PIOMAS vs CryoSat

[uniquorn]

CS2SMOS in this thread or preferably in the CS2SMOS thread.

A quick search didn't find the CS2SMOS thread in arctic sea ice

Comparison of PIOMAS and CS2SMOS, mar17-apr15

[Jim Hunt]

P.S.

https://GreatWhiteCon.info/2021/03/the-2021-maximum-arctic-sea-ice-volume/

[gerontocrat]

Here is a graph and maps prepared by the Polar Science Center comparing April 21 PIOMAS with CryoSat

Cryosat shows a tendency to decline while PIOMAS has it pretty steady since 2012,

[Paul]

Here is a graph and maps prepared by the Polar Science Center comparing April 21 PIOMAS with CryoSat

Cryosat shows a tendency to decline while PIOMAS has it pretty steady since 2012,

Even though there is different results, the overall trend is agreed with the volume anaomolies of thicker than normal ice over the Siberian side and below average thicknesses along the Atlantic front and the CAA.

Any positive or negative anaomolies near any landmasses I take with a little bit of skeptism due to overdoing compaction or if ice is moved away from the coasts and there is open water then of course it will show a negative anaomoly.

[WildFit]

Here is a graph and maps prepared by the Polar Science Center comparing April 21 PIOMAS with CryoSat

Cryosat shows a tendency to decline while PIOMAS has it pretty steady since 2012,

PIOMAS lives from its reputation and by now is way overrated.

IMO it's obvious that each year they are increasingly wrong. Probably they apply old
algorithms to a totally different situation for the purpose of comparability while I'd vote
for more accuracy over comparability.[/size][/font]

[gerontocrat]

Here is a graph and maps prepared by the Polar Science Center comparing April 21 PIOMAS with CryoSat

Cryosat shows a tendency to decline while PIOMAS has it pretty steady since 2012,

IMO it's obvious that each year they are increasingly wrong. Probably they apply old
algorithms to a totally different situation for the purpose of comparability while I'd vote
for more accuracy over comparability.[/size][/font][/color]

cryosat don't give us data files of numerical data day by day by regional sea (thanks Wipneus)

cryosat gives us nowt during summer
10-May-21: NRT Service Suspended during Arctic summer (May-Sept).

PIOMAS claim the model is tested against available actual readings and incorporates a range of data.

Nevertheless you may be right. Things may have changed so much that the algorithms are less able to reflect reality. But where is the alternative? A pretty picture is not enough.

PIOMAS has been extensively validated through comparisons with observations from US-Navy submarines, oceanographic moorings, and satellites. In addition model runs were performed in which model parameters and assimilation procedures were altered.  From these validation studies we arrive at conservative estimates of the uncertainty in the trend of  ± 1.0 103 km3/decade. The uncertainty of the  monthly averaged ice volume anomaly is estimated as ±0.75  103 km3. Total volume uncertainties are larger than those for the anomaly because model biases are removed when calculating the anomalies. The uncertainty for October total ice volume is estimated to be  ±1.35 103 km3 .  Comparison of winter  total volumes with other volume estimates need to account for the fact that the PIOMAS domain currently does not extend southward far enough to cover all areas that can have winter time ice cover.  Areas in the Sea of Okhotsk and in the Gulf of St. Lawrence are partially excluded from the domain.  Details on model validation can be found in Schweiger et al. 2011  and (here). Additional information on PIOMAS can be found (here)

A comprehensive library of sea ice thickness data for model validation has been compiled and is available (here)

[WildFit]

cryosat don't give us data files of numerical data day by day by regional sea (thanks Wipneus)

cryosat gives us nowt during summer
10-May-21: NRT Service Suspended during Arctic summer (May-Sept).

PIOMAS claim the model is tested against available actual readings and incorporates a range of data.

Sure, I said nothing about Cryosat because I don't have sufficient input about their results while I observe PIOMAS-Results and compare them with sat-images and other factual data since many years and always see things that are simply not right, wrong by too far. On the other hand, whenever it was hinted at, comparability was the killer argument but then, comparing wrong results with each other does not seem to be very useful.

You know, without science we would have no clue, hence science is important an a good thing.

But then sometiimes things become similar to religions where mentioning flaws is kind of heresy and therefore it takes too long to do a proper reassessment of formerly accepted norms.

However perhaps it does not matter much for now but one day we shall wake up with a big surprise because some greater area of thick ice disappears over night (because it was not that thick obviously)

 

[Paul]

Here is a graph and maps prepared by the Polar Science Center comparing April 21 PIOMAS with CryoSat

Cryosat shows a tendency to decline while PIOMAS has it pretty steady since 2012,

PIOMAS lives from its reputation and by now is way overrated.

