« Record dominoes 11: NSIDC September SIE | Main | More vids »


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


And the anomaly along the Norwegian coast is negative.

Jim Williams

Espen: "I[t] may be the Gulf Stream changing direction?"

Probably not, but given the way everything is changing now who can say?

There's been a bunch of things in the last year or so about the Labrador Current losing energy, and the Gulf Stream briefly flowing directly into the West Greenland Current.

The passages between Canada and Greenland are shallow, so only surface water could flow north in that direction....for whatever that's worth.


Espen & Jim: Precision please!

The Norwegian Coastal Current actually shows a positive anomaly, whereas the threshold between Labradoer Sea and Baffin Bay is close to 500 meters below the surface.

Cheers ;-)


I think Jan 2011 had a period when at least a portion of the Gulf Stream was running west of Greenland producing extreme highs temps in southern Greenland and Canada,



A guess at what record 13 might be: http://barentsobserver.com/en/arctic/towards-cargo-record-northern-sea-route-10-09

In other speculation: given that we're starting the freeze this year with 19% less ice by volume than we did last year, and that extent seems to be shrinking in a non-linear fashion, what are the odds that 2013 will have an even lower extent than 2012?

Jim Williams

I have this from Jan 2010 Terry: http://www.dailykos.com/story/2010/01/06/822520/-Freak-Current-Takes-Gulf-Stream-to-Greenland

Do we have another from 2011? (Once is happenstance. Twice is coincidence. Three times is enemy action.)

Jim Williams

P-Maker, Nares strait is about 360 meters. I'm less sure about the Archipelago but at least some of it is about 100 meters.

Are there any 500 meter deep openings to the Arctic west of Greenland?



You are right, I meant in the second lane (Fast Lane) of the coast of Norway!


2011 was the correct year, but it's possible that there were other reasons for the heat wave.


I remember them closing the schools, and record heat in Greenland at the same time. I't quite possible that I just assumed this was due to diverted gulfstream.


It was this incident, and the fact I could go outside at night comfortably in a Tee shirt (in Canada) that first peeked my interest in Global Warming - took a week or so to find this blog, and it's all been good since then.



The situation in Jan 2011 in West Greenland still shows up in the Baffin Bay longterm sea ice graph here...


...as by far the biggest anomaly of all time.

In Jan 2012, Baffin Bay is back to normal, but at the further Eastern edge of the Atlantic sector, this happens...


...a huge negative anomaly in the Barentsz sea.

Suggesting perhaps that the Gulf Stream may be meandering more? Delivering heat to the West of Greenland in 2010/11, and to the Barentsz Sea in 2011/12.

If so, then this coming winter could bring EITHER a repetition of of winter 10/11, bad news for Greenland, OR a repetition of winter 11/12, bad news for the sea ice in the Barentsz and Kara seas.

At the moment, based on the temp maps referred to on this thread, it looks like it could be Greenland's turn to take the heat in 2012/13.


I think there's a sill in Kane Basin of about 200m - where PII2012 is hung up. I also think Baffin Bay has passages considerably deeper.



Dr Muenchow has a bathymetric chart of Nares Strait at his site,


He has a great blog, was on the first (only) ship to get behind PII2012 while it was still in the mouth of Petermann Fjord, posts here from time to time, and doesn't seem to mind answering questions.



The refreeze this year will be fascinating. After, IIRC, a low pressure system moved west from the Labrador Sea in Dec 2010, remelting much of Hudson's Bay, it was for the first time not fully frozen at the end of the year.

Then earlier this year, the Kara Sea melted out in February.


It will be interesting (or horrifying) to see how much accumulated and advected heat goes into the Arctic this winter and where it goes. If the heat flows north of Norway then I would expect the Barents Sea to reman ice free. The Kara Sea would appear to be the next domino that could become ice free, or end up with very thin cover. With the "Laptev Bite" this melt season, a lot of heat appears to be flowing under the Kara and some churn mght well bring up warmer, saltier water to the surface. Perhaps the anomolus heat Jim Williams refered to earlier will end up off northern Russia. Or it could be bad news for the polar bears of Churchill.

