Support the Arctic Sea Ice Forum and Blog

Author Topic: Latest PIOMAS Volume update  (Read 1896978 times)

magnamentis

  • Guest
Re: Latest PIOMAS update (August)
« Reply #1050 on: August 05, 2016, 06:10:15 PM »
i know this will attract some heat again but i'm convinced that there must be a major flaw in the numbers piomas is providing. thickness overall and across the board is so much less that the volume numbers should be lower compared to other dates with the same area and extent levels. i expect a major correction one day in the future and/or a change to the algorithm. things have changed and i think that does not flow into their calculations.

this does not refer to graphs and other works based on the numbers, only started seriously doubting the source data a few months ago and each month that feeling gets stronger.

jai mitchell

  • Nilas ice
  • Posts: 2357
    • View Profile
  • Liked: 207
  • Likes Given: 60
Re: Latest PIOMAS update (August)
« Reply #1051 on: August 05, 2016, 07:32:53 PM »
i know this will attract some heat again but i'm convinced that there must be a major flaw in the numbers piomas is providing. thickness overall and across the board is so much less that the volume numbers should be lower compared to other dates with the same area and extent levels. i expect a major correction one day in the future and/or a change to the algorithm. things have changed and i think that does not flow into their calculations.

this does not refer to graphs and other works based on the numbers, only started seriously doubting the source data a few months ago and each month that feeling gets stronger.

I am seconding this perception, also not based on the graphs.  We have seen revisions before I think, it is very likely that we will see them in the future, though the process of 'failing on the side of least drama' may be a factor, I am not sure.

postscript edit:  however, looking at the total values, the estimation of difference (over) that I have is well within the margin of error, likely even within a single standard deviation (approximately 400 km^3 for July avg and (projecting) 650 km^3 for the Sept. avg.  I expect the Sept. Avg will come in around 5,350 km^3 but should read about 4,700 km^3.
« Last Edit: August 05, 2016, 07:47:44 PM by jai mitchell »
Haiku of Futures Passed
My "burning embers"
are not tri-color bar graphs
+3C today

magnamentis

  • Guest
Re: Latest PIOMAS update (August)
« Reply #1052 on: August 05, 2016, 09:38:55 PM »
@jai mitchell

thanks, you have more insight and experience and i'm glad for the more detailed elaboration.
as someone who in real life has nothing to do with ice, weather and the likes i often just see
that something is up but can't name it, hence i (we laymen) need the educated people like you.

looking forward with interest as to what the future will show as a correction but i have that
feeling it will be significant and the numbers you mention would well fall into that ball-park IMO.


Neven

  • Administrator
  • First-year ice
  • Posts: 9470
    • View Profile
    • Arctic Sea Ice Blog
  • Liked: 1333
  • Likes Given: 617
Re: Latest PIOMAS update (August)
« Reply #1053 on: August 06, 2016, 12:05:11 AM »
The enemy is within
Don't confuse me with him

E. Smith

icy voyeur

  • New ice
  • Posts: 63
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (August)
« Reply #1054 on: August 06, 2016, 12:21:12 AM »
Now the PSC has updated their PIOMAS  daily volume data, I updated my graphics, see the top post

Attached the current anomaly graph. July's curve of 2016 is closest to that of 2010, currently 3rd lowest year.


So to summarize what my weak eyes see.
The trend has been for an early dip in mid June. 2012 began that dip in the last week of May and was relentless through the 3rd week of June. 2010 began early in May and ran through the end of June.
2016 began mid May with a huge lead then stalled out mid June.
2011 also early May and kept surging through mid July.
As these are anomalies, it's all on a background of "average" melting.
This all seems to match the other metrics.
I'd interpret this as peripheral seas taking early hits with a stubborn core showing resistance. Weather plays its role as always and was a key factor in this year's reprieve. However, the overall trend is clear. The arctic is increasingly vulnerable to large ice volume losses in May and June.

Now, what is driving the "recovery" through July and into Aug? Is that an artifact of having melted the easy ice early? The reference line was still melting some of that easy ice in prior years during July and Aug? That's my read. The good news is that there's a core that is demonstrating some resilience. The bad news is that this core is absolutely facing a much earlier assault. Rebuttals and corrective analysis welcome.

Jim Pettit

  • Nilas ice
  • Posts: 1175
    • View Profile
  • Liked: 41
  • Likes Given: 11
Re: Latest PIOMAS update (August)
« Reply #1055 on: August 06, 2016, 12:46:33 AM »
I've updated my interactive 3D PIOMAS graphs with July's numbers (I also took the time to remove the somewhat confusing rectangular bottom grid from the 7-day "Death Spiral" graph and replace it with a more sensible circular grid). Click on any image for a direct link to the interactive version:

7-Day average PIOMAS volume "Death Spiral":
 


Monthly average PIOMAS volume:


Monthly average PIOMAS volume as a percentage of 1979 values:

S.Pansa

  • Frazil ice
  • Posts: 175
    • View Profile
  • Liked: 45
  • Likes Given: 5
Re: Latest PIOMAS update (August)
« Reply #1056 on: August 06, 2016, 06:23:20 AM »
You know, A-Team there are bigger problems with that map than the technical details of how it dithers and how is constructed but they have nothing to do with the person here who puts them together in spare time.

Take a look at the NASA MODIS products. The area north of Greenland where PIOMAS shows the thickest ice looks like hell on the MODIS products. There's no way the ice is as thick there as the PIOMAS model has it. The ice has been shattered and stretched by longshore movement and melted by weeks of normal to above normal temperatures after a warm winter.

I appreciate the maps all you guys make and I look at them very carefully.

