Near Real Time Sea Ice Volume

[Random_Weather]

Hi,

I build up an Near Real Time Sea Ice Volume (NRT SIV) product based on merged data from SMOS and Cyrosat2. To make it very simple, i decide to use mean sea ice thickness(SIT) and extent(SIE) data from NSIDC, beside this, it would also be possible by grid by grid methode, but since in winter, sea ice concentration (SIC) has not much variance over the domain, its seem unlikely that much differences would occur.

First things first, some Explantion:

SIV:  Daily
SIT:  5 five-day trailing average
SIE:  5 five-day trailing average

Its because of the merged SIT is just in a 5 five-day trailing average format accessible, therefor its also used for SIE and the final product SIV. What does this mean?

SIV(5 Nov): SIT(mean(1-5Nov))*SIE(mean(1-5Nov))


Some Plots:


SIV 2011-2018


NRT SIV vs. Cyrosat2 and PIOMAS (Okt to May)



Open Points for Future:
- Uncertainy-Bads from merged SMOS-Cyrosat2
- Testing other SIE like Uni-Bremen
- Melt-Saison-Model

Data-Source:
ftp://ftpsrv2.awi.de/sea_ice/product/cryosat2_smos/v202/nh/ ->analysis_sea_ice_thickness CS2SMOS merged sea_ice_thicknes
https://nsidc.org/arcticseaicenews/sea-ice-tools/ -> ea_Ice_Index_Daily_Extent_G02135_v3.0.xlsx -> 5 five-day trailing average

[Random_Weather]

SIV to 15 Dez:


PS: Update will occour every week or so

[gerontocrat]

It will be interesting to see over time how this product and PIOMAS compare over a whole year. Especially summer, when thin ice makes life so much more difficult.

[Aleph_Null]

Oh, I think this is all excellent! Though I'm sometimes accused of being easily impressed...

Thanks much for the explanation of your metric. Many of us are yearning for the best way to understand the condition of Arctic sea ice, while watching it get blown hither and yon. (I've unexpectedly absorbed a whole new feeling for ice dynamics from the Mosaic reports and data.) Efforts just like yours are the reason I can't break my ASIF habit.

[Random_Weather]

It will be interesting to see over time how this product and PIOMAS compare over a whole year. Especially summer, when thin ice makes life so much more difficult.

The Problem is, from 15th April to 21 Okt there will no data be available, therefore its just only winter product.

Meanwhile in compare to PIOMAS (used the same way as decribed before) from 21th Okt to 1. Dez 2019:

[Random_Weather]

And perhaps the Scattern Plots for Cyrosat 2 and PIOMAS for 2011 to 2019(april) mon_mean

PIOMAS vs. NRT SIV



Cyrosat2 vs. NRT SIV


No surprise, Correlation must better against Cyrosat2 because its related to it in main fraction, besides the infill from SMOS. Also to note, on higher SIV, products beginn to have larger variances

[Random_Weather]

Oh, I think this is all excellent! Though I'm sometimes accused of being easily impressed...

Thanks much for the explanation of your metric. Many of us are yearning for the best way to understand the condition of Arctic sea ice, while watching it get blown hither and yon. (I've unexpectedly absorbed a whole new feeling for ice dynamics from the Mosaic reports and data.) Efforts just like yours are the reason I can't break my ASIF habit.

Thx for your kind words, by the topic of MOSAiC, i use this here for track position
ftp://ftpsrv2.awi.de/sea_ice/mosaic/jpg/large/

greets

[blumenkraft]

Oh, I think this is all excellent!

+1

[Random_Weather]

Its also possible to make Sea Ice Snow Mass (SISM) for the arctic: ftp://ftpsrv2.awi.de/sea_ice/product/cryosat2/v2p2/nh/l3c_grid/monthly/

- Use Snow Thickness
- Use Snow Density
- Use NSIDC Extent

You get:


In other words, arround 190-200 Gigatonnes (Gt) Snow Mass every year on arctic sea ice, before melt saison beginns. This is quite stabile, havent though its that low variance

[Tor Bejnar]

Very interesting, RW.  Conceptually, the depth of snow will increase as the Arctic warms, but the winter ice area (~= extent) is decreasing, so a stasis is possible. 

I was wondering somewhere if increased winter ice mobility would cause more frequent lead openings and thus more opportunity for snow getting blown into (onto?) the water.  But maybe this possibility is an order (or two) of magnitude below the big picture.

[Random_Weather]

Cery interesting, RW.  Conceptually, the depth of snow will increase as the Arctic warms, but the winter ice area (~= extent) is decreasing, so a stasis is possible. 