Chris Biscan


Dr. Fraud has officially got his downward trend back.

Jim Williams

idunno: "If so, then this coming winter could bring EITHER a repetition of of winter 10/11, bad news for Greenland, OR a repetition of winter 11/12, bad news for the sea ice in the Barentsz and Kara seas.

"At the moment, based on the temp maps referred to on this thread, it looks like it could be Greenland's turn to take the heat in 2012/13."

Seems rather reasonable. If the Gulf Stream is wandering then the question would be: Why?

Aaron Lewis

Sea ice in the Arctic is an inherently unstable system at this point.

It is mechanically fragile, and subject to large mechanical stress as a result of cyclonic storms driven by the temperature differential at the ice's edge. My estimate is that the current sea ice has only 5% of the mechanical strength that it had only 7 years ago. Thus, now a storm is 20 times more likely to break the ice up into small pieces. Small pieces have a larger surface to volume ratio and melt faster.

It has lost mass and warmed, so that it has much less thermal inertia, and thus much less likely to survive short term warming events.

The summer albedo has dropped so there is more heat available to melt the residual ice.

And, the average temperature of the summer The Arctic is now above freezing so that latent heat from the south can be easily advected by the lower atmosphere to the sea ice where the water vapor condenses, melting the ice. All of the NH is now a heat collector for the Arctic sea ice. (With some heat diverted to other ice.)

This is not a ball sitting on a saddle; this is a system in free fall. It may do some weird orbital things, and it may collide with other objects, and it may be struck by objects going much faster, but it is in free fall. It has been falling for 60 years, so it has built up a lot of speed.

No, our climate models have not captured the essence of the issue.

We should have acted sooner. We did not. Now, we must act very aggressively. We have a stark choice. Do we want to mess up our economies for a few years, or do we want to mess up our economies forever?

Andreas Muenchow

The Gulf Stream is NOT changing direction unless you count or pick the right time and the right spot of the many wiggles, meanders, and eddies that are all a normal part of it. Once the Gulf Stream reaches the Grand Banks off Newfoundland (all the fog there is the meeting of the sub-polar Labrador waters inshore and the sub-tropical waters offshore) the tightly focused Gulf Stream becomes the more diffuse North-Atlantic Drift current. If you average long enough in time over a large enough regions, you will get the net heat flux that keeps northern Europe's climate so mild. It is not correct to talk about this broader more diffuse drift to the north-east (yes, it has on occasion southward elements here and there) as the Gulf Stream. I think Neven a few weeks ago summarized all this very eloquently .

Oh, and the "deepest" connection of the Arctic to the Atlantic is Ocean to the west of Greenland is indeed, as Terry points out, the 200-220 meter deep sill in Nares Strait where it is winter already with new ice forming.


real time current info - ocean currents - can be gotten here:


(one has to generate ones own plotting to see clearly)

yes it shows exactly what Andreas Muenchow just explained (focused Gulf Stream, then changing into a more broad flow field).

there are strong anomalies apparently in the SST in the Baffin Bay and even in the Greenland stream but one shouldnt confuse an anomaly map with an actual stream map.

The two branches of the north atlantic warm water inputting heat into the arctic (near spitsbergen and in the behindmost Kara sea) can actually be seen on the daily SST maps and they do show positive anomalies as well. I suspect if we talk about anomalous heat input into the artcic for this coming winter, from the ocean currents we should more focus on those two branches.



I hope your guarantees about the Gulf Stream is right, otherwise it becomes very cold here where I live!


I would like to draw people´s attention to some ocean buoy data, namely ITP-41. It is drifting in the Beaufort Gyre. ##


If I can believe what the data page says (I have no experience with this), then the buoy was deployed in 2010, and is still transmitting regular T&S depth profile data as of today. If I can take that at face value then the buoy has seen the whole melting season 2011, and now the whole melting season 2012. So one can compare.

