« On the move | Main | A new round of vids »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Artful Dodger

Thanks for this prompt update, Neven.

Just a quick note on English usage: the abbreviation "STD" generally is reserved for various forms of sexually transmitted diseases.

While the sea ice may be rotten, it wasn't sex that caused its disease, it was deviant fossil fuel use. ;)

The generally used abbreviation for "standard deviation" is Std Dev, or more simply SD.

Jim Williams

Lodger, I think it could be said that sex caused the disease. Now I need a fig leaf...

Not exactly a rebound, but volume does seem to have plateaued for a moment.

Rick Aster

“STD” has been used as an abbreviation for standard deviation since at least the 1960s, so it is older than the other meaning mentioned above and I would expect all statisticians to recognize it. On the other hand, for someone like me who works with statistics, you could write σ (omega) for standard deviation and I would understand it without stopping to think. But for the sake of the many readers who may not be well versed in such technical matters, especially this year, it might be worth the trouble to spell out things like standard deviation.

In that spirit, and for the benefit of those who may be scratching their heads, let me try to explain the significance of this statistic in a paragraph. We’re looking at this statistic mainly to answer the question, is the Arctic sea ice in a steady decline, or is the decline accelerating? 1 standard deviation means that the current level is not so far removed from what you would expect, based in this case on the long term declining trend. So the current reading is consistent with the idea of a steady decline in Arctic sea ice. When we go past about 2 standard deviations, as has happened in each of the past three years, that is not so consistent with the expectation of a steady decline — the decline may, instead, be accelerating.


Rick's point concerning the meaning "1 STD" (sexually transmitted or not:) is related to some other head scratching. The 1σ and 2σ bands on the Piomas graphs seem to only have meaning in the context of an assumed linear decline of the ice volume. But if the "true" trend should be represented an exponential or a Gompertz curve, it seems that the bands would be narrower (or am I wrong). Oh well, I do know one thing, σ is sigma, not omega! ;)


David Vun Kannon asked "what does "essentially ice free" mean in terms of volume?"

CT area in million km^2 at minimum divided by average thickness in m at minimum (PIOMAS Volume/CT area) seems to be staying roughly constant at about 1.72.

So if area=1M km^2 and the rough constant of 1.72 remains at 1.72 then CT area/t = 1.72
so average thickness(t) = 1/1.72 = 0.58 m
So volume = 1M Km^2 * 0.58 m = 580 Km^3

Note that 1M Km^2 area is rather more ice at minimum than if Extent = 1M Km^2 which is perhaps a more usual definition of 'essentially ice free'.

(I allowed for this in my calculation on the forum of using 21 K Km^3 volume reduction needed from 31 March rather than 21.612 K Km^3.)


And that makes perfect sense, since mass loss from Greenland last year was in the order of 600 cubic kilometers. You can't expect sea ice volume to go much lower than that.

Cheers P

R. Gates

I continue to watch two different critical time periods for ice thickness and ice volume. For peak thickness I'm watching the period of mid to late April. For peak volume, I'm looking at late May, and more importantly I expect the trend of volume to begin to really begin to diverge greatly from 2012 (unfortunately) around April 15th to 20th, with the line looking something like this:


R. Gates

In my previous post, to make it clear, the chart is of course the calculated average thickness based on the ratio of PIOMAS volume to CT Area, as so it is this ratio or calculated peak in average thickness I believe will begin to diverge from 2012 sometime around April 15-20, and reach a peak in late May before begin the real summer melt decline.

Rick Aster

Tim, a curve is more flexible, so it fits more closely than a straight line, and as a result it shows a smaller (narrower) standard deviation. And, thanks for the correction on the Greek letter name. I don't know how I made that mistake.

Susan Anderson

The bad news just keeps coming in. I noticed what has probably been obvious to you all forever, that the peak volume time has moved forward at least a month and a half since 2005! (from R. Gates figure)

Kevin O'Neill

Over at Dosbat I commented that we may be seeing a ceiling (or floor - depending on how you look at it) on the Max volume. Three consecutive years with almost identical maximum volume numbers.

This may be a quasi-stable value due to the physics and feedbacks of winter freeze under current conditions *or* it may be a PIOMAS error bearing in mind that PIOMAS overestimates thickness of thin ice and basically all we have left is thin ice.


Since the theory of Global warming comes to this: the earth's heat budget is out of balance, with more heat due to insolation arriving than escapes through infrared radiation. Given this more ice volume is a peculiar phenomenon. For since the budget is out of balance the earth should retain more heat this year than last, and lst years heat is still around. As this budget has become further and further out of balance in the last decade it rarely happened. It can only have a few causes.
1) The sun got dimmer either because of what happened on the sun or because of particulate matter in the atmosphere. But this would have to be a large effect actually pushing the budget in the other direction rather drastically. Neither of these things happened.
2) Arctic ice volume ceased to be a proxy for global warming. Glaciers calved so gigantically that the calves became significant in the measurement of arctic ice volume. Don't think so.
3) The ice became so rotten that ice volume was no longer a proxy for ice mass. So that even though ice volume increased, ice mass decreased. It's hard to imagine this happening in the winter.

R. Gates


Not sure exactly what metric you are using for the out of balance Earth energy budget, but the most accurate is the heat content of Earth's oceans which has been steadily increasing for decades.. Some of this heat is going to melt sea ice with a net overall decrease in sea ice volume over the past several decades and of course the reduction of mass in Greenland and Antarctica. Much of this excess heat however remains far below the surface of the ocean--too deep to affect sea ice. Still, overall in terms of visible impacts of a warming planet, the changes in cryosphere remain the most obvious and easy to see.


The southern Beaufort is well along to complete disintegration at an unprecedented point in the season:

 photo beaufortDisintegration_zps4c2d0390.jpg


Mdoliner 43, sea ice volume decrease is not linear. Also in theory there's a negative feedback that should assert itself (more and more) as there is more open ocean at the end of the melting season allowing more heat from the ocean to radiate out into the atmosphere and then into space. First year ice also grows faster than MYI.

But I think ice thickness distribution also plays a role, although Chris Reynolds is the expert on this one, actually looking into it. And ice thickness distribution differs from last year.


R. Gates
Yes you are of course quite right that most of the heat is in the deep ocean, but does the proportion in the deep ocean and that available change. If this proportion remains constant then the amount unavailable to melt ice is irrelevant to my argument.

Yes, Neven I wondered about this negative feedback a long time ago but can't remember anyone coming up with an answer. But I assume that it won't change the imbalance of the heat budget. If it did the earth would stop warming and the deniers would be right. And no, volume does not decline linearly, because there certainly iss some variation in the surplus heat retained each year. But heat always moved from hot to cold and therefore into the arctic. It can only leave the arctic if it leaves the planet, to space, even colder than the arctic. However, not all heat goes into raising the temperature. Melting ice and evaporating water take heat without raising temperature. But any additional heat must do one of these three things for the most part, though there are, of course other things that use heat in small amounts.


>"I wondered about this negative feedback a long time ago but can't remember anyone coming up with an answer. But I assume that it won't change the imbalance of the heat budget."

Reducing the thickness of the ice allows more heat out particularly in Autumn so it does alter the heat budget. So far the heat budget does not seem to have closed, more like it has got worse.