I was wondering somewhere if increased winter ice mobility would cause more frequent lead openings and thus more opportunity for snow getting blown into (onto?) the water.  But maybe this possibility is an order (or two) of magnitude below the big picture.

In the short period, (2011-2018) Trends

SISM:

Okt; -0.122Gt/y
Nov: -0.393Gt/y
Dez: -0.260Gt/y
Jan:  -0.452Gt/y
Feb:  -0.633Gt/y
Mar:  -0.870Gt/y
Apr:  -1.477Gt/y


Mean Snow Depth (MSD)

Okt; +0.07cm/y
Nov: +0.01cm/y
Dez: +0.11cm/y
Jan:  +0.06cm/y
Feb:  +0.04cm/y
Mar:  +0.11cm/y
Apr:  +0.09cm/y


Its a to short period for making any significant claim, but as the data would imply, loose of sea ice(could also be the leads or the drift of ice with very much snow) overcompensate increase snowfall that its turn out an negativ Trend of SISM

[gandul]

SIV to 15 Dez:


PS: Update will occour every week or so

This is really nice. Piomas just updated (or Wipneus just noticed) ,can you make an up-to-date comparison?

[Random_Weather]

Update to 21.12


@gandul
Nxt weeks i am most time at holiday, would make nxt up-to-date comparison in early january

[gandul]

Ok thx. Piomas is picking up much faster now, but that seems a consistent departure over the years.

[Random_Weather]

Just in short, if taken account for instrumental uncertainy, it would look like this: (up to 24. Dez)



Uncertainy becomes that large, that PIOMAS will all time time within the bands (Min/Max+Unc), i think thats more fair in compare with other products.

For those who may interest in, the .nc-Data also have the variable: "analysis_sea_ice_thickness_unc"

Explantion for the Figure: (SIV)
Mean_2011-18 = mean(thickness)*mean(extent)
Max+Unc= (Max(thickness)+uncertainy)*Max(Extent)
Min+Unc= (Min (thickness)+uncertainy)*Min(Extent)
2019(...)

[Random_Weather]

Also to note, as thicker the ice gets, as greater becomes the uncertainy of
the measurement. Here Scatterplot for the Mean of 2011-2018:
SIT= Sea Ice Thickness
SITU=Sea Ice Thickness-Uncertainy

[Random_Weather]

First of all, happy new year @ all

Nxt Part, SMOS at Melt-Saison:

First some notes:

1) SIT below 5cm is cut out, because there sometimes errors in the mapping at these low SIT
2) Its from 15. Apr to 1.09 Sep 2011-2019
3) SIT= Sea Ice Thickness
4) SIE = Sea Ice Extent

Some claims:
As often told, SMOS is not a physical use of SIT in the melt saison because of melt ponds cause interference which indicate the ice is very low in thickness, on the other hand, we can use this to claim, if SMOS SIT is very low at the begin of melt saison, its build up very much melt power, in other words, it preconditioning the outcome to september SIE.

Thats what i think i found in my analysis.. anyway some figures:

2019,2012 vs mean 2011-2018


Scattern Plot SMOS SIT(1.06 bis 15.07) vs. SIT (NSIDC September)


One more thing: If correlation is used for SMOS SIT after 15.07 the correlation gets near to 0 which also would imply the claims above

[Tom_Mazanec]

Thanks for the thread. Volume is more important than area or extant.

[Random_Weather]

Update on SIV (to 30.12.2019)

[Random_Weather]

With PIOMAS

[Aleph_Null]

With PIOMAS

This one is especially nice. A slight disagreement appears between your index and PIOMAS, as to whether volume is cleaving to the 2011-18 mean. Statistically insignificant, perhaps, but fun nonetheless: red versus grayish blue!

[Random_Weather]

its seem pretty normal, that sea ice volume in PIOMAS becomes greater then in the merged product of SMOS+Cyrosat2, its likely that PIOMAS ends near to the dotted line "Max+Unc"

greets

[Random_Weather]

Update until 9.01

[Tom_Mazanec]

So we are well below this past decade’s average.

[Random_Weather]

Y, past melt saison makes large damage on the thick ice layer, you only need to compare:

2019:
ftp://ftpsrv2.awi.de/sea_ice/product/cryosat2_smos/v202/nh/2019/01/W_XX-ESA,SMOS_CS2,NH_25KM_EASE2_20190103_20190109_r_v202_01_l4sit.png
2020:
ftp://ftpsrv2.awi.de/sea_ice/product/cryosat2_smos/v202/nh/LATEST/W_XX-ESA,SMOS_CS2,NH_25KM_EASE2_20200103_20200109_o_v202_01_l4sit.png

[Random_Weather]

In other words:

mean thickness at this date:

2019: 1.137332m
2020: 1.000787899m

while extent is nearly the same

[binntho]

In other words:

mean thickness at this date:

2019: 1.137332m
2020: 1.000787899m

while extent is nearly the same

That's quite a difference - and surprising, I had the impression that the Arctic had been unusually cold so far this winter. But perhaps not.