One should look at the "plot of ITP T&S contours". That gives a page showing Temperature and salinity data, as depth profiles versus time. There are a lot of disturbances some of which for sure are technical (unreliable data), which I am in no way qualified to even recognize. However, the general structure of the surface layer through a melting season is clearly visible. Especially, I want to draw attention to the top plot, of temperature, in the top 200m. One can actually see the heat pulse of a melting season. That of melt season 2011 starts at ca. day 570 and one can se the thermal pulse, as it moves deeper, and how it is followed by a cold pulse from the developing top winter, while it still moves deeper. And then from about day 900 on comes melt season 2012. (one can actually see the august storm as a short mixing event around day 950 being August 8 2012).

what strikes me about this figure is a comparison of the strength of thr 2011 thermal pulse versus that of 2012. This is solar energy input into the water (nothing to do with warm ocean heat flux from below). Now, as shown in the location data the buoy has drifted, and it does have moved south so the two seasons are not completely comparable, but the latitude difference may be no more than 3 degrees if I can believe the data. Given that, what strikes me is the apparently much more intense heat pulse of 2012 compared to 2011.

If anyone here has actual knowledge about arctic water science maybe they could comment? Is this all to be put down to the latitude difference, or does it indeed show that the solar energy uptake in the upper layer in 2012 (in that region) is much more intense trhan in 2011? If so, can one based on that expect that the winter ice growth should be seriously retarded there? Can this lead to a kind of cross-seasonal memory effect, setting up strong melt season 2013 now already - or are winters there still strong enough to erase such differences?

Jim Williams

Enno: "there are strong anomalies apparently in the SST in the Baffin Bay and even in the Greenland stream but one shouldnt confuse an anomaly map with an actual stream map."

We're not. The Gulf Stream per se can be clearly seen in anomaly graphs as being slightly north of the climatology from Cape Hatteras to about Massachusetts.

The extension of the Gulf Stream commonally called the North Atlantic Drift Has been trending northward for decades; which has generally tended to warm Northern Norway. However, in recent years the Drift has been so far to the north that occasionally the current has linked to the West Greenland current rather than the East Greenland Current.

We are specifically looking at the anomaly in the current from historical climatology.

Jim Williams

Espen: "I hope your guarantees about the Gulf Stream is right, otherwise it becomes very cold here where I live!"

I think he promised you that the future will be exactly like the past.

(It will only be cold until the GIS is gone.)


Jim, that´s good. However, the anomalies in water temperature in the current going around Greenland do not tell you that their anomalous warmth comes from the Gulf Stream. As far as I could see on the maps, already the southward flowing Greenland Stream between Greenland and Iceland has anomalies comparable to those in the Baffin Bay. I would therefore naively think that it can be that it brings its (relative to climatology) higher heat content with it, than that it gets it from the Gulf Stream. (This is entirely possible - that flow carries a significant heat return derived from the Atlantic warm input, nonwithstanding that it is a cold flow at its location). To make the specific Gulf Stream connection one would have to have more specific evidence that that is the case. Maybe there is such evidence but do you (or someone) have any? Up thread, it has been speculation; legitimate for sure but not to be confused with actuality.

The anomaly map at

does not seem to me to agree very well, moreover with that at


Basically, I can´t follow through the apparent data what exactly it is that brings you (or others) to suggest "the Gulf Stream", or a derivative of it would at this moment be delivering abnormal heat to Greenland waters. Well one could speculate that the cold countercurrent is abnormally warm because higher than usual heat has been input under the Arctic through the Barents and Kara (as is known it has); but to call that heat from changes in the Gulf Stream would be quite a stretch.


Hey guys. I noticed the debate RE how fast Greenland, Antarctica could melt. For the level GHG we have now, I'd think that the 1000 year study showing all Greenland melts out is slow. More like we see both West Antarctica and Greenland melt out in that period at the current forcing. At least that's what paleoclimate shows at 400 ppm CO2.

IF we get to 600 PPM CO2, the pace of melt should be much faster, on the rate of 3-10 meters per century seen in paleoclimate maximums.

IF we get to 800-1000 PPM CO2, I don't think there's anything keeping the ice from completely melting out within 300 years or less. Very rapid, very destructive. There's just too much energy for the ice caps not to soften up rapidly at that state.

For my part, the models are far, far too slow. Paleoclimate does better. But we need to go beyond paleoclimate and gain recognition that both the pace and force of human GHG emissions is far greater than that seen in geologic history.