The difference in heat flow between 1 cm thick ice and 21 cm thick ice is much greater than between 21 cm and 41 cm. So the negative feedback gets stronger as the ice gets thinner and should be able to do more to close the heat imbalance.

Removing sea ice from models causes ice to recover in 2 years. Should we expect similar in the opposite direction? i.e. heat imbalance should thin the ice and allow more heat out but how long should this take balance the heat budget?

How does that compare with other changes to heat budget like:
-Change in GHG levels
-More FYI in place of MYI allowing more heat absorption
-thinner ice having lower albedo
-temperature rises causing more heat emission


Big spaces with groups of years having little change just shows that a big "down year" is what determines much of the change. After 2007, there was a rebound 2008, and then a resuming of a down trend in 2009, confirming 2007, and then in 2010 another big down year.

So just consider random fluctuations along the trend, but the trend is still down very much.

One could view "up years" as compensation for the previous years' above average melt, yet a down trend continues.


I have been reading this blog since spring 2011, couldn't rate this site high enough. Commenting here for the 3rd time I believe.

I have a question. There has been much discussion about heat loss as an important negative feedback as ice area shrinks (assuming the process is most significant in the late fall and early winter). How significant is this feedback compared to the positive feedback of less albedo due to loss of area or condition of the ice (cracked ect.)

To answer my question in a way anyone could understand please use this example:

Normal ice area (1920's for example) during spring and summer melt days = 0 units of energy gained as an anomaly

Essentially ice free area (lets say 2018) during July-Oct = 100 units of energy gained as an anomaly (due to loss of albedo)

From that hypothetical 100 units of energy/heat gained per year from an ice free Artic, how much would you subtract due to the negative feedback of increased heat loss?

I know this would change as the oceans warm and the difference between water and air temperature changes through the end of the century.

Artful Dodger

Unsourced opinions are just that. Here are some guidelines for reporting statistics in scientific communication:


Abbreviations for the standard deviation are SD if the scores are from a sample and Σ if the scores are from a population.

Since PIOMAS is a sample, the appropriate abbreviation is SD, however...


Statistical abbreviations (e.g., M, SD) are only to be used within parentheses or at the end of sentences (i.e., when the abbreviation is not being used as a part of speech within the sentence). When the statistic in question is functioning as a part of speech in the sentence (e.g., as the subject of the sentence or the object of a prepositional phrase), then the statistic name must be spelled out as a word and not abbreviated, such as mean or standard deviation.

Since Neven uses STD as the object of the prepositional phrase "into territory" the correct approach is to write out "standard deviation" instead of using the abbreviation.

Can we put this to rest? Let's return to PIOMAS, else open a statistics topic on the Forum.


The notion that radiative cooling to space during winter is actually happening to the degree that it constitutes a negative feedback is still very much up in the air. There are serious papers showing that the temperature inversion caused by loss of sea ice (thick ice) during winter time actually contributes to atmospheric warming, most of the heat is trapped near the surface, and that this is just one more element of polar amplification.

One paper to reference for this view is:


The fact that this amplifying positive feedback may not currently have as strong or dramatic effect on sea ice as albedo loss during summertime does not deny the fact that it actually contributes to overall warming. You can see this process in the high temperature spikes during most of the past winters.

In general, I don't believe this is a negative feedback at all. Just a transfer of ocean heat to atmospheric heat. The result is a net gain for the latter. But atmospheric processes trap enough of the heat for it not to constitute an overall negative feedback and, in fact, it results in volatile warming of the Arctic during winter time.

Overall, I think we've been somewhat enchanted by the rate at which ice volume and extent rebound during winter. But if you look at Arctic ocean and atmospheric temps, you will see that these ice rebounds are not driven by a return to below average temperatures. To the contrary, these temperatures continue to increase and be much higher than normal. The state change, most likely, results from a rapid return to temperatures and low solar radiation states that are low enough to promote sudden refreeze over such a large area. In this case, it's not the end summer ice state that is most important. It is simply the return to low sunlight (the cessation of ocean heating) and to temperatures again low enough to freeze ocean water.

In short, the Arctic summer and winter are two entirely different worlds and should be considered in their own contexts. Loss of winter ice will come with the slower creep of influx of warmer air into the Arctic from the south and via ocean heat transfer.

Chris Reynolds

Robert Scribbler,

During periods of rebound atmospheric temperatures should be higher than at other times precisely because more latent heat of fusion is released, and thinner ice is more prevalent than the thicker ice it has replaced, so heat flux through it is greater.

Here's a few years of NCEP/NCAR surface temperature for the region north of 70degN.

2000 -17.663
2001 -16.640
2002 -16.518
2003 -16.872
2004 -18.333
2005 -16.262
2006 -16.516
2007 -15.853 Peak
2008 -16.743
2009 -16.019
2010 -16.190
2011 -16.626
2012 -15.632 Peak

Note that autumns of 2007 and 2012 both show a peak of warming at the surface, this warming is also apparent a bit further up, but is essentially surface-hugging. This warming is associated with open ocean freezing and forming massive amounts of new ice. The heat that is emitted to the atmosphere warms the air, but the warmer air then radiates more heat energy which is ultimately radiated to space.

To be clear; here we are talking about the thickness/growth feedback, a powerful negative feedback in which thinner ice allows more heat to be conducted from the warmer ocean to the colder surface. The issue of radiation to space increasing with temperature increases is a separate and equally fundamental negative feedback.

I agree that summer and winter are totally different regimes and must be considered as such.

In my March sea ice status update, here, I digress to consider possible implications of the apparent reduction in later winter volume in the last three years.

In a nutshell:

Under current trends the average peak thickness (April) will meet average seasonal thickness loss (April -> Sept) in around 2017, the same date as volume projects using an analogous technique.

However most of early the meeting of the thickness trends is due to winter thinning, not increases in spring/summer melt.

IF the recent three years of winter volume being equal (simplistically implying that thinning my be stalling) is due to ice growth in the autumn in response to low sea ice, this could reduce the winter thinning and result in the following situation:

The volume loss trend could flatten because autumn/winter regrowth could be stronger than spring/summer loss increases.

I am not saying this will happen, I am saying it is a possibility that needs to be borne in mind. By this time next year we should have a better idea of whether this is feasible. But it is feasible that the years to come could see a longer tail than many anticipate.


I believe the PIOMAS max will be reached quite early this year. Just an accumulation of data, nothing concrete to point to, but early/mid March would be my guess. Actually, I think I'm thinking of sunlight, and that in a few weeks this will be 24hrs in places that are already weak.


MAY argh


Alright, alright, I've changed the sentence into: "The anomaly trend line has climbed a bit further into 1 sexually transmitted disease territory."


;-) :-P

John Christensen

With the unusual cold temperatures we saw in Feb/March combined with very extensive cracking, we should expect high volume increase as discussed in the last PIOMAS update.

So now we have the ice, but it is cracked ice, so should still be a more fragile state than last year.

Did anyone speculate on causes for the cold spell in Feb/March?

It is interesting that apparently the low automn/early winter SIA would cause higher air humidity and precipitation causing the very high NH snow cover positive anomaly in Dec 2012 - could this to some extent delay the otherwise warming climate, at least for some winter months in the NH, or were air temperatures as much above normal as usual for recent years??