For what it's worth, I am inclined to agree with that - and TOPAZ4 seems to do as well. Below is the thickness graph from TOPAZ4 for July, 31st & PIOMAS for the same date (both 1-4m). Of course I have no evidence to back this up - just a feeling that I get when I look at Worldview and the concentration maps. So rock solid science right there.  8)

seaicesailor

  • Guest
Re: Latest PIOMAS update (August)
« Reply #1057 on: August 07, 2016, 01:09:02 PM »
Not wanting to fall into the fallacy of criticizing a product because it doesn't show what one expects or wishes for. But I put the July 31 PIOMAS map by Wipneus (hope he/she does not mind) against the Uni Hamburg AMSR2 of the same day.
Am I the only one I am not the only one with the impression that, given the level of fragmentation of the pack at the end of July, PIOMAS is over-predicting thickness in many areas of Greenland Sea, ESS, CAB, Chukchi and Beaufort where UH is showing almost complete open water?
Just an impression.
Edit: I just read posts above of S Pansa and FishOutOfWater. I tend to agree as well.
« Last Edit: August 07, 2016, 01:22:41 PM by seaicesailor »

Nick_Naylor

  • Frazil ice
  • Posts: 291
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (August)
« Reply #1058 on: August 07, 2016, 01:36:52 PM »
I've had the same suspicion as well. It's obviously difficult to assimilate thickness data when good data doesn't really exist. I suspect though that PIOMAS may include an energy balance component that might help offset that deficiency - not that measuring that with any great precision is a piece of cake either.

However, to the extent energy considerations are included, this could lead to an overall volume estimate that is closer in aggregate than the gridded data suggests. It would be good to hear from someone more familiar with how PIOMAS works.

OSweetMrMath

  • Frazil ice
  • Posts: 137
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (August)
« Reply #1059 on: August 10, 2016, 07:44:05 PM »
Here is my PIOMAS anomaly graph with the current data. The graph basically shows that although the ice considered plunging in mid-May, since then it has returned to the typical level for this year (after accounting for the trend). As previously, the thick black line is 2016, 2010-2015 are colored red through purple, and earlier years are thin black lines.


icy voyeur

  • New ice
  • Posts: 63
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (August)
« Reply #1060 on: August 11, 2016, 05:35:50 AM »

So to summarize what my weak eyes see....

Was I incoherent, did I babble, I feel so alone ...

abbottisgone

  • Frazil ice
  • Posts: 297
  • "...I'm a rock'n'roll star,...... YES I ARE!!!!!!"
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (August)
« Reply #1061 on: August 22, 2016, 08:06:14 AM »
<<..

I'm reading your analysis now and I'm half way through... sounds good so far and well explained.

I'll give it some more thought but how is this for interesting: check out the weirdness of years 2007/2008 and mentally compare it to the wackiness of 2006.

It would seem a definate phase change is what you are looking at in the following years with the phase change itself being graphically recorded in the years 2006/2007/20008.

I only started looking for that after reading your summation of events. I will get back to reading the rest of it soon! (Good stuff!!  ;))
..
But I left school and grew my hair
They didn't understand
They wanted me to be respected as
A doctor or a lawyer man
But I had other plans..........

Nick_Naylor

  • Frazil ice
  • Posts: 291
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (August)
« Reply #1062 on: August 28, 2016, 05:55:29 PM »
Here's another way to look at the PIOMAS data. I converted the volume gain/loss to Btu's of heating cooling, which is admittedly not very interesting numerically, but it is kind of cool to think that July ice melt is providing a cooling service equal to 200 billion window air conditioners running flat out all month. (based on 12,000 Btu/hour units).

The chart shows that May and June cooling have been steadily growing since 1979 as the global temperature increases, in spite of the lower availability of ice.

However, by July and through August, the available ice area/extent has fallen so much that in spite of elevated temperatures, less cooling is available.

In future years - perhaps not too far distant - the inability of the ice to provide cooling in August will become painfully obvious.

FishOutofWater

  • Nilas ice
  • Posts: 1088
    • View Profile
  • Liked: 696
  • Likes Given: 332
Re: Latest PIOMAS update (August)
« Reply #1063 on: August 28, 2016, 06:25:48 PM »
Thick ice that has been driven by storm after storm towards the CAA is now moving south through the channels, especially the main channel. The weather is quite warm there because the sun is reasonably high that far south and there is no snow. Thick sea ice that has gone into the main channel has melted away and more is flowing in. The Coriolis effect is keeping the moving ice on the right side of the channel while the left side stays ice free. This situation is also allowing fresh water that piles up under the Beaufort high to drain from the Arctic. On the Atlantic side, warm fresh water is entering and thick ice driven towards Svalbard and Franz Joseph Island is rapidly melted by the thick layer of warm Atlantic water.

I suspect that the amount of volume loss that is happening now is unprecedented for so late in the melt season except, perhaps, for 2012.

Nick_Naylor

  • Frazil ice
  • Posts: 291
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (August)
« Reply #1064 on: August 28, 2016, 07:32:51 PM »
I suspect that the amount of volume loss that is happening now is unprecedented for so late in the melt season except, perhaps, for 2012.

That's an intuitively appealing interpretation, but it's unlikely we're setting any records for volume loss rates. Look at 2012 for example. Compared to the 1979-1999 period averages, 2012 lost only 86% of the average for the full month, and 52% of the average for that last week in August.

Because there wasn't all that much there to begin with.

