because it takes no account of ice thickness and there are so many other factors that can affect ice between January and September.

Over this time period a january average of circa 13.7 was followed by a range of Sept Mins of anywhere between 3.39 and 5.05 million km2

Ah yes, a classic example of cherry-picking the data to arrive at your preconceived conclusion.  The question one must ask is why did you choose 2006 as a starting point?  Simple, it shows the lowest correlation.  Added just five previous years of data shows a much stronger correlation.  Ten shows even more.  Ironically,  dropped the first two years of data from your plot shows a better correlation.  Amazing how so many posters were hoodwinked by your graph!

binntho

  • Nilas ice
  • Posts: 1518
    • View Profile
  • Liked: 528
  • Likes Given: 118
Re: The 2019/2020 freezing season
« Reply #823 on: January 31, 2020, 03:02:51 PM »
Ah yes, a classic example of cherry-picking the data ...

Here's a graph covering the years 1979-1919 showing March average on Y axis vs. September average on X axis, from NSIDC.

The correlation is very strong but equally meaningless. Both winter max and summer min have trended downwards with time, and the plot says absolutely nothing more than Niall's graph further up.

I.e. it is not possible to predict with any certainty what the summer mininum will be based on the winter maximum that year.

« Last Edit: January 31, 2020, 05:08:40 PM by binntho »
because a thing is eloquently expressed it should not be taken to be as necessarily true
St. Augustine, Confessions V, 6

Glen Koehler

  • Frazil ice
  • Posts: 350
    • View Profile
  • Liked: 337
  • Likes Given: 757
Re: The 2019/2020 freezing season
« Reply #824 on: January 31, 2020, 09:30:11 PM »
binntho
  To resolve the earlier discussion, can you graph January Extent on X axis vs. September Extent on Y axis for 1979-2019?  Even better if you show not just the slope equation but the R2, F and df (all included in the output table if you are using Excel).  That would let us estimate statistical significance. 

    The March-Sept correlation is middling, but the original question was about predicting Sept. Extent from prior January, which is a lot more tenuous.  And of course, Extent is only part of the story.  (I think the main issue is ASIF folks with cabin-fever looking for something to argue about while waiting for the 2020 melt season).

     I may have over-reacted to the weak looking ice thickness map for Jan. 2020 vs. 2019 and 2018 as shown on Jim Hunt's Great White Con post. 
http://greatwhitecon.info/2020/01/wheres-the-thickest-arctic-sea-ice-gone/
The early January 2020 ice thickness map shown there may be simply be showing that the 2019-2020 freeze season got off to a late start.  A lot could change before March-April. 

     Any single year is short term noise around what really matters - the long term trend for less Arctic sea ice.  But 2020 is a particularly interesting year to watch. 

    By the long term volume trend, 2020 has about a 50:50 chance of going below the 2012 record low.  The long term trendline for Extent gives a 2020 estimate that is still roughly 20% above the 2012 record low.  The Great Arctic Cyclone caused a greater loss to Extent than it did to Volume. 

     As both values diminish, the Extent trend slope has to bend down faster to catch up to Volume - because at the end zero Volume dictates zero Extent.  The fact that thinner ice melts faster contributes to that.  An Extent maximum of 14.5 km2 in March 2020 will be more vulnerable to melt than the same Extent in earlier years when average thickness was thicker.  However, I don't think that effect has really taken hold yet, given that the March vs. September Extent correlation appears to still fit a straight line, with no suggestion of an increasing curve in the correlation that would/will become apparent when/if declining thickness causes faster Extent loss.

grixm

  • Frazil ice
  • Posts: 337
    • View Profile
  • Liked: 202
  • Likes Given: 65
Re: The 2019/2020 freezing season
« Reply #825 on: January 31, 2020, 10:39:56 PM »
Here I have plotted the normalized average extents in january, februrary and march, compared to september. Normalized meaning the value divided by the average value for the whole year. This means that the long-term trend of general melting is removed.