Chris Reynolds

John Christensen,

To add to my post above: Those cold temperatures may be part of the reason for a failure in volume drop this year. It's worth adding to my post above that CT area peaked at around the same level for three years following 2007 before dropping since.


I can heartily second Artful D's comments. We really need to stay clear of the alphabet soup style of writing -- a sea of parochial acronyms is very offputting to people just coming on to the site.

Climate science is very much cross-disciplinary yet most individuals are not. So yes SD, FYI, CAA, SAR, M. arctica might be widely used in statistics, Arctic ice, Canadian geography, remote sensing, and diatom biology but would they be familiar to an atmospheric chemist or glaciologist?

A post should be self-contained -- spell out the word in is first use. No extra keystrokes are needed -- your word processor can expand your favorite acronyms.

In real life I write incomprehensible academic research articles that I can hardly understand myself after a few years have passed. Here posts need to be understandable by the broadest possible community.



Trend has moved day of maximum from 113 to 105. So early/mid May would seem rather late. About day 100 (10 April) would seem an appropriate day if you expect it to be earlier than the trend. This would probably mean little increase from 21.612 on day 90.

Gain day 60 to 70 = 0.761
Gain day 70 to 80 = 0.451
Gain day 80 to 90 = 0.455

2012 (late max) had 0.858 0.497 0.498
2011 (early max) had 0.602 0.554 0.431

Probably not sensible to read too much into this but 0.455 of last 10 days is nearer the 0.431 of 2011 which had an early max.


One of the key assumptions in the original 2003 Zhang and Rothrock paper describing the Piomass model is wrong.

For clarity, let me repeat that. One of the key assumptions in the original 2003 Piomass model of Zhang and Rothrock is wrong. For this and other reasons, the model is no longer particularly applicable to the ice we have today.

Piomass solves the 'ice momentum equation governed by a viscous–plastic rheology with an elliptical plastic yield curve" as described in Hibler 1979, which in turn is based on the AIDJEX model of Coon 1974. The Los Alamos CICE sea ice model Hunke and Lipscomb 2008 also uses a very similar isotropic elastic-visco-plastic rheology.

At the time, these improved on primitive rheologies assuming a freely drifting, compressible viscous Newtonian fluid with zero shear strength ice pack However VP isotropic rheology cannot and does not capture observed shear, vorticity, divergence, thickness distributions, trends in velocity and export, or deformation-driven ice production.

Looking through the rear view mirror for 'Inside Baseball' statistics will not get at the ice end game at all. This is just 'negotiating' in Kübler-Ross terms -- denial, anger, negotiating etc have no effect on actual outcome.

Jerome Weiss. Drift, Deformation and Fracture of Sea Ice SpringerBriefs in Earth Sciences DOI: 10.1007/978-94-0007-6202-2.5

Jim Hunt

Hi A-Team,

I've taken the liberty of copying most of your post over to the "Distributed Arctic Sea Ice Model" thread on the ASI forum, where some of Danny Feltham's other recent work is already under discussion:


Chris Reynolds


Reduced ice speed implies reduced Fram Flux, which implies PIOMAS over states Fram Export.

Increased thickness of that model implies PIOMAS is under projecting volume.

Both mean that if the authors are correct, the ice cap is in a better state than PIOMAS shows.

It will be interesting to see how this ties up with Laxon et al, because they show that absolute PIOMAS volumes are below ICESat while the change of volume is about half that shown by the difference between ICESat and Cryosat.

Chris Reynolds

PS, ridging in PIOMAS is separately calculated using a parameterisation.

Kevin O'Neill

A-Team, PIOMAS is not simply the result of GCM model runs, but constrained by observations. The assimilation of observational data appears to have a significant effect on the results:

"Note that the Model-Only integration, which does not use assimilation, shows little seasonal variation in the amount of explained variance. In contrast, the IC-SST and IC integrations which assimilate data to constrain the model have lower values of explained variance, particularly in spring and fall. This suggests that the assimilation process introduces a substantial amount of variance into the ice thickness outside the DRA (Sumarine Data Release Area ed.) and that ice thickness variability outside the DRA is much less controlled by the dynamics and thermodynamics captured by the model and forcing data than in the DRA. PIOMAS is apparently getting significant help from the assimilation procedure."

Schweiger et al, 2011, "Uncertainty in Modeled Arctic Sea Ice Volume."

Doesn't the assimilation of observational data negate many of the deficiencies that might exist in the GCM itself?

If we were looking at PIOMAS model output for the next ten years - where obviously no observational data exists - then the GCM deficiencies would be worth considering. But these shortcomings in the GCM used would seem to have little relevance to PIOMAS as a tool to examine the current state of ice thickness/volume.

Aaron Lewis

To extend A-Team's remark above, PIOMAS also does not consider the fact that the ice is no longer as uniform as it was in the past. Prior to 2000, the ice was subject to very slow circulation and mixing. Now, ice that circulates into some parts of the Arctic simply melts.

PIOMAS considers ice fungible. However, today some ice may be partially melted and contain films or pockets of water or air. Thus, while the 2013 and 2012 volumes may be similar, the actual energy required to melt the ice is less in 2013. My evidence for this it the weakness in the ice resulting in this year's cracking events.

Yes, there were some wind events, but the take away was not that the wind was so strong, but that the ice was so weak, e.g., the ice was already well on its way to melting. The extent of rapid cracking says there is a lot of ice that is well on its way to melting.

Dr Tskoul

I agree with Chris. The constraining of the model with observations goes a long way in reducing its problematic rheology.

Dr Tskoul

..erhh "the error introduced by its problematic rheology".


Two points here.

Point 1. Whilst volume is vitally important in measuring decline and the mechanics of decline, surely it has almost no relevance in "ice free" terms? After all, when we are talking ice free we are talking about watts/sqm absorbed by the sea over a period of time, whether it be days, weeks or months. In that scenario, surely, we are only interested in area and no other figure?

Point 2. Whether we get "good news" about the volume reaching, or even exceeding, 2011; is it not a salient point that from September to March volume was significantly below 2011? A late volume surge is of no more significance than a late area surge in the scheme of things, surely? Unless it is a case of 2 year ice being created rather than more area of <1 year ice, then surely all we are seeing is a large volume of very thin ice; which will add to the very rapid April to July ice loss?

If that is the case then we have gained nothing but numbers and the sad story of decline continues unabated and questions as to "Why is this happening in the face of more heat sequestration?" are just misleading.


@John C

“Did anyone speculate on causes for the cold spell in Feb/March?”

Several have asked whether Stratospheric Sudden Warmings (SSWs) are related to cracking events in the Arctic. I think they are and here is why:

As shown previously by R. Gates, major SSWs occurred in 2006 (~ day 30) and in 2013 (~day 40)

This paper - https://www.e-education.psu.edu/worldofweather/files/worldofweather/file/Likovich_etal.pdf - points to a clear link between the onset of SSW events over the past 20 years and a subsequent rapid anticyclone development in the Beaufort Sea. It turns out that SSWs lead to clear skies all the way from the surface to outer space, which immediately leads to a dramatic temperature drop at the surface ( in the order of 10-20 K in a few days).