IMO it's obvious that each year they are increasingly wrong. Probably they apply old
algorithms to a totally different situation for the purpose of comparability while I'd vote
for more accuracy over comparability.[/size][/font]

So is it wrong in terms of overdoing thickness of the ice or underdoing the thickness or even both?

PIOMAS is not perfect but for the 2018 and 2020 melt seasons, the distribution of ice volume seemed to of added up in terms of what happened in those melt seasons. For example in 2018, PIOMAS had the Siberian side thicker than average and the Beaufort lower than average and despite alot of heat on the Siberian side, melt was quite slow, especially the ESS. The Beaufort was slow to melt but onmly because it was under favourable conditions but despite this, large portions of ice did eventually fragment and melted which indicated in 2018, if the ice was hit by warmer weather  then it would of been quite rapid potentially.

In 2020, PIOMAS had the total opposite anaomolies, the Siberian side was below average and the Beaufort somewhat above average and lo behold, the Siberian side melted at rapid pace(you can even see it on Worldview during May how weak the ice looked) and the Beaufort was alot more resilient to melt away. This year, the Siberian side does look thicker than average, probably not as thick as. 2018 and I accept the deep red anaomoly over the ESS could be a bit overdone in terms of compaction against but looking at the weather patterns we had this winter, it seems to make sense too me the Siberian side should have positive anaomolies and near the CAA and Atlantic front negative anaomolies.

Crysosat meanwhile was picked up by Neven and others of a potential flaw of snowcover perhaps fooling the sensor as thicker ice. It does not make meteological sence too me that a stormy warm 2016/17 winter has more ice volume than a colder(although nothing exceptional and does not include the Atlantic section) drier winter of 2020/21. The only conclusion I can think of there is not as much snowcover on the ice pack this year and definately not as much as in 16/17.

I do favour PIOMAS over cryosat but I do accept the flaws of resolution being quite low and perhaps overdoing ice thicknesses near any landmasses.

[wdmn]

Beyond individual melt seasons the important thing about volume is the area that doesn't melt out. It is a measure of how vulnerable that ice is becoming.

Thickness/volume anomalies in places where the ice melts out are only relevant in that it will take more/less energy to melt that area out this year than normal. The longer term consideration of the ice is entirely about the area that does not melt out each year.

Area/extent are better measures for single season.

Thickness/volume is the better measure for long term analysis, but only if we focus on the areas that don't usually melt out.

Sorry to state the obvious.

[oren]

PIOMAS might or might not be overrated at this point. But looking at the comparison map between the model and the measurement of Cryosat (which also involves some models of course) I am struck by the good match. Thinner ice and thicker ice in more or less similar locations. Truly the differences could have been much worse. So I see this as actual validation of PIOMAS, which then is very useful during summer when the measurements become unavailable.
No surprise that Cryosat doesn't see what PIOMAS sees on the Siberian coast, as I doubt Cryosat can measure pressure ridges. So if they are there or not it's hard to know, but in all probability they are there IMHO, and a lot of volume could be hidden in them to compensate for the differences between PIOMAS and Cryosat. The ice movement against the coast tends to create these ridges, and according to Wikipedia they can be up to 40% of total ASI volume.
OTOH I don't think this matters too much, some remnant ice area might remain thanks to these ridges, but this will not delay the advance of the melting front into the CAB anyway because their total area is not very high. In addition the ridges are deep in the water and once the fast ice breaks and ice movement occurs on a large scale, I believe they would tend to a higher melt rate than normal flat ice.

[interstitial]

Piomas was completely wrong about thick ice near the pole when the mosaic mission visited. It shocked scientists and has lead to efforts to develop new algorithms from CryoSat Data. CryoSat algorthims are still a work in progress but are a direct result of Piomas failure.

[Paul]

Piomas was completely wrong about thick ice near the pole when the mosaic mission visited. It shocked scientists and has lead to efforts to develop new algorithms from CryoSat Data. CryoSat algorthims are still a work in progress but are a direct result of Piomas failure.

I don't remember there being thick ice near the pole on the anaomoly charts, the central basin tends to show average thicknesses(give or take either way slightly) but last year, I remember there being alot of red against the CAA/Barants sea* and the Beaufort sea and alot of blue in the Siberian seas and looking at the sat images and how the melt season went, then it seemed pretty accurate.

* Note the red in the Barants sea does not fully mean thick ice, it more likely because in the last 10 years, this area has struggled to gain and thicken ice especially the area to the north of Svalvard so any positive anaomolies here I tend not to focus on too much.