There seems to indeed be a strong correlation of high extent early season = low extent late season, for all three months graphed.






binntho

  • Nilas ice
  • Posts: 1518
    • View Profile
  • Liked: 528
  • Likes Given: 118
Re: The 2019/2020 freezing season
« Reply #826 on: February 01, 2020, 04:47:59 AM »
binntho
  To resolve the earlier discussion, can you graph January Extent on X axis vs. September Extent on Y axis for 1979-2019?  Even better if you show not just the slope equation but the R2, F and df (all included in the output table if you are using Excel).  That would let us estimate statistical significance. 
Sounds complicated. Besides grixm seems to have done the necessaries

Here I have plotted the normalized average extents in january, februrary and march, compared to september. Normalized meaning the value divided by the average value for the whole year. This means that the long-term trend of general melting is removed.

There seems to indeed be a strong correlation of high extent early season = low extent late season, for all three months graphed.

I for one am surprised that there is any correlation at all. And if this is real and not statistical noise, what could be the mechanism behind it?

Not only does it seem that a large extent in winter might indicate a low extent in September. Looking at the difference between the three months, if I read the graphs correctly, the longer the winter extent stays high, the stronger the correlation with low September extent.

EDIT:
The March-Sept correlation is middling

According to grixm's graphs, the March-Sept correlation is the strongest, while the Jan-Sept is the weakest of the three.
because a thing is eloquently expressed it should not be taken to be as necessarily true
St. Augustine, Confessions V, 6

mabarnes

  • New ice
  • Posts: 52
    • View Profile
  • Liked: 22
  • Likes Given: 48
Re: The 2019/2020 freezing season
« Reply #827 on: February 01, 2020, 06:30:57 AM »

I for one am surprised that there is any correlation at all. And if this is real and not statistical noise, what could be the mechanism behind it?


A possible mechanism?  The smaller the floating ice "cap" on the Arctic Ocean, the more open water is able to radiate heat to space, especially once the sun sets for a long period - six months at the pole, less further south but still substantial (112 days at Svalbard).  Perhaps the greater amount heat that "bleeds" from the ocean during this period influences the January extent?

Just takin' a shot.  Question for grixm - my slightly baked theory (microwaved?) would suggest September extent as the independent variable (x-axis) ... is it the other way around because your postulating "more in spring -> less in September" ...?  Thanks

binntho

  • Nilas ice
  • Posts: 1518
    • View Profile
  • Liked: 528
  • Likes Given: 118
Re: The 2019/2020 freezing season
« Reply #828 on: February 01, 2020, 06:48:44 AM »

I for one am surprised that there is any correlation at all. And if this is real and not statistical noise, what could be the mechanism behind it?


A possible mechanism?  The smaller the floating ice "cap" on the Arctic Ocean, the more open water is able to radiate heat to space, especially once the sun sets for a long period - six months at the pole, less further south but still substantial (112 days at Svalbard).  Perhaps the greater amount heat that "bleeds" from the ocean during this period influences the January extent?

It seems reasonable that any possible mechanism has to do with the insulating properties of ice since insolation does not enter the picture in winter. The faster the onset and larger the area of surface freezing, the more heat gets trapped. Thus bottom melt becomes a bigger factor in the following melting season.

All pure speculation. And I do not follow the logic of your last sentence. I'd think that this putative mechanism possibly causing a large winter extent to be followed by a small summer extent would precisely be caused by the ocean not bleeding the excess heat.
because a thing is eloquently expressed it should not be taken to be as necessarily true
St. Augustine, Confessions V, 6

psymmo7

  • New ice
  • Posts: 68
    • View Profile
  • Liked: 32
  • Likes Given: 14
Re: The 2019/2020 freezing season
« Reply #829 on: February 01, 2020, 09:09:33 AM »
Thx grixm for focussing the argument

grixm

  • Frazil ice
  • Posts: 337
    • View Profile
  • Liked: 202
  • Likes Given: 65
Re: The 2019/2020 freezing season
« Reply #830 on: February 01, 2020, 09:15:14 AM »
Overnight I realized that my graphs has a big flaw, the fact that the years are normalized with themselves can induce a lot of autocorrelation. For example, if it is the case that a low september minimum is completely random, that would still increase the ratio of normalized winter extent vs normalized september extent, because the low september extent would drag the average for the whole year down, thus increasing the normalized winter extent. Furthermore, since calculating the normalized extent requires knowledge of the average for the whole year, it is not possible to predict anything with it beforehand.