As I alluded to in a couple of earlier posts, frost cracking requires a sudden temperature drop in order to proceed effectively. According to DMI: http://ocean.dmi.dk/arctic/meant80n.uk.php both 2006 (-20 K from day 30-40) and 2013 (- 10K from day 40-50) saw sudden temperature drops immediately following the onset of the SSWs in those years. It is thus very likely, that this year’s SSW event led to a sudden drop in surface temperature in the Beaufort Sea, which led to frost cracks, which then led to an early ice break-up as documented by A-team and others. However, air temperatures have remained above average during the whole winter, so PIOMAS volume should stay well below average despite the cold spell.

@Aaron Lewis

I simply don’t buy your remarks about melting sea ice at air temperatures somewhere between -30 and -50 degrees C.


"@Aaron Lewis

I simply don’t buy your remarks about melting sea ice at air temperatures somewhere between -30 and -50 degrees C."

Where is it that cold ? Canadian Arctic is under a heat wave +15 C above normal....

Its more like -10 to -20 C now at many locations, and I have shown on my website that ice freezes at -11 C, and -10 C is a key marker when even under thicker ice the bottom ice starts disintegrating. There is also a lot of black ice, thinner and much more vulnerable.

Not counting greater Algae growths needing thinner ice to do so. I have captured the darkening on Barrow strait ice much having same thickness as the overall Piomas calculated thickness.


Seems that my comment went nowhere, will try again,

"@Aaron Lewis

I simply don’t buy your remarks about melting sea ice at air temperatures somewhere between -30 and -50 degrees C."

Canadian Eastern Arctic is under a heat wave with temps in excess of +15 C above normal, hard to believe these -30 to -50 C temperatures. Aaron is very good with sea ice. I have shown on my website that ice freezes at -11 C, and many places in the Arctic are warmer than this temperature.

The comments here about overall thicker sea ice does not make sense unless it was truly colder, but that was not the case for most of the winter past. Ridging from much thinner ice may have contributed to an apparent increase in thickness, so many thousands more leads meant so much more ridging.

Also Algae darkens the ice from under its surface, making it more a black body, as I have captured just recently and showed on my blog. http://eh2r.blogspot.ca/


@ Chris Reynolds

Note that the paper states heat radiated creates an inversion in which most of the warmer air is trapped closer to the surface and not radiated out into space. The paper also notes that this feedback is positive and not negative.

The other mechanisms you mention are not covered. However, a rising winter temperature (both ocean and atmosphere) pushes even winter ice toward deleterious tipping points. So far, the only convincingly powerful negative feedback, in my view, is fresh water infusion.

@ All

Someone mentioned air temps at -30 to -50C. They're not quite so low now. The coldest averages I'm seeing are -20. Ocean temps are a different matter with some regions showing -1.5 to 0.

Chris Reynolds

Robert Scribbler,

This works in the model due to cooler bodies emitting less IR than warmer bodies. It's worth noting the Boe et al find a set of models produce an inversion that is stronger than reality.

However ultimately this heat will not be kept at the surface, it will be radiated to space throughout the winter.

If the inversion keeps the heat there then how come ice growth increases exponentially with greater end of season open water?

Paul Beckwith

Geoengineering replacement of sea ice...Suppose we make gazillions of these balls white and put them in the Arctic Ocean when the sea ice vanishes. Would that restore a more "normal" climate, reduce extreme weather events, and buy a bit of time to slash emissions? Why wait, we can do it now. Is there any law against deployment? http://photography.nationalgeographic.co.uk/photography/photo-of-the-day/ivanhoe-reservoir-los-angeles-ludwig/

Chris Reynolds

Boe et al, 2009, Current GCMs’ Unrealistic Negative Feedback in the Arctic.

Jim Hunt

I keep repeating this in various places, but here goes again.

A-Team is questioning (justifiably in my view) whether either the "modelling" or the "assimilation" in PIOMAS have a good handle on what's been going on for the past few months and hence what will happen over the next few months.

My anecdotal justification for agreeing with him is that recently I listened to Rear Admiral Jon White of the United States Navy say that his branch of the service needed "10 years and more data" to get a decent handle on modelling sea ice. I always thought the US Navy had a better handle on ice thickness than the rest of the scientific community?

Wayne is now suggesting another feedback mechanism which the models don't have a good handle on.

In view of all that is "albedo geoengineering" a good idea, legal or otherwise?

John Christensen

@Paul Beckwith,

The idea itself that we should geoengineer ourselves out of this problem to me, is the whole reason why we are having this problem.

We wanted safe agricultural lands, so we cut down the forest and eliminated a number of wild animal species (at least in most of Western Europe).

We wanted energy so we started burning coal, oil, and then natural gas.

We wanted more energy, so in addition to fossil fuel, we enrich uran, establish wind power, hydroelectric plants with redirection of large water systems, etc. with further significant consequences to the environment.

In addition, we don't like that material deteriorates, so we created plastics.

What these developments have in common is that we never realized the full consequences, or wanted to be bothered with them.

Even with what is unfolding in front of us related to AGW, plastics seems to be a more comprehensive threat to humanity in the long term, as every part of the Earth is being infested with these hormon-impacting molecules that do not break down in centuries.

Geoengineering is just us trying to fix an immediate problem without consideration of the consequences - a perfect example of human intervention.


@Paul Beckwith,

the Arctic Ocean is not exactly a lake, also it leaks into the Atlantic which is connected to all other oceans. So, in order to solve the missing sea ice problem, your idea would cover earth up to 70% with white balls, reflecting enough sun light to turn the planet into :


I don't think your plan was to kill all land based life, have you considered to stop emitting CO2?

Might have far less dangerous side effects.

Jim Hunt

PS - My most recent comment is currently invisible, and John and Noiv's were invisible to me when I wrote it. If I'd seen theirs I would have phrased mine slightly differently!


I'm checking the spam filter as much as I can, it's in my daily routine now. Again, TypePad has switched to a new spam filter system and supposedly it takes time for it to get going.

Sorry for the inconvenience. For some strange reason the spam filter likes to hold on to comments by certain commenters more than others.

Jim Williams

The spam filter is a denier.

Artful Dodger

Robertscribbler.wordpress.com wrote | April 08, 2013 at 06:57

Note that the paper states heat radiated creates an inversion in which most of the warmer air is trapped closer to the surface and not radiated out into space.

Agreed, Robert. Here is some more recent science published on this topic:

Sterk, H. A. M., G. J. Steeneveld, and A. A. M. Holtslag. "The role of snow‐surface coupling, radiation, and turbulent mixing in modeling a stable boundary layer over Arctic sea ice." Journal of Geophysical Research: Atmospheres (2013).

For interested readers, there is also a freely-available PDF. From the Abstract:

... Results indicate a shift in process significance for different wind regimes. For low wind regimes, the model sensitivity is larger for surface coupling and radiation, while for high wind speeds, the largest sensitivity is found for the turbulent mixing process. An interesting non-linear feature was found for turbulent mixing for frequently occurring wind speeds and low wind speed cases, where the 2 m temperature increases for decreased amounts of mixing.

So it seemly very likely that decreased 80N surface temps during the Feb/Mar 2013 breakup episode were the result of enhanced wind-mixing of the surface layer with colder higher air.