   Volume Loss   

         Last Week   Month
         of August  of August
1979     -.456     -3.282
1980     -.437     -3.189
1981     -.666     -3.825
1982     -.433     -2.964
1983     -.474     -3.448
1984     -.489     -2.802
1985     -.481     -2.825
1986     -.393     -2.908
1987     -.557     -3.382
1988     -.472     -2.833
1989     -.432     -3.059
1990     -.242     -2.551
1991     -.419     -3.109
1992     -.194     -2.470
1993     -.374     -2.762
1994     -.321     -3.050
1995     -.312     -2.816
1996     -.364     -2.458
1997     -.275     -2.804
1998     -.442     -3.327
1999     -.599     -3.913
2000     -.453     -2.811
2001     -.220     -2.643
2002     -.314     -3.009
2003     -.375     -2.710
2004     -.425     -3.219
2005     -.442     -2.710
2006     -.261     -2.544
2007     -.214     -2.520
2008     -.461     -3.349
2009     -.307     -2.578
2010     -.419     -2.567
2011     -.145     -2.372
2012     -.218     -2.619
2013     -.200     -2.221
2014     -.296     -2.360
2015     -.473     -2.631

seaicesailor

  • Guest
Re: Latest PIOMAS update (August)
« Reply #1065 on: August 28, 2016, 08:25:09 PM »
Adding the August 27 UH image to July 31 PIOMAS and UH (sorry for the misalignment).
It is only August 27 ...
Volume only counts real melting, and real melting there was at periphery while storms dispersed ice.  Note Beaufort, the Wrangel arms, and ESS bite. Then we have the enormous transport of ice toward the Atlantic that seems to disappear as fast as it arrives.
Not sure how to compare this with other years, but watching the gif, the volume decrease should go well beyond 2012, since as Nick says, in August 2012 there was no so much left for real melting (at peripheral seas).
Last month I suspected that PIOMAS estimate was a bit on the high side. Taking a correction of this into account, I think volume may go below 2011, and later the minimum may get very close to 2012 when the Atlantic front advances once the ice stops being pushed towards the front.
MYI also getting dropped into the Greenland sea.

Edit: getting mesmerized by the overall effect of the mega cyclones. This was not ice that "was going to melt anyway"
« Last Edit: August 28, 2016, 08:48:38 PM by seaicesailor »

FishOutofWater

  • Nilas ice
  • Posts: 1088
    • View Profile
  • Liked: 696
  • Likes Given: 332
Re: Latest PIOMAS update (August)
« Reply #1066 on: August 29, 2016, 01:35:39 AM »
Of course, I was never suggesting that total volume loss rates in any August are anywhere near the loss rates in May, June and July. I am suggesting that the loss rates for so late in the year, the last 2 weeks of August, may be unprecedented.

We will see the numbers pretty soon and it will be very interesting.

A-Team

  • Young ice
  • Posts: 2977
    • View Profile
  • Liked: 944
  • Likes Given: 35
Re: Latest PIOMAS update (August)
« Reply #1067 on: August 30, 2016, 12:32:51 AM »
Looking now at Jim Pettit's admirable 3D interactives of monthly and weekly PIOMAS in #1055 (works in Firefox but not latest Opera) but not thrilled with lack of grayscale 'floor' and use of gregorian calendar (winter solstice off, lunar cycle off and irrelevant, weeks and months arbitrary units), I chased down the original daily data set at UW (which consists of 38 years x 365 days in three columns year-day#-km3 covering 1979-2016.

http://psc.apl.uw.edu/research/projects/arctic-sea-ice-volume-anomaly/data/

Bringing this into a year by day rectangular array, fleshing out the missing months of 2016 based on July 2016's average fraction of 2015 (0.884), normalizing by multiplying every entry by the quotient of the global maximum in 255 and rounding to integers in [0.255] or rather in this data set [85,255], this text file can be imported as a 32-bit grayscale using Import --> Text Image... in ImageJ, which converts correctly to 8-bit (provided Edit --> Options --> Conversions has 'Scale when converting' de-checked, preventing rescaling to [0.255]).

The txt attachment below provides the final numerical array used to make the initial image.

Back in olden times, everyone was familiar with the exact 1-1 correspondence of numerical data arrays and pixel arrays of the same dimensions (pictures) as numbers in the graphics engine drove the monitor display voltages. Today however it is surprisingly hard to find software other than matlab and imageJ comfortable with the back-and-forth (which extends with three array stacks to RGB, HSV etc).

To reiterate, all 38 years of daily Piomas data can be displayed as a simple 14k grayscale that is 365 pixels wide and 38 pixels high (UW drops day 366 of leap years). The grayscale value of each pixel is proportional to the volume of ice in the year of its row and day of its column.

If you read out the grayscale value by mousing over it, say 246 gray for day 71 of 1979, and divided by scaling normalization used (7.719), you recover UW's tabled ice volume for that day and year, 31.839 km3.

The Piomas data scarcely has the level of precision indicated so 232 values ranging 0 to 4,294,967,295 is also excessive. The actual ice volume on that day might have been somewhere between 30 and 33 instead of the 31.839 km3 reported. But let's play along in 32-bit because day to day relative values might be better and then drop to 8-bit at the end.

When should a numeric array be recast as an image? Always: it's far easier and faster to manipulate data using image tools and see patterns and trends. For example, the Piomas data can be contoured in interactive 3D right from ImageJ's Analyze --> 3D surface Plot --> Isolines.

The initial image is quite small at 13,870 = 38 x 365 pixels and a little tricky to faithfully enlarge without unwanted interpolation: chose 'none' instead of bicubic etc and restrict to integral multiples to say triple the size of individual pixels without them being affected by neighboring values, as in the 3rd image below (which had to be removed because its width was throwing off forum display of the other images).

You can always drop down to the underlying spreadsheet and replicate an image operation with numbers (eg for line graphs or statistical analysis) but when was the last time you looked at the numerical array underlying a Landsat image or built a 36 layer spreadsheet for the Modis channels?

The animation shows a simple pattern analysis. The 3rd image was blurred slightly and posterized to two levels, then rescaled to forum limits. This shows as the years went by, ice volume really started shrinking in the summer months (darker grays).