Therefore, I experimented with instead normalizing the years to their predicted average extent from a linear regression of all the years. Now, it should be truly neutral, and also you can make predictions. Unfortunately, doing this does reduce the correlation a lot, but it is still there.

Now, we can make  a prediction for 2020 based on the january value. The high extent compared to the ever-decreasing trend makes this year stand out a lot, the normalized january extent is an all-time high: 1.322. See the red area on the january graph. Will this mean the september extent  will be very low like the graph suggests? Or does it mean the correlation will break down? If we trust the graph naively, the expected normalized september minimum average for this year is 0.42, which is 4.30 Mkm^2 (which is third lowest of all time, behind 2012 and barely 2007), with a lower uncertainty bound of ~2.97 Mkm^2 and a high bound of ~5.33 Mkm^2.



« Last Edit: February 01, 2020, 10:37:25 AM by grixm »

binntho

  • Nilas ice
  • Posts: 1518
    • View Profile
  • Liked: 528
  • Likes Given: 118
Re: The 2019/2020 freezing season
« Reply #831 on: February 01, 2020, 10:17:25 AM »
grixm: Brilliant! Good statistics, but what's more, actual predictions to add spice to the winter dullness. Can't wait to see how February and March turn out - and of course, to match it up to the eventual September average extent.

Which brings me to a niggle: Your graph for January predicts a September average of 4.3 Mkm2 and not a September minimum of 4.3. Correct me if I am wrong! NSIDC September average for 2012 was an amazing 3.6 but both 2007 and 2019 came in at very close to 4.3.

As for the possible mechanism behind this statistically apparent correlation: The capping of excess ocean heat by an unusally rapid freeze and larger winter extent has already been mentioned. Another mechanism could be to do with weather, in two (possibly related) ways: The same winter weather that produces rapid freezing also results in stronger preconditioning come spring, or alternatively, a bigger extent at maximum increases the changes of stronger preconditioning weather in spring.
because a thing is eloquently expressed it should not be taken to be as necessarily true
St. Augustine, Confessions V, 6

uniquorn

  • Young ice
  • Posts: 2726
    • View Profile
  • Liked: 1224
  • Likes Given: 242
Re: The 2019/2020 freezing season
« Reply #832 on: February 01, 2020, 10:22:15 AM »
Higher extent could imply more export.
unihamburg-amsr2-uhh, pacific side, jan22-31

grixm

  • Frazil ice
  • Posts: 337
    • View Profile
  • Liked: 202
  • Likes Given: 65
Re: The 2019/2020 freezing season
« Reply #833 on: February 01, 2020, 10:36:06 AM »
Which brings me to a niggle: Your graph for January predicts a September average of 4.3 Mkm2 and not a September minimum of 4.3. Correct me if I am wrong! NSIDC September average for 2012 was an amazing 3.6 but both 2007 and 2019 came in at very close to 4.3.


Yes that is correct, I've edited the post

gandul

  • Grease ice
  • Posts: 703
    • View Profile
  • Liked: 164
  • Likes Given: 10
Re: The 2019/2020 freezing season
« Reply #834 on: February 01, 2020, 06:01:51 PM »

HI,
An R^2 of 0.3 or 0.4 means a 60 to 70% of variability  has to be attributed to other sources. So any prediction of summer based on this is similar to pissing around strong winds, you may think you are safe you may end up very wet.

The Forum sentiment for years is that it is worthless to project until May, and taking into account other stuff, continental snow, early area loss by ponds, extent in May, etc.

Steven

  • Grease ice
  • Posts: 633
    • View Profile
  • Liked: 198
  • Likes Given: 17
Re: The 2019/2020 freezing season
« Reply #835 on: February 01, 2020, 07:16:59 PM »
Here I have plotted the normalized average extents in january, februrary and march, compared to september. Normalized meaning the value divided by the average value for the whole year. This means that the long-term trend of general melting is removed.