If you download the PDF, do have a look on page 11 at Figure 5: "The modeled skin and 2m temperature (K) after 9 h for different geostrophic wind speeds (ugeo (ms1)) using varying turbulence intensities where mixing is adjusted in both boundary and surface layer."



I don't think you can dismiss geo-engineering based on historical data. In fact, i'd suggest the opposite: if it gets as bad as some of the projection indicate, geo-engineering is going to happen and the best choice is to prepare for it.

Why will it happen? Because historically we don't do diddly until we have to. But the time 'have to' comes, what are the world leaders going to do? panic and geo-engineer. So discount it all you want from some intellectual perch but if you want to be productive, start coming up with plans and trying to sort it out in advance.

As far as plastics go, wouldn't sheets of cellulose work just as well but be biodegradable?


@Paul Beckwith,

To me the only acceptable form of geoengineering would be bio-char. It would take CO2 out of the atmosphere and return it to where it belongs - the earth.

John Christensen

The nature is probably the best engineered solution to take care of excess carbon:


But this is a laissez-faire solution, which does not appeal to our problem-solving approach to fix the world, so I agree with the comments that humankind will end up engineering something that will be much less effective than the ecosystem, and will have other negative consequences - like the plastic balls..

Mike Constable

Can you imagine the resources needed for geo-engineering plastic balls etc - and white-anything will provide a new environment for growing something (probably a dark Cyanobacteria!). Even if it contains toxins it would get coated with dark bodies of things it had killed.
I was once puzzled by slight circular depressions in tarmac paths - then I noticed there was not enough traffic on them to stop lichens from growing there. The sun shines, frazzles the lichen, which shrink, pulling up the softened tarmac underneath! Life???

Mike Constable

Making bio-char takes energy, produces CO2 and other toxic wastes - and finally could be seen as a good source of energy if it was too concentrated in the environment!! Coal is better, left in the ground (with no energy consumption used mining it!).
Energy efficiency must be the only way to go to get out our consumption addiction.


Models in climate science have taken on a reality over time for people immersed in them, a reality so powerful that real reality becomes a distant, inconvenient and annoying background buzz, one that can be tuned out by issuing predictions a safe 90 years out or endlessly discussing ill-defined non-observables such as sea ice volume.

This isn't unique to climate modelling -- population biologists 'measure' an allele's fitness as the increased number of surviving offspring many generations out in the future. Not risking any confrontation with data here either, not during your lifetime.

We are primarily interested in the decline of multi-year ice over this coming melt season. For forward prediction, it is much better to have a daily update than being a month and a half behind (end-February Piomass on 09 Apr).

The animation below has extracted what polarized color radar thinks is multi-year ice in the Arctic Ocean from 20 Feb 13 to 08 Apr 13. This corresponds very well with other determinations but has much more detail.

The top of the image is cut off at the point of no return entry to the Fram Strait. Most of the ice in the upper left is caught in the Transpolar Drift current and is headed out the door. Multi-year ice along the CAA coast is mostly sloshing back and forth. Channels and land have been masked out.

It is straightforward to quantitate area and ice class by posterizing and counting pixels in each category. However the Jaxa-Ijis satellite has a hole in coverage inconveniently placed about the north pole and is not available for previous years, whereas Ascat imagery, though grayscale, does not have either issue.

 photo myiMonthB2_zpscb0b2bba.gif


Typepad just *loves* me, never even asks me to read those confirmatory letters any more. Must have something to do with number of previous validations. Or that I never sign out. Could it somehow measure post quality (polysyllabic words?).

About shouting 'fire' in a crowded theatre when there is no fire. That's not covered by the right to free speech -- people could rush to the exit, get seriously trampled. In fact, not shouting fire when there actually is a fire might work better because as others gradually become aware, they exit at manageable rates.

Now what about shouting 'ice-free by mid-Sept' in an April theatre when you're quite sure a big melt season is coming. Rather than a rush to the exits, this might cause a rush to the grocery store -- where we all know there's not enough to go around.

This is less of an ethical conundrum than it first appears because people are so immersed in their chatter, texting and ear pods that they're hardly watching the movie, not worrying about meteors or a couple degrees of warming in 2100, much less melt in some remote location they never planned to visit anyway.

Predicatbly though, come September, they will indignantly ask, 'why didn't you give us some warning'. So it is a good idea to keep putting them on the record, as many people here have already.

Espen Olsen


We already located a hot spot in the middle of North East Greenland:


Espen Olsen


FYI We have just located a "hot"spot in the middle of North East Greenland, now you are warned:



Someone expressed a concern that this site is becoming mainstream-scientific. Never fear, while there are other excellent climate science blogs, these are mostly chasing after the latest press release for the latest journal article. With rare exceptions, those were written 6 months ago -- with luck, the 2012 melt season gets covered. NSIDC provides -- appropriately -- cautious coverage of events and trends of the preceding month, after some perspective has set in.

So that leaves us as providing the sole real-time coverage of breaking events. I for one could certainly improve my posts if I took several months to compose and analyze each, in consultation with a half dozen distinguished colleagues and reviewers. But I don't. It seems better to get a 'pretty good' product out the door and trust that people can fill in the dots for themselves. This is going to be a very fast moving melt season.

Here is the histogram animation that goes with the multi-year ice movement above. Just a teaser, it shows that a random square of 15,000 pixels in the center of the relic multi-year ice is changing fairly systematically in its dielectric constant (salinity, polarized radar reflectance) over the course of 48 days of this spring (20 Feb to 08 Apr).

The drift is giving us some sublte indication of the progression of the melt season. What exactly, that would take broader distribution of effort.

 photo histogramMYIb2_zps226ebbd9.gif

Martin Gisser

biochar prodcution releases energy - energy that photosynthesis had stored when the plant was growing. The resulting char does not rot, thus fixates some carbon - at least for a few centuries. (Proof: Look at pre-Columbian Amazonian Terra Preta soil.) Otherwise the plant would rot and release essentially all its carbon back into the great carbon cycle (as CO2 if not methane).

In North America there's a huge resource for biochar production: Forests killed by bark beetle and/or drought. Not making biochar from this amounts to a continuation of grotesque stupidity (hence to be expected...).

I've done a little experiment and math: Here in Germany, in winter 2011, the "fossil fools cost" of one metric ton of perfect gardening biochar made from wood pellets was -343€/t (incl. VAT). In U.S.$ (VAT substracted, 1.36€/$) -378$/t. Yes: minus. That is, if compared with the cost of home heating oil and 25% energy left unburned in the char.

Meanwhile biochar (Terra Preta) got quite fashionable amoungst German nonstupid farmers. Problem is, where to get the char from. There is one large producer, Carbon Terra who charges +700€/t -- which according to my maths is grotesque. Plus, they don't actually use the energy (in theory it could be used). It's mostly from processing Romanian forestry "waste". At least their system produces clean biochar: no dioxins, no PAH. It seems easy to switch wood pellet home heating systems to produce some biochar. (I'm actually trying to get such a machine constructed. Alas, engineers aren't much interested in this stuff.)