[Random_Weather]

As note in literature, thickness is not only a result of temperature but also from compaction, even more above (i hope i remember correct) 2m, the ice itself gets so isolated from Temperatur at surface that 10-20K more or less is no more effects thickning .

And last summer kills a lot of the thickest sea ice

[gerontocrat]

In other words:

mean thickness at this date:

2019: 1.137332m
2020: 1.000787899m

while extent is nearly the same

That's quite a difference - and surprising, I had the impression that the Arctic had been unusually cold so far this winter. But perhaps not.

To find thickness I divide volume by sea ice area, not extent. In my opinion, though not perfect (though my opinion is), it's a better measure.

[Random_Weather]

But is not important here, because SIT is a direct meassurement of every grid over detected sea ice

[Random_Weather]

Here the Histogramms:

09.01.2019

09.01.2020


In 2020, SIT above 3m not exists

[Phil.]

Would be better to show it with the same axes, the first impression is that 2020 is thicker.  Then when you look at the axes you realize what's happening.

[Random_Weather]

Hope that makes it clearer:

[psymmo7]

Much clearer thanks!

[Phil.]

Hope that makes it clearer:

Perfect!

[Glen Koehler]

    It could just be another example of year to year variation, but seen as a trend indicator, losing 13% of thickness over a one year span is ominous.  Really makes me wonder what the upcoming melt season will bring. 

    Apart from the direct effect on total volume, the change in ice quality does not bode well for ice survival.  Maybe the remaining 2020 winter will reverse the striking difference between 2019 and 2020 in fraction of ice over 3 meters thick.  If not, 2020 melt season will start with the ASI preconditioned to match or exceed previous record for amount of summer ice melt, and from a low volume starting point.  It is both interesting and devastating to witness this process unfold.

    The NASA and NOAA GISS temperature observations for December 2019 will come out next week and thus statements about the 2019 full-year global average surface temperature.  By the face value, 2019 will almost certainly rank #2, just a bit (ca. 0.02C) below 2016. 

    But 2016 was a monster El Nino year and at the peak of upward solar cycle influence on temperature.  Filtering out the effect of the ENSO ONI3-4 sea surface temperature, solar, and aerosol influences that cause year to year variation reveals the underlying trend.  Using a preliminary GISS-equivalent temperature for December and filtering out those short-term influences shows 2019 as the warmest year since modern records began.

[binntho]

Using a preliminary GISS-equivalent temperature for December and filtering out those short-term influences shows 2019 as the warmest year since modern records began.

Ocean temps apparently reached their highest levels yet, according to this article in the Guardian.

2019 shows a marked jump on the year before, also interesting is the apparent change in slope after 1990.

[Tammukka]

How it comperes with El Nino years? Quit well I would guess.

[Random_Weather]

@ Glen

Melt-Saison has variance and if we look back over time, its not done yet, in the data i have, a good chance for all volume is gone after melt saison, is when spring volume is below 15km^3.

So we now at 13.37km^3 and usually there will grow arround 4.5 km^3, if so, there have to be the strongest melt saison ever to get this volume down

[Gray-Wolf]

I think we need be mindful that 'melt season' does not only consist of 'in situ' melt but also involves 'export'?

[RoxTheGeologist]

Thanks for the thread. Volume is more important than area or extant.

Well, except for aldebo and the whole Arctic amplification thing.

[Rodius]

How it compares with El Nino years? Quit well I would guess.

Not really.
https://www.nature.com/articles/s41612-018-0020-1

The El Nino years in recent decades have been:
1972–73
1982–83
1997–98
and 2014–16

[binntho]

How it comperes with El Nino years? Quit well I would guess.

If you mean, how well does ocean heat content correspond with El Nino years, then I think that a correlation would be expected: During El Nino years, the oceans do not absorb as much heat in the Pacific, even lose some, while in neutral or La Nina years the oceans absorb more heat. Whether the global ocean heat content reflects this is perhaps less certain, but seing as how the Pacific is a hell of a lot of ocean, one would expect to see some correlation between the Pacific and the global.

Looking at the graph, the years when the red columns are lower than the preceding year are often closely preceded by El Nino events. The 2015/16 El Nino was the biggest so far, the 97/98 the second biggest. The 2009/10 El Nino was significant, while the 2006/7 wasn't really noticeable.