Not really.  Your method is skewed by the fact that the long-term extent losses are happening more rapidly in September than in January/February/March.

September extent in the last few years is about 40% lower than in the 1980s, whereas March extent has decreased by only 10%.  So your "normalized" March extent has an upward trend over the last few decades, whereas the normalized September extent has a downward trend.  So it's not surprising that you get a negative correlation between them.  But that correlation is spurious.

A more meaningful method is to detrend the data (see e.g. here for some background on detrending).  It turns out that the correlation between the detrended March extent and the detrended September extent is very weak: the correlation coefficient is  -0.029.

edmountain

  • New ice
  • Posts: 68
    • View Profile
  • Liked: 40
  • Likes Given: 3
Re: The 2019/2020 freezing season
« Reply #836 on: February 01, 2020, 11:52:23 PM »
Here I have plotted the normalized average extents in january, februrary and march, compared to september. Normalized meaning the value divided by the average value for the whole year. This means that the long-term trend of general melting is removed.

Not really.  Your method is skewed by the fact that the long-term extent losses are happening more rapidly in September than in January/February/March.

September extent in the last few years is about 40% lower than in the 1980s, whereas March extent has decreased by only 10%.  So your "normalized" March extent has an upward trend over the last few decades, whereas the normalized September extent has a downward trend.  So it's not surprising that you get a negative correlation between them.  But that correlation is spurious.

A more meaningful method is to detrend the data (see e.g. here for some background on detrending).  It turns out that the correlation between the detrended March extent and the detrended September extent is very weak: the correlation coefficient is  -0.029.
This.

Also normalizing the data in above fashion creates a singularity as the mean extent tends towards zero. The non-linearity associated with this singularity further skews the statistics.

Pavel

  • Frazil ice
  • Posts: 247
    • View Profile
  • Liked: 40
  • Likes Given: 9
Re: The 2019/2020 freezing season
« Reply #837 on: February 03, 2020, 11:35:55 AM »
The snow cover extent is about 3 mln sq km below average. It's important because we have lack of snow in the lower latitudes and albedo is important.

blumenkraft

  • Guest
Re: The 2019/2020 freezing season
« Reply #838 on: February 03, 2020, 05:40:39 PM »
Last week's 7-day hindsight mean temperature anomalies.

blumenkraft

  • Guest
Re: The 2019/2020 freezing season
« Reply #839 on: February 03, 2020, 05:41:23 PM »
Last week's ice drift map.

blumenkraft

  • Guest
Re: The 2019/2020 freezing season
« Reply #840 on: February 03, 2020, 05:43:53 PM »
Fram export GIF is very, very bad this week due to a missing day and otherwise pretty shitty flight paths. I made it anyway for consistency's sake...

The Walrus

  • Grease ice
  • Posts: 677
    • View Profile
  • Liked: 94
  • Likes Given: 42
Re: The 2019/2020 freezing season
« Reply #841 on: February 03, 2020, 07:04:32 PM »
Now that January is in the books, we can test out the theory of September minimum based on January extent.  After all, that is how science works.  The January average this year was 13.56 sq. km.  Based on an analysis of the entire four decades of measurements, the September minimum would cone in around 4.9 sq. km.  Using only the past 20 years of data, the minimum would come in slightly lower at 4.8 sq. km.  Comparing to the more recent past, this year’s average falls between 2014 and 2015, which were 13.56 and 13.53 sq. km. respectively.  The minima for those years were 5.0 and 4.4.  The uncertainty is rather high, as the correlation is only fair, so that a likely range would be 4.2 - 5.5.

Tom_Mazanec

  • Young ice
  • Posts: 3610
    • View Profile
    • Planet Mazanec
  • Liked: 622
  • Likes Given: 307
Re: The 2019/2020 freezing season
« Reply #842 on: February 03, 2020, 07:08:51 PM »
Walrus:
You mean million sq. km.?
SHARKS (CROSSED OUT) MONGEESE (SIC) WITH FRICKIN LASER BEAMS ATTACHED TO THEIR HEADS

The Walrus

  • Grease ice
  • Posts: 677
    • View Profile
  • Liked: 94
  • Likes Given: 42
Re: The 2019/2020 freezing season
« Reply #843 on: February 03, 2020, 10:06:14 PM »
Walrus:
You mean million sq. km.?