One essential caveat: Don't put fresh char into soil. That would suck up soil nutrients in the first years. The char needs to be preloaded with nutrients, e.g. soaked in CAFO cesspools (or e.g. human urine, as I do it). And don't mix more than 20% vol. into soil - good biochar has an immense water holding capacity, too much of it risks root rot. And of course with too much char the soil could be used for burning. (The suicidal Greenland Norse did that even with pure soil).

Paul Beckwith

Regarding the white balls for geoengineering:

Yes, there are details to be worked out in any scheme. Some have questioned the longevity of the method and whether we can produce gazillions of the balls.

I agree that after a while they would turn blue, shrink and shrivel up since the water is very cold.

Can we make enough? I think that industry could rise to meet the production challenges required. After all, it would look pretty bad for the planet to go belly up simply because we did not have the balls to do anything about it.

The nice thing about putting out ideas for geo-engineering is that people have no idea whether or not you are serious or joking…

In my view, given the rapidity of changes in the sea ice conditions, the only realistically practical method that could be deployed immediately is the anthropogenic Arctic volcano…
We know that it has precedence for working. Physics, specifically conservation of angular momentum dictates that the atmospheric jets slow and go wavy due to albedo collapse in the Arctic, so increasing the albedo there has the best shot at temporarily restoring a more "normal" climate. Whatever method is used, one can only hope for the best and prepare/expect for the worst, at this late stage of collective human lethargy (excluding AMEG people http://www.ameg.me/ and some people here, needless to say). Crops in Britain are failing and there is no end in sight for drought in the US mid-west and now in California. Global grain supplies are near record low amounts (about 70 days supply). Meanwhile methane emissions are sharply rising in the Arctic. etc. etc.


A-team, hi,

Nice pic, as usual! I've got the boundaries of a related area on CAD/ASCAT:

 photo ASCATonCAD010108042013Model1verysmall_zpsa57fd0c5.jpg

The light blue line is 0101 2013. It spanned 1.81 Mkm2. The dark blue line is 0804, holding 1.98 Mkm2.

This ice was what was left of the more or less coherent, "mesh-pattern" pack mid-September. June 2011 I digitized a part North of Greenland and found that just 40% consisted of unified floes larger than 15 km2. The rest of the area was rubble filled leads.
That's why I think just about a third is good MYI now. The rest is MYI rubble frozen in again.
That this area grew a bit during the last part of winter is due to the fragmentation you presented so well.

Fairfax Climate Watch

Don't know if this has been seen by everyone yet, but this video discusses PIOMAS and leads: http://www.youtube.com/watch?v=NwVxxn1-Ll0

Just Testing

Regarding geo-engineering:

We're already doing that. We're geo-engineering a new climate using a few very simple molecules in a distributed setup. It should be apparent to anyone advocating this same mechanism as a countermeasure just how unimaginably difficult it is, how much resources it consumes, how long it takes and how many hidden feedback systems there can be.
And before we try to model a solution with a novel kind of geo-engineering maybe, I dunno, we could try to get a grip on this one and develop models with a margin of error that is less than 5 decades.

"Just reduce CO2" was kinda too difficult, or too expensive, or not profitable, or Obama is a socialist, or it'll delay the new iPhone.

Face it folks, this is old-school Darwinism. And you're part of a gene pool that's just not fit.

Jai Mitchell

With regard to geoengineering. Ice-free June arctic models use +25 w*m-2 and have an increase in longwave escape in autumn (-44w*m-2). If there is an induced temperature inversion that prevents longwave radiation escape then the globally averaged forcing would be an annual average additional +2 w*m-2. This is within the thin tail range of a doubling of CO2. There is a potential that it is much higher on a long-range plot.

Geoengineering with char, white plastic balls (can't imagine the environmental impact on this one btw) or even anthropogenic arctic volcano are inconsequential and would lead to a further delay in mitigation strategies.

I am doubtful that we will reach some kind of equilibrium, 500 mb height anomaly patterns indicate a further weakening of the polar jet which will lead to more subtropical heat transport to the arctic region.

Fairfax Climate Watch

A-Team: expanding on your comments... If you're shouting "ice-free by mid-Sept' in an April theater" and no one's listening, then it's really more like whispering fire in the crowded theater to the guy next to you. In my opinion it's best to make a quick exit before people's pants start catching fire. Nothing says "this crowded room is on fire" like a screaming running man in actual flames. But with competition from Angry Birds and whatever else they're playing on their iphones and Androids now, maybe even a dancing flaming man wouldn't get the audience's attention.

Artful Dodger

Good day, Geo-engineers

While your comments are interesting, this is NOT the appropriate thread for that discussion. This thread is for PIOMAS.

Please continue you discussions in the appropriate place on the ASI Forum:


There already exists a topic titled "Saving the ice cap with geo-engineering? Futile?"

We now return you to your previously scheduled discussion on Arctic Sea Ice. ;^)

Mike Constable

Martin, you point out some of the complexity of bio-char production and use, the original source would be better used as fuel in place of mined coal? Saves transporting it around too.
Jai, I agree geo-engineering just deals with the symptoms, the thought delays action on the problem.
Just Testing, afraid I agree.

The simplest way of extracting CO2 from the atmosphere is photosynthesis, no factories or human chemistry involved! What we have to do is live within our energy supply without dumping CO2 into the atmosphere.(We might be able to turn CaCO3 rocks into bicarbonate and keep the oceans alkaline to help things with shells - but would the dissociation complicate things???)
At the moment aircraft require the high energy content of hydrocarbons, many (most?) other activities can be done with lower grade fuels if we move more slowly?

Jim Hunt

More bad news re the science of Arctic sea ice:


Dr Katharine Giles, a lecturer at University College London (UCL), was on her way to work when she collided with a tipper truck near Victoria Street.

She had travelled to the Arctic and the Antarctic to study the sea ice.

Dr Giles had a "bright future" and was "ready to provide the next generation of leadership" in the field, UCL said.

She was the second cyclist to be killed in the capital this year.

A statement from the head of the earth sciences department, Prof Phil Meredith, said: "Coming so soon after the accidental death of Katharine's own closest colleague, Seymour Laxon, we are all left with a sense of the outrageous unfairness with which some of our best colleagues have been taken from us.

Chris Reynolds

Dr Schweiger has just emailed me (well emailed me on Monday and I've only just got round to dealing with my emails).

PIOMAS gridded thickness data is now available for January to March 2013.

Chris Reynolds

Plots of January to March PIOMAS thickness available here.


Nice work there, Werther. These rubble-filled leads will generate many free floes before very long, greatly increasing side-melt perimeter (and time of exposure) over previous years.

Can the coming export of multi-year ice through the Fram in spring and summer 2013 be understood by projecting forward behavior of the ice from yesteryear? No. We had a record low of this ice left in Sept 2012 plus its character has evidently changed.

What's happening here is not some malarky from catastrophe theory but rather a simple crossing of a threshold: the ice is now too thin, too rotten, too salty and too warm relative to the same old forces acting on it, so it fails.

It is no different than a hiking trail -- thousands can walk along it with very little impact but a single horses can do huge damage because the pounds per square inch from a horseshoe exceeds the material strength of the soil.

I ask that people refrain from invoking bifurcation theory here unless they can demonstrate a research Ph.D not just in pure mathematics but specifically in geometric topology, plus firsthand familiarity with Thom's old papers. Anything less, you are just blowing smoke.