Because most scientists are cautious, they are less likely to be accurate than a threat analyst, say.



More heat transport to Greenland. Joy...



(It will only be cold until the GIS is gone.)

Then I am OK, I will be gone by then, at least I believe so!
PS; My girlfriend always tell me, not to worry too much!


Does anyone know of a site that explains how to read the data that that the buoys provide?

The second from the top charts both for temp and salinity don't seem to match the upper ones, the temp salinity "profile" page is a complete mystery.


Protege Cuajimalpa

Rob Dekker:

We agree that models are important. It is just that most of them have fail too much and I believe that discussing what it going to happen at the end of the XXI or XXII centuries, is misleading. Almost everybody don’t care what is going to happen at 2100, so instead of promoting their interest, the models seem to make the job that deniers love to do. That is, making believe that there is time left to react.
I would like to see a new generation of models that will focus on what is going to happen at 2030 and 2050 (both), taking into account that the Arctic Sea Ice will melt at 2018. I would like to see that they change the methane carbon equivalence from 20-25 times CO2 (hundred years equivalence) to 72 times (twenty years equivalence). Those models will make people think and react against climate change, because they will learn consequences that are going to happen in only 17 or 18 years (consequences that will affect our lives). But at this moment, the facts are the ones that are telling us the true about how fast global warming is going to happen (even that we learned with the models, that is also true).
Anyway, I want to end this comment with a today’s news: “Leading climate scientist warns that vulnerable island nations may need to be evacuated within a decade as evidence shows polar ice is shrinking at greater speeds than models predicted”:


Protege Cuajimalpa


Thank you [again] for your answer. I want to use some of your graphs, with your permission. ¿Is that ok?



If we look at ice mass loss in Greenland, melt is nonlinear in much the same way that sea ice melt has been nonlinear. But, for Greenland, we're still pretty high on the curve.

That said, if end of century predictions can be accelerated to 2030-2050 (as has happened with sea ice melt), we could reasonably see 1-3 meters or more of sea level rise by that time. The problem is that these rises are most likely to occur in pulses. A silver lining, if you can consider current conditions to be a silver lining, is that we would likely notice very large ice lakes forming in the interior of the ice sheet, then breaking out via an ice damn collapse. Another scenario could include multiple melt lakes at varying altitudes on the ice sheet. Once the number grows to a certain point, a collapse of one melt lake far up the ice sheet could set off a domino effect as the flood overflows numerous ice lakes. The result, down glacier, could resemble a glacial tidal wave pulsing off of the glacier.

There is geological evidence for large melt lake collapses -- English Channel flooding, the mounds deposited by melt across North America, and trees ripped from lands in Canada and Alaska and deposited thousands of miles away on islands off the coast of Russia. Slurry remains of mammoths and other animals, all mixed together and broken into bits as in a blender, have also been considered evidence of catastrophic melt pulses by a number of Arctic researchers.

Everyone, the scientists, even the threat analysts are generally cautious about making too many dire predictions. No one wants to be seen as a doomer or someone predicting the end of the world.

My personal opinion is that people who live on or near any coastline should consider leaving about five years from now. At some point, it's going to become pretty obvious that the coasts are very precarious locations. At that critical mass, it becomes impossible to sell homes/real estate. As for actual risks due to increased storms and sea level rise -- areas of the East Coast and Gulf Coast are already at heightened risk. The coasts are sinking and the seas are rising. A bad combination. In addition, the storms are becoming more intense.

At ten years, I'd say there's a 50% chance that sea level rise has doubled or worse. At 20 years, I'd say that Greenland is at risk of a moderately large melt pulse, At 30 years, I'd say that Greenland's risk is much higher and you've got to start putting Antarctica into consideration. By 2050, under BAU, with no sea ice in the Arctic at any time during the year and with Antarctica starting to see dramatic net outflows, things will probably start to get hairy.