Yes, of course.

ArcticMelt2

  • Grease ice
  • Posts: 929
    • View Profile
  • Liked: 192
  • Likes Given: 37
Re: The 2019/2020 freezing season
« Reply #844 on: February 04, 2020, 09:39:51 AM »
https://twitter.com/AlaskaWx

Quote
January was a persistently cold month across nearly all of Alaska, though only a few records were set. Much of mainland Alaska was the coldest January since 2012 but Panhandle was coldest since 2004. Lots of snow some near-coastal areas



Quote
Kodiak saw the 5th coldest January in more than a century  of records: average temperature was 21.9F (-5.6C). This is 8.6F (4.8C) below 1981-2010 normal. January in the 21st century has seen big swings, with 3 of the coldest and 2 of the warmest.



The coldest January on Kodiak is the forerunner of the big summer ice loss? Top 10 includes January 2007 and 2012.

ArcticMelt2

  • Grease ice
  • Posts: 929
    • View Profile
  • Liked: 192
  • Likes Given: 37
Re: The 2019/2020 freezing season
« Reply #845 on: February 04, 2020, 09:43:41 AM »
Quote
The January average temperature at Bethel of -5.5F (-20.8C) was 12.1F (6.7C) below 1981-2010 normal but not low enough to break into the top ten coldest. There is no trend at all for January temps at Bethel past 95 years.


ArcticMelt2

  • Grease ice
  • Posts: 929
    • View Profile
  • Liked: 192
  • Likes Given: 37
Re: The 2019/2020 freezing season
« Reply #846 on: February 04, 2020, 09:45:36 AM »
Quote
Bering Sea #seaice extent up to about 87% of 1981-2010 average for the start of February in @NSIDC data. Stormier weather pattern already underway for the region. This is likely to slow growth & modify ice distribution in the coming week.



Quote
Average #seaice extent in the Bering Sea for January from @NSIDC  data was a bit lower than 2019 and only 81% of the 1981-2010 average. Weather was favorable for ice: a late start to ice-up & above average #sst slowed ice growth.


ArcticMelt2

  • Grease ice
  • Posts: 929
    • View Profile
  • Liked: 192
  • Likes Given: 37
Re: The 2019/2020 freezing season
« Reply #847 on: February 04, 2020, 09:50:27 AM »
https://twitter.com/Climatologist49

Quote
The NCEP/NCAR (R1) Reanalysis agrees with the JRA55, January 2020 was the 2nd warmest on record behind 2016. For the Lower 48, it was the 5th warmest since 1948 (and probably top 5 to 8 since 1900). For Europe, possibly 2nd warmest.


El Cid

  • Nilas ice
  • Posts: 1236
    • View Profile
  • Liked: 523
  • Likes Given: 100
Re: The 2019/2020 freezing season
« Reply #848 on: February 04, 2020, 10:48:55 AM »
"The coldest January on Kodiak is the forerunner of the big summer ice loss? Top 10 includes January 2007 and 2012."

I think there might be some truth to that. Alaska is cold during the winter and Europe/lower 48 US is warm when the polar vortex is well behaved and no "cold-spills" reach them. Due to this there is usually little snowcover in NH midlatitudes so when spring comes they should warm up fast. This could of course lead to a fast meltout of the periphery in the Arctic which - given good weather - would lead to fast ice loss especially as there is not much old ice nowadays.

There are many ifs though...but I think we have a good chance of seeing some "fireworks" during this summer in the Arctic

uniquorn

  • Young ice
  • Posts: 2726
    • View Profile
  • Liked: 1224
  • Likes Given: 242
Re: The 2019/2020 freezing season
« Reply #849 on: February 04, 2020, 11:04:23 AM »
A quick look at the Atlantic side, jan10-feb3, uni-hamburg amsr2uhh