A major La Nina from mid 1999 through 2000 might explain the sharp rise after 2000, while  the the two El Nino events in 2002 (average) and 2004 (small) could explain the "plateau" 2003 - 2005.

Edit: Ocean temperatures and atmospheric temperatures show correlation with the ENSO cycle, but in opposite directions: An El Nino will warm the atmosphere but cool the oceans, and vice versa.

[Rodius]

How it comperes with El Nino years? Quit well I would guess.

If you mean, how well does ocean heat content correspond with El Nino years, then I think that a correlation would be expected: During El Nino years, the oceans do not absorb as much heat in the Pacific, even lose some, while in neutral or La Nina years the oceans absorb more heat. Whether the global ocean heat content reflects this is perhaps less certain, but seing as how the Pacific is a hell of a lot of ocean, one would expect to see some correlation between the Pacific and the global.

Looking at the graph, the years when the red columns are lower than the preceding year are often closely preceded by El Nino events. The 2015/16 El Nino was the biggest so far, the 97/98 the second biggest. The 2009/10 El Nino was significant, while the 2006/7 wasn't really noticeable.

A major La Nina from mid 1999 through 2000 might explain the sharp rise after 2000, while  the the two El Nino events in 2002 (average) and 2004 (small) could explain the "plateau" 2003 - 2005.

Edit: Ocean temperatures and atmospheric temperatures show correlation with the ENSO cycle, but in opposite directions: An El Nino will warm the atmosphere but cool the oceans, and vice versa.

While what you say about what the El Nino and La Nina do heat wise is correct... the rest is a little off.

And I honestly can't see how you can see a correlation between either event and the increase  ocean temps given all it is doing is going up quickly.... it is totally ignoring EL Nino and La Nina.

I linked a study about it one comment up.

*edit* - here is a link to a thread discussing this exact topic... but have a read
https://forum.arctic-sea-ice.net/index.php/topic,2872.0.html

[binntho]

Rodius, I assumed that your earlier answer was regarding any correlation between sea ice and ENSO (which is why I started my answer with a caveat) - and as far as I know, it has repeatedly been shown that no such correlation exists.

And I am not sure what it is in my previous post that you think is a little "off" - variance in global sea temperatures correlate well with the ENSO cycle.

To recap: The ENSO cycle does NOT correlate with variance in sea ice cover, but it DOES correlate with fluctuations in ocean temperatures as well as fluctuations in atmospheric temperatures.

Which is not to be confused with the trend which is relentlessly up where the temps are concerned, and relentlessly towards less and less sea ice cover - and the ENSO cycle has absolutely no effect on the trend.

[Tom_Mazanec]

binntho:
If I read that chart right, your hypothesis says there will be no El Niño this year, right?

[binntho]

binntho:
If I read that chart right, your hypothesis says there will be no El Niño this year, right?

Nothing I've said would indicate any prediction as to any appearence of the siblings this year, whether it's the nino or the nina.  So I'm a bit puzzled about your question. What hypothesis exactly is it that you think you can read from my post?

And I'd like to ask a counter question: Does anybody really doubt that the ENSO cycle will be discernible as noise in a graph showing ocean temperatures? And if so, why?

The trend is a straight line (or as near as makes no difference). So what causes the year-to-year variability if not ENSO and similar cyclic events?

[Tom_Mazanec]

Sorry binntho, I misread your writing as a drop in temperature preceding an El Nino

[Rodius]

Rodius, I assumed that your earlier answer was regarding any correlation between sea ice and ENSO (which is why I started my answer with a caveat) - and as far as I know, it has repeatedly been shown that no such correlation exists.

And I am not sure what it is in my previous post that you think is a little "off" - variance in global sea temperatures correlate well with the ENSO cycle.

To recap: The ENSO cycle does NOT correlate with variance in sea ice cover, but it DOES correlate with fluctuations in ocean temperatures as well as fluctuations in atmospheric temperatures.

Which is not to be confused with the trend which is relentlessly up where the temps are concerned, and relentlessly towards less and less sea ice cover - and the ENSO cycle has absolutely no effect on the trend.

Looks like I misread your words.
Sorry, still, it is a good reminder for me and good to read the material again.

Sorry

[Random_Weather]

Thanks for the thread. Volume is more important than area or extant.

Well, except for aldebo and the whole Arctic amplification thing.

Thats also a way to simply view, ice is a good but not perfect isolator, if ice gets thin enough, energy can be advected to the water below the ice surface, also young and fresh ice in winter times allows atmosphere to interact with the ocean below the ice. Thats why T2m-Anomaly often over fresh sea ice is very large also if the weather pattern would imply cold.

So sea ice thickness is not just about thickness ist also about Arctic amplification