The problem is that we aren't used to dealing with nonlinear problems. So we're apt to be surprised once events start moving rapidly. Look at the 2012 melt. It shouldn't have happened. Storms didn't turn hundreds of thousands of square miles of sea ice into slurry in the past. There were minor instances. But not this. Now we have a multiplication of the number of conditions that have resulted in melt. Storms, consolidation, spreading out and flushing out of the Arctic, winds, calm, ocean current incursion. It seems that anything can cause melt in the summer time now. It would seem that forces acting to melt the ice sheets would multiply as climate warms and insulation is lost.

Finally, I'd like to ask a couple of questions. Does anyone have the final numbers for Greenland melt this summer? And does anyone have solid melt numbers for Antarctica (the ice sheet) through 2012? Would be very helpful.

Rob Dekker

About ITP41 :
Normally, ITP data is extremely interesting to analyze, since it tells a lot about what is going on beneath the ice or at least beneath the sea surface.

However, it seems that currently Woods Hole has a defect in their plotting software, so you CANNOT TRUST the plots they put out currently.
For example, ITP41 profile shows a surface temperature around 0 C and warming (which is really bizarre at this time of the year.

But if you check the data in the file "itp41grddata.zip" on the main page for this buoy :
you will find that the profile data itself mismatches with the plot, and frankly the data makes much more sense.
For example, here is the close-to-the-surface data for ITP41 from one of the latest runs :

%ITP 41, profile 1468: year day longitude(E+) latitude(N+) ndepths
2012 279.25006 -137.7083 73.9674 223
%year day pressure(dbar) temperature(C) salinity
2012 279.25039 192 -1.5238 33.2166
2012 279.25049 194 -1.5174 33.2530
2012 279.25062 196 -1.5083 33.2781
2012 279.25074 198 -1.4970 33.3023

This suggests that surface temp is a realistic -1.5 C, which means that the plot is way off right now.

I actually got bitten by this defect in the plotting software, when ITP57 in the Laptev bite started to show some really odd behavior suggesting that there was significant eddying and bottom melt going on. Until I found out that the data itself did not support the plots presented :

It seems that Woods Hole fixed the ITP57 plotter (it now shows a normal profile) but ITP41 is sure way off right now.

I'll send and email to Woods if this problem is still present by monday.


Rob Dekker,

What is your take on this issue ?

First, I am familiar with "control systems" by profession. So please use the terminology if it suits you.

only an unstable system would cause an time-exponential derivation from the mean (like a ball on a saddle will roll off) and make it collapse or expand to extreme extent

I would raise the following questions:
1) perhaps the system _is_ unstable;
2) if the system _is_ stable, the "mean" can still be highly non-linear.

I am inclined to think the latter is the case, after all ice/water is a infinitely non linear function of temperature at the melting point. But I have an open mind toward possible tipping points or bi-stable behavior etc.

I have often made the case that exponential extrapolation does not have any basis in physics (since it implies an inherently instable system)

I read this before and I cannot comprehend how that makes sense. Its 40 years since "The Limits to Growth", that is full of information on exponential curves.
I don't think that it is disputed that with current atmospheric composition a permanently frozen "snowball" earth is a second stable condition. As a whole the earth is not stable, a large enough push will start an transition that undoubtedly can be modeled by exponential functions at some point.

These have all a place in physics. They differ from the mathematics that the growth/decline will not go on forever. But that is nothing to do with exponential functions. A ballistic curve of a projectile is still a parabola, even if everyone understands that the parabola must end somewhere.

To summarize: the exponential fit is more a question than an answer. What is settled is that for 80% it is an accurate discription. For the next 20% it is one of the calculated speculations.


I want to use some of your graphs, with your permission. ¿Is that ok?

Fine with me, I am glad you think they are useful.

Rob Dekker

Terry, Enno,
Let's take another look at your questions once Woods Hole fixes the defect in their plotting software.

Meanwhile, please check out the new T/S profiles of the new bouys that were recently installed in the Beaufort : ITP63, 64 and 65. As far as I can see, their plots match the profile data.
There are some interesting differences in the upper halocline between these buoys and diffences in salinity profiles, which suggests that the upper 200 meter sea surface has stirred up rather irregularly during the end of the melting season this year. Some heat/salt from great depth has made it to the surface at some places but not all...

Better analysis can be done once Woods Hole fixes the plotter software for the other buoys..