A-Team

  • Young ice
  • Posts: 2977
    • View Profile
  • Liked: 944
  • Likes Given: 35
Re: Latest PIOMAS update (August)
« Reply #1068 on: August 30, 2016, 12:48:53 AM »
Here are a couple of quick views taken from the interactive 3D viewer in ImageJ. These involve opening the text file attached above as a graphic. The first replicates Jim Petitt's view but from a volume-quantitative grayscale of daily Piomas values. The second shows the annual trend as a contour pattern: the volume loss season has gotten markedly longer.

The second image shows one of the many things that D. Tschumperle's gmic site can do, the Piomas image was processed to 15 contours. The years range from1979 at the top to 2016 along the bottom. The outlier years show up quite clearly in this.
« Last Edit: August 30, 2016, 12:59:59 AM by A-Team »

Neven

  • Administrator
  • First-year ice
  • Posts: 9470
    • View Profile
    • Arctic Sea Ice Blog
  • Liked: 1333
  • Likes Given: 617
Re: Latest PIOMAS update (August)
« Reply #1069 on: August 30, 2016, 12:45:22 PM »
This open-source paper looks highly interesting (if only because of the authors): Comparison of Arctic Sea Ice Thickness from Satellites, Aircraft, and PIOMAS Data
The enemy is within
Don't confuse me with him

E. Smith

oren

  • Moderator
  • First-year ice
  • Posts: 9805
    • View Profile
  • Liked: 3584
  • Likes Given: 3922
Re: Latest PIOMAS update (August)
« Reply #1070 on: August 30, 2016, 01:35:43 PM »
This open-source paper looks highly interesting (if only because of the authors): Comparison of Arctic Sea Ice Thickness from Satellites, Aircraft, and PIOMAS Data

It seems there are biases to both directions. However, one important sentence from the abstract caught my eye:
Quote
All satellite-retrieved ice thickness products and PIOMAS overestimate the thickness of thin ice (1 m or less) compared to IceBridge

As the Arctic is reverting to thin scattered ice all over the place, I get the feeling PIOMAS is overestimating remaining volume, based on past correlations that are no longer accurate.

Jim Pettit

  • Nilas ice
  • Posts: 1175
    • View Profile
  • Liked: 41
  • Likes Given: 11
Re: Latest PIOMAS update (August)
« Reply #1071 on: August 30, 2016, 05:11:53 PM »
Looking now at Jim Pettit's admirable 3D interactives of monthly and weekly PIOMAS in #1055

Thanks for the kind word; 'admirable' is good. ;-)

(works in Firefox but not latest Opera)

I tested in many browsers, but out of necessity ignored Opera. Chrome and Firefox together comprise 89% of all browser traffic, while Opera comes in at just over 1%. Not that I like to dismiss anyone, but, well, you know...

not thrilled with lack of grayscale 'floor'...

I can certainly desaturate the graphs. But I think the colors, along with the Z-axis time series, do a pretty good job of showing the state of things, everything being equal.

use of gregorian calendar (winter solstice off, lunar cycle off and irrelevant, weeks and months arbitrary units)...

I've had a number of people over the years request that I place the annual maximum--or the annual minimum, or the winter solstice, or one of the equinoxes--of my various "death spiral" charts and graphs at the top, the reason given being that using calendar months is arbitrary. I suppose it is, but January 1 seems logical as do many annual graphs (including popular derivatives such as Andy Lee Robinson's). Months are used as guideposts so the viewer knows roughly where in the annual cycle a particular batch of data lie. And a 7-day period is highly arbitrary, I'll admit, but given that even the most powerful browser or tablet would likely suffer under the math and memory required to keep track of and move 15,000 or so screen points in real time, I chose what I thought was the best bend of performance and precision--and thus weeks.

(FWIW, I experimented with a flat, non-interactive graph using just daily data, and it looked much like yours, though in color. But as soon as I tried moving it around in three-dimensional space, by computer balked. Perhaps when processors get a little better... ;-)

greatdying2

  • Grease ice
  • Posts: 574
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (August)
« Reply #1072 on: August 31, 2016, 04:21:55 PM »
Here is my PIOMAS anomaly graph with the current data. The graph basically shows that although the ice considered plunging in mid-May, since then it has returned to the typical level for this year (after accounting for the trend). As previously, the thick black line is 2016, 2010-2015 are colored red through purple, and earlier years are thin black lines.

So which of these is wrong?
Or is the ice unusually thick this year?  :o
The Permian–Triassic extinction event, a.k.a. the Great Dying, occurred about 250 million years ago and is the most severe known extinction event. Up to 96% of all marine species and 70% of terrestrial vertebrate species became extinct; it is also the only known mass extinction of insects.

Nick_Naylor

  • Frazil ice
  • Posts: 291
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (August)
« Reply #1073 on: August 31, 2016, 05:33:37 PM »
gd2, there are two reasons for the difference:

1. The area chart includes August, while the volume data ends at July.
b. You posted the CAB area chart - not the whole Arctic. We have more than the usual amount in the peripheral areas this year.

If you go back one month on the "total" area chart, 2016 looked reasonably "normal" by post-2012 standard.

greatdying2

  • Grease ice
  • Posts: 574
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (August)
« Reply #1074 on: August 31, 2016, 10:39:51 PM »
Nick, Then it follows that we can expect a large drop in the next PIOMAS update and a large jump in the negative anomaly.  ;)
The Permian–Triassic extinction event, a.k.a. the Great Dying, occurred about 250 million years ago and is the most severe known extinction event. Up to 96% of all marine species and 70% of terrestrial vertebrate species became extinct; it is also the only known mass extinction of insects.