The animation below defines the flowline of theTranspolar Drift over the period 01 Feb 13 to 06 Apr 13. If this persists out into May, we'll have a bit of predictive handle on the extent of doomed ice.

The final frame shows the ice to the 'left' of the north pole moving as a consistent velocity stream, most of it destined to exit the Fram unless something changes.

However the adjacent multi-year ice -- over to the archipelago coast -- just sloshes back and forth with little net effect. This ice is not headed out the Fram as things stand but will experience viscous spreading later in the season as buttressing first year ice melts, putting older ice out in the domain of the Transpolar Drift and at risk for export depending on time available before fall freeze-up.

 photo blueFram2B_zps89796605.gif


March Volumes from cell thicknesses:
2013 2012
20.73 20.69
This is presumably average for March, so very similar and could be less at end of March based on daily numbers.

A quick analysis:
Volume in excess of 0.5m thick in any cell
2013 2012
14.1 14.0
Very similar

Volume in excess of 1m in any cell
2013 2012
8.4 8.6

Volume in excess of 1.5m in any cell
2013 2012
3.9 4.3

Volume in excess of 2m in any cell
2013 2012
1.4 1.6

Volume in excess of 1.67m in any cell
2013 2012
2.8 3.25

Volume got down to 3.26 last year equivalent of melting 1.67m from the March thickness of each cell. Not sure how good a method it is to apply that to leave 2013 minimum volume at 2.8 K Km^3.

Anyway, the trend of less ice volume in the thickest cells continues.


A map of thickness differences between March 2013 and 2012:



Thanks, Wipneus!

I take it that corresponds to this entry on the Forum:


Things are becoming clearer concerning the prognoses for the melting season!


Thanks, Wipneus!

That corresponds with this entry on the Forum:


Things are becoming clearer, concerning prognoses for the melt season!


Chris, is it feasible to post Piomass imagery as plain png -- and without the hideous green coordinate system?

The tiny ice depth scale rectangles are so dithered that ice depth cannot really be read off the image as things stand.

I would like to overlay satellite imagery on the Arctic Ocean portion of it. But here is what the image looks like now in the Bering Strait region:

 photo piomassFix_zpsfe4c9204.png


Wipneus has got this right on the March 2013-2012 difference map and done something interesting beyond.

First note the png is not dithered. That means clicking on a key color with zero color space tolerance cleanly picks out all the ice of that color and no other.

Second, deleting each color in turn should completely empty the map, ie all the ice is of one of the designated thicknesses and no other.

What happens however is the contour boundary lines are left, a good thing because if the contour lines are not wanted the tolerance can be raised slightly, as would be done for computing areas by pixel count.

The legend might include pixel percentages (~ area) in each category as a convenience to the viewer.

 photo piomassDiff_zps3167e756.png


Here is an amended key to Wipneus' difference map for March 2013-2012 Piomass ice thicknesses, as graphic and comma-delimited text. This is the reason for not dithering scientific images.

 photo piomassMarDiffPixels_zps5e0afe70.png


Shared Humanity

Werther......"The light blue line is 0101 2013. It spanned 1.81 Mkm2. The dark blue line is 0804, holding 1.98 Mkm2."

Looking at this, it seems to me that more of the MYI was pushed into the Beaufort Sea by the gyre than was transported out the Fram?

Chris Reynolds


Sorry but I auto-save off Excel and I use jpg as they're not really meant for what you're doing. They're normally scaled down on my blog.

Here's a png.

It does reduce smear, but it's still not as good as having the actual plot in Excel, which is what I use for anything but rough estimates of thickness, for a start in Excel the mouse cursor tells you the thickness when hovering. I'd normally run code on the source data to get any quantitative numbers.

Furthermore, these have been scaled manually by clicking on right lower corner and dragging. Scaling by setting the plot size produces a skewed map due to the presence of the legend and title.

Chris Reynolds

Typepad is eating my words again!

Spreadsheet with sea ice broken down into thickness contributions from 25cm thickness categories available here:
Updated to March 2013. Goto cell CI:74 for a user selectable (drop down list) set per month and region for all years.



I have this data organized in neat arrays in an R environment.
If there is anything you would like to be calculated or plotted I can see what I can do.

I have said it before, anyone who interested in the code can contact me "wipneus on freenet dot de".

BTW, the map uses a perspective mapping, perspective from 3 radius-earth distance. That is probably not what you expect from a scientific image. Shall I change it to stereographic?


I think Chris Reynolds is using Excel to generate his images. Not sure whether there is any good way of using Excel to generate undithered image. Each pixel has a different area so it wouldn't be too difficult to calculate areas and volumes much better from Excel than by counting pixels off an image.

Chris Reynolds

So Typepad ate my last two replies, I'll try again.

A Team,

Crandles is correct, I use Excel.

Here's a PNG:
Actually I might redo the lot in PNG. Still dithers though. By 'green coordinate system' do you mean the land? Yes I can take that off.

But I really think the best thing to do is to get to grips with the base data. When I have the time I'll be writing a piece on the forum about doing that in Excel. Using images as an intermediary is really not satisfactory.

BTW my images are polar stereographic but they're scaled by hand so aren't really to be used for what you're doing. The only way around that would be for me to remove the legend and make the title an overlay.


Evening SH,

If you extract that conclusion from my ASCAT pic post above, yes, a larger part was pushed into the Beaufort direction.
Don’t attach too much importance to the numbers. They’re just for comparison and have no true meaning.

But if we’re talking comparative numbers… CAD shows the "area" mainly lost by the Fram Strait-express as app. 330K. The circulation/fragmentation into the Beaufort direction 499K.

Keep in mind that the large part is still bound within the CAB. Just 104K is actually in the Beaufort Sea now:

 photo PackonArcticSeas11042013verysmallb_zps9119d919.jpg


Chris, Wipneus -- thanks so much for the clarifications. If Wipneus could please post the stereographic png, not of the 2012-13 difference but just the March 2013 thicknesses, same cell style, that would be very helpful.

For most purposes, I wouldn't worry about a few percent error here and there given the ambient overall degree of departure from true conditons (accuracy).

For example, are the billions of little brine channels part or not part of Piomass volume? Does all ice, regardless of temperature and temperature gradient, count the same toward volume? Off by a meter in some places? -- well, that is a colossal error.

It is not feasible to recast thousands of satellite images as excel spreadsheets in the Piomass curvilinear coordinates without critical blurring of already borderline resolution -- note the image release ratio to Piomass releases is ~900:1.

Since it is imperative to integrate all available significant data, better to export Piomas out into the real world.

I think you will be quite shocked to see what the observational ice-penetrating radar looks like when we finally get it overlaid on a high quality Piomass display.

Has no one ever done this in all the decades of Piomas releases? Totally bizarre -- it was their job all along to provide overlay-compatible releases.

It makes me think something very fishy is going on here. I've seen that before: theorists sidestepping comparison to observation.

And please, what date corresponds to a Piomas 'March' release -- first/last/average? March is a time interval, not a time.

The point here that Jaxa-Ijis imagery, which only started last summer, sits in a profoundly deeper informational universe than the old scatterometer reads.

Models have killed us; observational data has to call the shots from here on out. No more tail wagging the dog.