I originally completely bought into the exponential loss of sea ice volume. The reasoning seems simple more open water absorbs more heat increasing the rate at which more open water appears. Physical phenomina where the rate of change is porportional to the amount are exponential. It is important to remember that ice volume is an effect of energy into the system and energy leaving the system. But the arctic is much more complex than this. Some of the extra energy into the open water in the system simply heats up the water and does not get close enough to the ice edge to melt any ice. So we have a chaotic system where storms and weather act to move energy around which sometimes bring more energy into the ice. I wanted to increase my understanding so I retrieved the images from the US Navy arc volume graphs 2010 - 2012 and compared the ice thickness from May 15, July 15 and Sept 15 for those years. The story they tell is that the massive volume loss occurs from the edge of the ice inward. The melting in the central part of the ice is a very small percentage of the overall volume loss. In fact if I compared the thickness in the ice that survived this year with ice in the same area in Sept 2011 and 2010 I was surprised to discover that much of that ice was a little thicker in 2012 than in 2010. Not by much 0-20 cm but it was not thinner except close to the edges. So as the central ice gets smaller the area of the band at the edge where most of the rapid melting occurs gets smaller and the extra heat in the rest of the arctic needs more chaotic weather to reach the edge. This could be one of the contributors to creating a sigmoidal loss instead of an exponential crash to zero in the near future. I do not know the answer and I can only wait for more data to indicate the most likely path. Also, since the arc graphs are designed to show the maximum thickness instead of the average this may be misleading me. However, if real this more complex view of the ice may be leading some scientists to believe 2020 or 2030 is more likely than 2016.

Chris Biscan


Chris Biscan

Those Navy ARC graphs are worthless, don't use those.

Andy Lee Robinson

I've just uploaded a new PIOMAS animation:
(I made a cross-your-eyes stereo 3D version too!)


Cool, Andy. I've added it to the More vids post.



My problem at this stage isn't that the plotting may not be reflecting the data, but rather how to read the graphs themselves. I don't understand what the second graph down represents. It can't be simply an extension from the top graph, nor is it a larger, less detailed representation of the water column showing the extra depth.

I've got other questions about the graphs, but resolving this one would be a start.


Steve C

I'm no expert, but maybe I can provide insight. Let's look at ITP-41 profile image at:

Yes, the plots are in pairs. I think understanding them demands careful attention to the vertical scales. In the two pairs, the top chart has a smaller range of depth -- it's simply a kind of close-up of the layering of the top 200 meters. The one below displays the full-depth profile.

So why does 200 meter color on the top chart not match the 200 meter color on the lower? The color scales are different, but you might miss that if you don't pay close attention to the negative signs for temp on the two different scales. It seems they do this so that the color scales can be used to better show the stratification over the different plots.

Hope that helps.

Steve C.


I have updated the sea ice concentration and thickness for Sept 30, 2012 from NCOF/Godiva2. Also I have added (as an experiment) a column with the MMAB/AMSR-E sea ice concentration images for 083112, 091712, and 093012. If they are of interest, I can add more.

The NCOF kmzs are included for 093012.



Thanks so much Steve. As usual I'm embarrassed at my not noticing something so obvious once it was brought to my attention.


Protege Cuajimalpa

Thank you, Robertscibbler, for your interesting explanation.
Some questions: Why the East Coast and Gulf Coast are sinking? And can you tell us the speed rate in which they are sinking?



Sorry for the inexact description. In short, some areas, like Hampton Roads in SE VA are settling due to their geologic make-up (sitting on fractured rocks etc), in addition, the SE is pivoting back due to crustal rebound. This is a pretty slow process, though. Lastly, additional sea level rise is occurring in the region. Some scientists have written that this may be due to the gulf stream backing up or to melt from Greenland moving the currents around.

Chris Biscan

Hey Terry, how nice or refreshing is it to post here and not have folks treat you like dog bleep if they disagree or you are wrong or post a mistake.

Rob Dekker

Wipneus, thanks for your response !

I am familiar with "control systems" by profession

That's good, since for me it has been 25 years since I took the course.