Nick_Naylor

  • Frazil ice
  • Posts: 291
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (August)
« Reply #1075 on: September 01, 2016, 12:41:47 AM »
I agree, but perhaps it will not be too exciting. Perhaps like this, or possibly record low, although I wouldn't bet on it.

epiphyte

  • Frazil ice
  • Posts: 387
    • View Profile
  • Liked: 22
  • Likes Given: 21
Re: Latest PIOMAS update (August)
« Reply #1076 on: September 01, 2016, 06:50:04 AM »
The PIOMAS model usually gets the thickness right when it is zero. I think they call it "forcing" ;)

... so when you see an unexpectedly rapid PIOMAS volume drop just as (uniformly thin) area is falling off a cliff, it seems to inform us that the earlier modeled thickness was incorrect, or the melt was underestimated, or both. But it doesn't of itself tell us which.

abbottisgone

  • Frazil ice
  • Posts: 297
  • "...I'm a rock'n'roll star,...... YES I ARE!!!!!!"
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (August)
« Reply #1077 on: September 01, 2016, 07:17:46 AM »
This open-source paper looks highly interesting (if only because of the authors): Comparison of Arctic Sea Ice Thickness from Satellites, Aircraft, and PIOMAS Data

It seems there are biases to both directions. However, one important sentence from the abstract caught my eye:
Quote
All satellite-retrieved ice thickness products and PIOMAS overestimate the thickness of thin ice (1 m or less) compared to IceBridge

As the Arctic is reverting to thin scattered ice all over the place, I get the feeling PIOMAS is overestimating remaining volume, based on past correlations that are no longer accurate.

Good find by Neven then. Overestimation of volume, as you point out as the corrolary of that abstracted quote,  would have to seem like the worlds most serious problem at the moment I dare say.

Combine that with the well established trend toward zero we have the possibility of REAL panic on our hands sooner rather than later.

MYI disappearing possibly this year was the number 1 water cooler conversation in government departments around the world- including Australia- and that was back in January...



..
But I left school and grew my hair
They didn't understand
They wanted me to be respected as
A doctor or a lawyer man
But I had other plans..........

marcel_g

  • Frazil ice
  • Posts: 127
    • View Profile
    • Art by Marcel Guldemond
  • Liked: 54
  • Likes Given: 425
Re: Latest PIOMAS update (August)
« Reply #1078 on: September 01, 2016, 02:41:47 PM »
So basically if PIOMAS is overestimating the thickness of thin ice, and since most of the thick MYI has melted out since 2007, and especially since 2012, then can we surmise that the volume situation is even more dire than the PIOMAS numbers are showing?

Especially considering the distribution of the thick vs thin ice means that for 2016 there has been a greater proportion of thin ice to thick ice, so that the overestimation becomes larger as the thicker ice disappears. I'm thinking of the Beaufort arm last year, which was thick MYI and didn't survive the melt season, and the thick ESS ice that has all but melted out this year, and the MYI that's currently headed out of the Fram right now.

The overestimation would also increase as the melt season progresses and the ices thins out.

As well, could we say that the minimum volume might be approaching 2012 again, since 2012 at this point was a relatively compact contiguous area that didn't appear to be too thin, whereas a lot of this year's area is thin dispersed rubble and slush?


Richard Rathbone

  • Nilas ice
  • Posts: 1730
    • View Profile
  • Liked: 387
  • Likes Given: 24
Re: Latest PIOMAS update (August)
« Reply #1079 on: September 01, 2016, 04:33:34 PM »
So basically if PIOMAS is overestimating the thickness of thin ice, and since most of the thick MYI has melted out since 2007, and especially since 2012, then can we surmise that the volume situation is even more dire than the PIOMAS numbers are showing?

No, because if you look elsewhere in the same paper you can draw the opposite conclusion (PIOMAS says the ice is a lot thinner than ICESAT does in the autumn as well as saying thin ice is a bit thicker than ICEbridge does in the Spring)

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Latest PIOMAS update (September)
« Reply #1080 on: September 03, 2016, 11:16:03 AM »
PIOMAS data is in, both the official daily volume and the gridded thickness data.

For the volume data I updated my graphics, see the top post

Wipneus

  • Citizen scientist
  • Young ice
  • Posts: 4220
    • View Profile
    • Arctische Pinguin
  • Liked: 1025
  • Likes Given: 0
Re: Latest PIOMAS update (September)
« Reply #1081 on: September 03, 2016, 11:18:53 AM »
The really last post.

From the gridded daily data, the animation.

slow wing

  • Grease ice
  • Posts: 823
    • View Profile
  • Liked: 155
  • Likes Given: 546
Re: Latest PIOMAS update (September)
« Reply #1082 on: September 03, 2016, 11:53:25 AM »
Thanks, Wipneus!

Maybe a final, final post showing the ice at the last day of the month?  :P


Best of luck with your move.  :)


EDIT: here's the screenshot anyway if Wipneus is already lost inside a packing crate...
« Last Edit: September 03, 2016, 12:02:56 PM by slow wing »

greatdying2

  • Grease ice
  • Posts: 574
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (September)
« Reply #1083 on: September 03, 2016, 02:46:35 PM »
I'm confused. That is supposed to model the same ice represented here (1 day diff)?
The Permian–Triassic extinction event, a.k.a. the Great Dying, occurred about 250 million years ago and is the most severe known extinction event. Up to 96% of all marine species and 70% of terrestrial vertebrate species became extinct; it is also the only known mass extinction of insects.