There's an extra PIOMAS update, but folks can continue discussing here as well.


I'm not sure that everyone above is aware that Photoshop etc consist solely of a pleasant front end to a spreadsheet that conducts all its operations behind the scenes. Which is exactly where you want it.

Anything you can usefully do to an array or stack of arrays in excel can be -- and often is -- simply ported to a Photoshop script registry. Any rectangular numeric array can be re-normalized, one way or another, to an image.

Any image format can be deconstructed into a stack of Photoshop spreadsheets. Any reasonable combination of arithmetic and logic operations eg convolution can be performed and captured on any combination of layers without leaving Photoshop.

Mouse over in excel to see ice depth? That's called the histogram eyedropper. Only it's a whole lot better -- you can mouse over a cell (or any describable selection of cells) and see ice depth simultaneously for the three years of your choice plus the average, std deviation, mean and median. Plus capture all that as standalone new array products with jone click of that mouse.

I'm seeing Piomass arrays with a preposterous number of significant digits -- six when charitably there might be two. Thus nothing is lost going to 8 bit imagery, no need for the 16. There rarely is.

The Arctic Ocean is intrinsically a geospatial array. As is all the data, including Piomas. Since there is way too much data volume, the whole issue boils down to cognizant compression.

Sea ice algae? Tundra veg changes? That's why they put 36 wavelength channels on Modis. Five years of that ... you want 5x365x36 = 65,700 excel arrays at 2,250,000 cells each? Good luck with that and bon voyage.

People tend to pooh-pooh the human brain as obsolete technology but actually it is quite extraordiary at making sense of image arrays, especially moving and multi-channel. I'll get more out of Modis just watching the movie -- kick back and let the brain sort through the clutter on auto-pilot, it spent the last billion years learning how to do just that.

It is for these synergetic reasons -- compression enabling intuitive wheat-from-chaff interpretation -- that the Photoshop user always run circles around the Excel user: faster, better, deeper. Some days I put out 4-5 pretty decent products -- try that in excel. I'm not seeing the posts.

Artful Dodger

Hi A-Team,

Okay, I will be the first one to say this. It is time to revisit your forum nickname.

I think from now on it should be A+Team.

EXCELLENT rant! Bravo! Encore! :^)



Okay, I will be the first one to say this. It is time to revisit your forum nickname.

I think from now on it should be A+Team.

Because of the name and the style I always see this image of John 'Hannibal' Smith in front of me:

If he keeps going like this, I will have to ask him to marry me at one point. Hope my wife (and his) won't mind. :-)

Great stuff, A-Team!

Chris Reynolds


What interests me is process.

It might take me an evening or two to get the meta-data into a spreadsheet so I have numbers I can analyse. Then it can take literally weeks of messing about with alternate interpretations, checking papers, doing something else while my subconscious works on the problem. I have about three posts in this process at present.

I think images alone are limited. For example, can you do an image which tells us if we face a long(ish) tail, or a rapid crash to zero area?

I suspect the answer is already in the data, the problem is my limited mathematical ability.



"I suspect the answer is already in the data, the problem is my limited mathematical ability."
A-Team did put it nicely above - the human brain is tailored for images. It is normaly not tailored for numbers. That is why you think a fit would tell somthing - it makes just an image to help your brain. If there is no model to test, a fit is as good as any other line drawn in a brain-friendly way between the data points. There is no other information in that line but brainfriendlyness. The two only scientific reasons for a fit are either to cancel a hypothesis by reality or to get another number from a model which is allready prooven long ago by other means.

The pictures are better for the brain, because they are 5D - 3 colors and 2 axis. And putting images together in time you get 6D for a movie. The math is the exactly same for images as for lines and numbers, but the brain gets much more information from images.

To get to your question "tail or no tail": I have not seen a model description here resulting in a tail so there is no need for a Gompertz-fit to learn something. But statistics could do that tail-trick easily instead. You can see it in the movies: The pack moves some days and stays there some other days. If you put it in numbers, you call that "noise" and "trend" and things are complicated. If you watch the movie (while knowing live and weather...) you understand - you integrate the motion and you can predict easily, where it will go.

To conclude - if there will be a tail, it is because of noise. Albedo feedback results in exponential decline. Area can not get below zero. It will get zero, when volume gain in winter gets maximum. There are things are cold winters and cold summers /ENSO-things and so on. So - after area will have become zero, there will be some years again above zero and others hitting zero earlier. Your data will result in a tail by averaging a positive definit value at a specific time. Your brain instead would integrate that movie to an understandable crash-behaviour because it can compute time and area in parallel.

Steve Bloom

A-Team, to give the Piomass team a bit of a break, their funding is probably pretty limited and the satellite verification of their method being so recent probably resulted in a long list of tasks. But maybe cut to the chase and just get in touch with them about this specific issue?

Chris Reynolds


I disagree about least squares fits, they often differ from the line of best fit I would draw because, for example, they don't weight recent change as much as I would. When fitting one is faced with a variety of base functions to use, the question then becomes what basis function to use and why.

The brain is known to be easily fooled when looking for pattern, this is why numbers are better.

"To conclude - if there will be a tail, it is because of noise."

I disagree, tail or no-tail are outcomes of the interplay of positive and negative feedbacks. In the simple model I've been playing with it is the summer loss / winter gain in ice that determines tailness, noise in sea ice area doesn't make or remove a tail. I will blog more on this in due course.


"the question then becomes what basis function to use and why."

In physics we use least square fitting to data to test, if a model function could describe the data or not. If the sum of the error squares is similar to the degree of freedom (number of data minus fitting parameters), the function is a possible description. There is no way a proof, that this function ist the "real on", because there are endless numbers of other function which could be fitted as good. Therefore, fitting is a tool to ask, if you have to refuse a function or if it would be ok to use it. By no means you can get an answer from fitting if the function is the right one. It may just be a possible one.

In the case of positive definite data like ice volume or area it is dangerous to apply simply fitting ignoring that boundary condition: A tail or a base line are very similar in that case to describe that "positively cut noise" - the covariance matrix is large and the probability to exclude your test function gets low. So it will be very hard for you even in 10 years to exclude the exponential, which is still the most simple function based on an axcepted feed back model. That simplest function should be excluded first before trying to exclude all other functions without any need. That just makes no sense to me.


I am sure that the quest for modelling the september minimum volume trend is completed since last year. The next goal is to decribe winter/spring volume trends.

Using the exponential and including a base line of e.g. 2.000 km2 gives >95% confidence until either
1. Sept. volume will stay significantly >4.000 km3 for some years or
2. the big error of the PIOMAS modelling is reduced by using volume data from A-Teams images using e.g. JAXA data. Unfortunately it will take 5-10 years until we will have the data to do the fitting test.

So - for the next 5 years it is unlikely to proof me wrong. And if Sept. volume really will increase in future, the Gompertz and all the other decreasing functions are out, too. Therefore, a quest for a better function describing sept. minimum is futile and a waste of time.

Instead, we should concentrate to find a simple model description of winter and spring volume for the time after max. volume gain is reached (about now). Will it stay stable or how will it decline? If we are in a transition time, it can not be stablelized allready now - so the question left is actually, how fast will max. volume decline in future?

The comments to this entry are closed.