[HapHazard]

Piomas was completely wrong about thick ice near the pole when the mosaic mission visited. It shocked scientists and has lead to efforts to develop new algorithms from CryoSat Data. CryoSat algorthims are still a work in progress but are a direct result of Piomas failure.

This is essentially correct, and there was some talk about it here at the time.

[vox_mundi]

Arctic Sea Ice Thinning Faster Than Expected
https://phys.org/news/2021-06-arctic-sea-ice-thinning-faster.html

Sea ice in the coastal regions of the Arctic may be thinning up to twice as fast as previously thought, according to a new modelling study led by UCL researchers.

Sea ice thickness is inferred by measuring the height of the ice above the water, and this measurement is distorted by snow weighing the ice floe down. Scientists adjust for this using a map of snow depth in the Arctic that is decades out of date and does not account for climate change.

In the new study, published in the journal The Cryosphere, researchers swapped this map for the results of a new computer model designed to estimate snow depth as it varies year to year, and concluded that sea ice in key coastal regions was thinning at a rate that was 70% to 100% faster than previously thought.

They found that the rate of decline in the three coastal seas of Laptev, Kara and Chukchi seas increased by 70%, 98% and 110% respectively, when compared to earlier calculations. They also found that, across all seven coastal seas, the variability in sea ice thickness from year to year increased by 58%.

... "Previous calculations of sea ice thickness are based on a snow map last updated 20 years ago. Because sea ice has begun forming later and later in the year, the snow on top has less time to accumulate. Our calculations account for this declining snow depth for the first time, and suggest the sea ice is thinning faster than we thought."

To calculate sea ice thickness researchers used radar from the European Space Agency's CryoSat-2 satellite. By timing how long it takes for radar waves to bounce back from the ice, they can calculate the height of the ice above the water, from which they can infer the ice's total thickness.

In the new study, researchers used a novel snow model previously developed by researchers at UCL and Colorado State University, SnowModel-LG, which calculates snow depth and density using inputs such as air temperature, snowfall and ice motion data to track how much snow accumulates on sea ice as it moves around the Arctic Ocean. By combining the results of the snow model with satellite radar observations, they then estimated the overall rate of decline of sea ice thickness in the Arctic, as well as the variability of sea ice thickness from year to year.

"Faster decline and higher variability in the sea ice thickness of the marginal Arctic seas when accounting for dynamic snow cover", The Cryosphere (2021)
https://tc.copernicus.org/articles/15/2429/2021/

[josh-j]

On the face of it then, both PIOMAS and Cryosat volumes could be overestimated?

[Killian]

On the face of it then, both PIOMAS and Cryosat volumes could be overestimated?

Pretty much guaranteed. I wonder if this new data can be used to revise historical data or will be used to simply adjust now and keep going, like adding satellite data to paleodata, e.g...?

[oren]

My concern is that the new calculation is also based on a snow model, though a much improved one. Models in the Arctic are suspect as actual data is severely lacking and the climate and environment are changing rapidly. I would love to see some in situ confirmation of the calculations by way of some local measurements.

[Steven]

The PIOMAS website has a comparison of October 2022 thickness anomalies for PIOMAS vs Cryosat:
http://psc.apl.uw.edu/research/projects/arctic-sea-ice-volume-anomaly/

[oren]

Interestingly, both PIOMAS and Hycom place the thicker ice as adjacent to the CAA, while Cryosat places it around the Pole. Cryosat should be more correct as it is a measurement, however as noted above it also uses some modeling to determine the measurement. So it's difficult to know the actual state of the ice.
A mystery.

[Glen Koehler]

     A mystery indeed.  While the PIOMAS and Cryosat values are reconcilable on the Russian side, on the North American - Greenland side they are not only different but in places in opposing directions relative to the 2011-2020 average.

[oren]

Actually, after thinking on Steven's stats of Nares export, and the difficulty for PIOMAS to model Lincoln Sea to Nares ice movement due to coarse grid resolution and NSIDC artifacts, I think Cryosat is surely more correct this year in the Lincoln Sea. We had an open Nares all year and nonstop export, and yet PIOMAS maintains a high volume there, which I am sure is much lower in reality.
The only way for PIOMAS to get rid of this volume anomaly is if the ice area in Lincoln disappeared entirely for a few days, or is massively exported north and east, as in such a case PIOMAS recalibrates itself using NSIDC concentration data. (Hope my thinking is correct here).

[Steven]

Comparison for mid-January 2023.  The first image shows the thickness anomaly for PIOMAS, the second image for CryoSat-SMOS .  There are still some major discrepancies between them in some regions.