Neven

  • Administrator
  • First-year ice
  • Posts: 9470
    • View Profile
    • Arctic Sea Ice Blog
  • Liked: 1333
  • Likes Given: 617
Re: Latest PIOMAS update (September)
« Reply #1084 on: September 03, 2016, 02:51:29 PM »
Thanks, Wip. Here's PIOMAS September 2016.
The enemy is within
Don't confuse me with him

E. Smith

seaicesailor

  • Guest
Re: Latest PIOMAS update (September)
« Reply #1085 on: September 03, 2016, 05:47:24 PM »
I'm confused. That is supposed to model the same ice represented here (1 day diff)?
I am confused too. Last month same happened and PIOMAS shows now a big decrease. This in part may be due to the total melt out of areas where thickness was overestimated. I expect the same to happen during this month.
This gif compares PIOMAS August 31 with UH AMSR2 and MODIS 7-2-1 same date. Increased contrast in Wipneus' map to enhance the core of 0.75 - 1 m effective thickness along the Wrangel arm. This is a lot of ice that actually isn't there.
PIOMAS may benefit from better models of edge zone final melt out, seems authors are looking at it
http://onlinelibrary.wiley.com/doi/10.1002/2015JC010770/full
In this 2015 paper Zhang and others develop theories of floe break up and final melting. In fact the  abstract begins with:
To better describe the state of sea ice in the marginal ice zone (MIZ) with floes of varying thicknesses and sizes, both an ice thickness distribution (ITD) and a floe size distribution (FSD) are needed. In this work, we have developed a FSD theory ...
and finishes with:
... the ability to explicitly simulate multicategory FSD and ITD together may help to incorporate additional model physics, such as FSD-dependent ice mechanics, surface exchange of heat, mass, and momentum, and wave-ice interactions
Thanks to @Wipneus for making the PIOMAS maps and all the work, good luck with the move
« Last Edit: September 03, 2016, 06:02:26 PM by seaicesailor »

Jim Hunt

  • First-year ice
  • Posts: 6268
  • Don't Vote NatC or PopCon, Save Lives!
    • View Profile
    • The Arctic sea ice Great White Con
  • Liked: 893
  • Likes Given: 87
Re: Latest PIOMAS update (September)
« Reply #1086 on: September 03, 2016, 06:10:36 PM »
I'm confused. That is supposed to model the same ice represented here (1 day diff)?

In brief, yes. Here's the high res AMSR2 version for August 31st:

http://GreatWhiteCon.info/2016/09/the-2016-arctic-sea-ice-metric-minima/#Sep-03
"The most revolutionary thing one can do always is to proclaim loudly what is happening" - Rosa Luxemburg

greatdying2

  • Grease ice
  • Posts: 574
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (September)
« Reply #1087 on: September 04, 2016, 08:46:07 AM »
Yeah, the Wrangle arm is way off.

There is also a large area just to the Russian side of the pole (in the middle of the "broken heart") that is modelled at contiguous 1.5m+ thickness when in fact it is full of holes. Can what ice that does remain there really be that thick? I got the impression from watching ice movement during the cyclones that this ice got pushed toward the Atlantic (while the ice initially closer to the Atlantic melted out), and so this are should have thinned dramatically, which is what the satellite photos show.

I presume that unprecedented changes to the ice pack are difficult to model, but you would think that more weight could be given to data that clearly shows areas of deterioration, rather than assuming that areas that used to be thick will remain to be thick, which seems to be what the model puts out.

It's really unfortunate because volume is what matters, and the lack of a believable model (one that doesn't dramatically disagree with available data) means that focus may continue to rest on extent.  :-\

Edit: What I mean about the ice movement is better described my A-Team's animations and Oren's comments today in the Homebrew thread:

The interesting thing is that while the Barents front looks roughly the same during the animation timeframe, the area to the north keeps losing concentration and polynyas are opening up. A naive observer might assume that the melting area is near the pole.
Early in the season it has already been noted that the Atlantic sector was the killing ground, but the effect took a long time to materialize visually since the inner CAB had lots of ice to supply when the season started. Had the ice been less mobile and fractured this year, it might have survived in much greater numbers.
« Last Edit: September 04, 2016, 08:54:46 AM by greatdying2 »
The Permian–Triassic extinction event, a.k.a. the Great Dying, occurred about 250 million years ago and is the most severe known extinction event. Up to 96% of all marine species and 70% of terrestrial vertebrate species became extinct; it is also the only known mass extinction of insects.

greatdying2

  • Grease ice
  • Posts: 574
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (September)
« Reply #1088 on: September 04, 2016, 09:11:10 AM »
This (Sept. 25)...

Looking at Wipneus' PIOMAS animation, I see no sign at all that it models this ice movement out from the middle to the edge. Instead it maintains a continuous gradient from the pole to the Atlantic.
The Permian–Triassic extinction event, a.k.a. the Great Dying, occurred about 250 million years ago and is the most severe known extinction event. Up to 96% of all marine species and 70% of terrestrial vertebrate species became extinct; it is also the only known mass extinction of insects.

epiphyte

  • Frazil ice
  • Posts: 387
    • View Profile
  • Liked: 22
  • Likes Given: 21
Re: Latest PIOMAS update (September)
« Reply #1089 on: September 04, 2016, 09:21:58 AM »
Yeah, the Wrangle arm is way off.

There is also a large area just to the Russian side of the pole (in the middle of the "broken heart") that is modelled at contiguous 1.5m+ thickness when in fact it is full of holes. Can what ice that does remain there really be that thick? I got the impression from watching ice movement during the cyclones that this ice got pushed toward the Atlantic (while the ice initially closer to the Atlantic melted out), and so this are should have thinned dramatically, which is what the satellite photos show.

I presume that unprecedented changes to the ice pack are difficult to model, but you would think that more weight could be given to data that clearly shows areas of deterioration, rather than assuming that areas that used to be thick will remain to be thick, which seems to be what the model puts out.

It's really unfortunate because volume is what matters, and the lack of a believable model (one that doesn't dramatically disagree with available data) means that focus may continue to rest on extent.  :-\

It's just a model...  not an Oracle. This is the only time of year when its predictions for the melt season can be validated... Or not. There could be years worth of overestimates in there which wouldn't become apparent until suddenly there was nothing where it thought there was something - like today. Also, thin, fragmented ice in summer is pretty much chaotic - so as the volume shrinks and becomes more uniformly distributed, it seems inevitable that the significance of unmodeled behavior becomes more apparent.

Andreas T

  • Nilas ice
  • Posts: 1149
    • View Profile
  • Liked: 18
  • Likes Given: 4
Re: Latest PIOMAS update (September)
« Reply #1090 on: September 04, 2016, 10:23:32 AM »
.....

I presume that unprecedented changes to the ice pack are difficult to model, but you would think that more weight could be given to data that clearly shows areas of deterioration, rather than assuming that areas that used to be thick will remain to be thick, which seems to be what the model puts out.

It's really unfortunate because volume is what matters, and the lack of a believable model (one that doesn't dramatically disagree with available data) means that focus may continue to rest on extent.  :-\

....
Too much criticism of the PIOMAS model shows very little understanding of what the model does. It is a wrong assumption to think the model "assumes that areas which used to be thick will remain to be thick". The model attempts to work out how physical processes which are known will affect the thinning and thickening of the ice.  It can not model individual holes and gaps in the ice, but should show a thinning of the average ice thickness where thick ice becomes more dispersed.
You are not showing available thickness data but you are choosing to believe that you can guess ice thickness from looking at a satellite image.
I don't doubt the model may have difficulties to match what is happening in these dispersed areas away from the ice edge if it can not resolve or model the processes there well enough, but to dismiss its findings so readily is doing more to discredit the criticism than the model.

oren

  • Moderator
  • First-year ice
  • Posts: 9805
    • View Profile
  • Liked: 3584
  • Likes Given: 3922
Re: Latest PIOMAS update (September)
« Reply #1091 on: September 04, 2016, 11:22:57 AM »
It can not model individual holes and gaps in the ice, but should show a thinning of the average ice thickness where thick ice becomes more dispersed.
Indeed. But eyeballing the PIOMAS animation, at least it seems that the apparent reduction in concentration in the area north of the "Barents front" is not translated into a reduction in average thickness. Where concentration is dropping sharply (due to motion towards the front) average thickness should drop as well, since the remaining ice is not growing any thicker. An AMSR2 animation side by side with PIOMAS covering the same area should clearly show whether that's happening in the model or not.

greatdying2

  • Grease ice
  • Posts: 574
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: Latest PIOMAS update (September)
« Reply #1092 on: September 04, 2016, 12:46:00 PM »

Too much criticism of the PIOMAS model shows very little understanding of what the model does. [...] but to dismiss its findings so readily is doing more to discredit the criticism than the model.
I never claimed to understand how the model works, and I do not. I am simply stating from a naive observer's perspective that the model output does not match observations.

To your other point, yes I think it is clear that the ice in that area is nowhere near 1.5m thick (on average). The animation provided by Wipneus of PIOMAS data shows that this area does not thin as the ice in the middle of the pack splits and the right half gets pushed toward the Atlantic, but rather shows it gradually thinning as if such an event never occurred. My naive guess is that this process is not in the model.

People are free to draw their own conclusions, naive or or otherwise. I was rather hoping that someone with a better understanding of the model would comment, rather than simply criticize the observation.
The Permian–Triassic extinction event, a.k.a. the Great Dying, occurred about 250 million years ago and is the most severe known extinction event. Up to 96% of all marine species and 70% of terrestrial vertebrate species became extinct; it is also the only known mass extinction of insects.

seaicesailor

  • Guest
Re: Latest PIOMAS update (September)
« Reply #1093 on: September 04, 2016, 02:44:31 PM »
...
Too much criticism of the PIOMAS model shows very little understanding of what the model does. It is a wrong assumption to think the model "assumes that areas which used to be thick will remain to be thick". The model attempts to work out how physical processes which are known will affect the thinning and thickening of the ice.  It can not model individual holes and gaps in the ice, but should show a thinning of the average ice thickness where thick ice becomes more dispersed.
You are not showing available thickness data but you are choosing to believe that you can guess ice thickness from looking at a satellite image.
I don't doubt the model may have difficulties to match what is happening in these dispersed areas away from the ice edge if it can not resolve or model the processes there well enough, but to dismiss its findings so readily is doing more to discredit the criticism than the model.

I am personally not dismissing PIOMAS but agree with greatdying2 that this month there are regions where ice is overestimated. Carefully inspecting MODIS and AMSR2 images, it is not possible that for instance the central area of the Wrangel arm has a thickness of 0.75 to 1 m.

Perhaps a very good introduction to understand PIOMAS to some of us total amateurs is Chris Reynolds blog. For instance
http://dosbat.blogspot.com.es/2015/04/what-is-piomas-gice.html
http://dosbat.blogspot.com.es/2013/02/cryo-sat-2-and-piomas.html
http://dosbat.blogspot.com.es/2016/05/piomas-april-2016.html

PIOMAS has deviations with respect to observations, but these are consistent, meaning it is known to a great extent when, where, and why they happen. The amount of scientific work over many years (since the original development early-nineties based mostly on a theoretical model from the seventies!), modeling improvements, comparison with submarine, airplane, and sat measurements is overwhelming.
One of the things that Chris Reynolds exploits most in his blog is looking the PIOMAS sub-grid ice thickness distribution. PIOMAS does not simply consider the ice as a single layer of stretched broken or glued ice, so to speak. Within each grid cell, the ice is considered a discreet distribution of ice elements with different thicknesses. For each grid cell, the model solves a number of equations equal to the amount of ice thickness bins. The huge benefit of this is that mechanical and thermodynamical processes can be partly resolved as a function of sub-grid thickness and not just modeled as a whole. This includes ridging, fracture, ice temperature, melting rate... The conservation equations not only dictate how each category of ice thickness changes, but rather in which proportion ice mass and energy is transferred from one bin to the adjacent ones.
http://psc.apl.washington.edu/zhang/IDAO/ZhangModelingGlobalIce.pdf
PIOMAS assimilates NASA SSMI concentration and velocity data
http://psc.apl.washington.edu/zhang/Pubs/Lindsay_Zhang_ic_assimilation_2006.pdf
but in the paper they explain that this assimilation has to be done with great care to not contradict model equations and deliver unphysical results, and to avoid assimilating errors inherent to the concentration itself, so a match of PIOMAS-derived ice extent and observed ice extent is not possible. Also they explain that, whenever ice melts out and ice extent goes to zero in a grid cell, the ice volume that disappears is greater than the the ice volume that reappears, by a large a margin. Perhaps this in part explains why PIOMAS lags in refreezing with respect to Cryosat 2 (see Chris Reynolds entry). No comparison with Cryosat in summer, however it may well happen that it lags in thickness until extent suddenly goes to zero.
Analysis of uncertainties and errors of PIOMAS have been published by the creators of the model themselves and others,
http://onlinelibrary.wiley.com/doi/10.1029/2011JC007084/full

Zhang himself recognizes the need of correctly modeling sub-grid floe size distribution and not only sub-grid ice thickness distribution (see paper I cited in a previous post). This is important precisely where the ice is highly fractured and at the edges of the ice pack. If the model was capable of modeling this distribution, it could resolve better the effect of waves, lateral melting, final melt-out ..., and it would probably improve the modeling precisely where we complain it shows some differences, in a year where ice edge has been extraordinarily affected by storms without completely melting out as in 2012. These differences are probably resolved once the marginal ice zone is assimilated as part of the ice pack during refreezing.
I don't doubt of PIOMAS consistency (and accuracy too within its limitations!!) given the incredible, original, rigorous scientific work behind it.
Just pointing out some evident but localized deviations in August and September 2016.

EDIT: Note, on the other hand, that PIOMAS shows a very large reduction of volume during August compared to other years, as most of we expected! This is the consistency of the model. And given the current large extension of the so-called marginal ice zone in scientific literature (the very broken ice edge during melting season), I presume that it will show a significant reduction until minimum. We'll see...
« Last Edit: September 04, 2016, 03:28:38 PM by seaicesailor »

FishOutofWater

  • Nilas ice
  • Posts: 1088
    • View Profile
  • Liked: 696
  • Likes Given: 332
Re: Latest PIOMAS update (September)
« Reply #1094 on: September 04, 2016, 04:57:15 PM »
Yes, we knew there were large areas of thin ice in vulnerable regions where the surrounding water would contribute to bottom melting. The large August volume melt was expected by most of us. I even wrote a comment somewhere around here predicting it.

I appreciate the comment discussing the details on the capabilities of the PIOMAS model.

A-Team

  • Young ice
  • Posts: 2977
    • View Profile
  • Liked: 944
  • Likes Given: 35
Re: Latest PIOMAS update (September)
« Reply #1095 on: September 04, 2016, 10:30:03 PM »
Quote
incredible, original, rigorous scientific work behind it
I am really skeptical that modeling has added any real value. The physical processes are far too complex, seldom studied, rarely revisited, and little understood, a familiar refrain all across climate modeling. Mostly it is just fed-in observational data that itself is not great.

The animation below rescales and rotates the crappy little Arctic map that they use to fit AMSR2 microwave data, both for the 31st of August. The 3rd frame shows the alignment quality and the 4th 'goodness of fit' with microwave-observed ice locations for the same day. It doesn't even have the surface ice right. In terms of thickness, to say the ice is 1 m thick with an error >0.5 m does not seem insightful given how severely the range is restricted from the get-go.

Richard Rathbone

  • Nilas ice
  • Posts: 1730
    • View Profile
  • Liked: 387
  • Likes Given: 24
Re: Latest PIOMAS update (September)
« Reply #1096 on: September 04, 2016, 11:16:51 PM »
It doesn't even have the surface ice right. In terms of thickness, to say the ice is 1 m thick with an error >0.5 m does not seem insightful given how severely the range is restricted from the get-go.

You do realise that the data you are comparing it against has just the same sort of errors in it? Concentration errors are 15% at low concentration, so the place where you are putting the ice edge is being determined by a signal that is the same size as the error in it. 

seaicesailor

  • Guest
Re: Latest PIOMAS update (September)
« Reply #1097 on: September 05, 2016, 12:24:30 AM »
Well, there is abundant world-class scientific work behind it, that much can be said.

Artful Dodger

  • Frazil ice
  • Posts: 459
  • The traps have got him, and that's all about it!
    • View Profile
  • Liked: 24
  • Likes Given: 127
Re: Latest PIOMAS update (September)
« Reply #1098 on: September 05, 2016, 06:19:42 AM »
Hi folks,

A few points on PIOMAS:
  • it's silly to compare PIOMAS and AMSR2 charts. PIOMAS is based on NSIDC data (you know, the 25km res. data?). It's produced from the same source as the "blue marble" SIC chart from Aug 25, 2016 attached below
  • PIOMAS claims +/- 1K km^3 accuracy. They are primarily concerned with representing the trend reliably, not the varagies of seasonal wx
  • if you REALLY want some insight into current volume, you should be looking at SMOS thickness data. When sea ice is <50 cm thick, SMOS is quite accurate, it's not a model (it's DATA), it's released daily, and there's lot's of <50 cm thick sea ice out there ATM.
Have fun storming the Castle, boys (and girls)!

Cheers,
Lodger

« Last Edit: September 05, 2016, 12:03:59 PM by Artful Dodger »
Cheers!
Lodger

budmantis

  • Nilas ice
  • Posts: 1220
    • View Profile
  • Liked: 7
  • Likes Given: 34
Re: Latest PIOMAS update (September)
« Reply #1099 on: September 05, 2016, 06:36:08 AM »
Hi folks,

varagies


Cheers,
Lodger

Lodger:

I've enjoyed reading your posts here as well as on the asib over the last few years. At the risk of being anal, I think you meant "Vagaries".

Cheers,

Bud