I would raise the following questions: 1) perhaps the system _is_ unstable; 2) if the system _is_ stable, the "mean" can still be highly non-linear.

I am inclined to think the latter is the case,

I agree with you.
After all, IF the system is unstable, than it would be hard to explain why ice extent did not collapse over at least the past two thousand years, not even before that, when Arctic irradiance was significantly higher than today.

Which leaves "if the system _is_ stable, the "mean" can still be highly non-linear", and this seems to match observations.

However, if the mean is highly non-linear, this would suggest that a small forcing is able to "knock" the "stable" ice extent from the past 2 millennia into some non-linear state which allows 80 % of volume to be reduced without any significant "push-back" from any sort of negative feedback getting it back in line.

If that is the case, then what would be the constraints on further ice melt ?
Will it just melt through to virtual ice-free state, followed by further reduction / ice free conditions in ice cover in August, July and then into June or even earlier ?

From a "control systems" point of view, which behavior would Arctic sea ice have to show that there are constraints to the "non-linear mean" that you suggest as the most probably option ?

Rob Dekker

Terry, I'm sorry I misunderstood your questions. I hope that Steve C. answered then satisfactory. The scale on the Wodds Hole ITPs is not just differently colored, but also non-linear. They seemed to have constructed that non-linear scale to emphesize the temp/salinity differences at the upper and lower halocline.

Let me also add that at this point I'm not sure if my assertion (that the problem is the plotting software) is correct.
After futher investigation, it seems that some of the ITP data for a single run is split between two files. For ITP41 specifically, it may be that the data is actually consistent with the plots if these two files are merged, but even then the data for the lower halocline as presented in the plots still seems incorrect.

Needless to say that at this point I would not trust any of the ITP data (or plots).

Let me send Woods Hole an email this monday, before we draw any conclusions from any of the ITP plots or data sets.


Rob Dekker,

Very interesting questions, I don't think the answers are there yet.

It would be hard to explain why ice extent did not collapse over at least the past two thousand years, not even before that, when Arctic irradiance was significantly higher than today. (...) However, if the mean is highly non-linear, this would suggest that a small forcing is able to "knock" the "stable" ice extent from the past 2 millennia into some non-linear state

I don't think you can directly compare the Nothern Hemisphere Summer Insolation difference with green house gas forcings. Over the globe average irradiation at the top of the atmosphere did not change, just the distribution over the latitude and the seasons. Higher irradiation yes, but less heat transport from lower latitudes.
I think the stability is well explained by relative stable effective forcings. You could read the work of Ruddiman. He present proof that the growth of icesheets on Canadian islands after the peak in irradiation did stop and revert about 500 years ago. You don't have to support his explanation (anthropogenic warming associated with earliest civilizations) to accept the fact that there is something balancing the reduction of irradiation.

From a "control systems" point of view, which behavior would Arctic sea ice have to show that there are constraints to the "non-linear mean" that you suggest as the most probably option ?

A control systems point of view does not tell you this, just from the shape of the measured response. Yet for better effective control it is essential to know the cause. That avoids the "inertia" that is always in between a cause and its effect. Inertia’s or "lags" as control engineers call them are the enemies of effective control.
I am afraid the measured decline in ASI is ultimately caused by greenhouse gas emissions with a lag of 20-40 years. That means effective control is nearly impossible (effective = save significant portion of the ice). Attempts with geo-engineering will when employed in such measures that the ice will remain, will overshoot and plunge earth into the other direction of freezing. Of course geo-engineering can be employed once again, but that will overshoot with even worse warming and decline in sea ice.

BTW, when I saw this graph first: http://upload.wikimedia.org/wikipedia/commons/f/f7/Five_Myr_Climate_Change.svg it struck me as it could have come from a textbook on control. It shows the destabilization effect that the ice and snow on the northern hemisphere has on earths climate. This all brought to you by a few ppm's of CO2. We are bringing the CO2 levels now back to those of 4 million years ago in the middle of this century. That is likely to melt virtually all perennial northern ice, not bad for a small forcing.

Seke Rob

Looks like that latter chart expresses the Milankovitch cycles, per the notes right top.

Personally, I've never considered for TOA to vary at any point of earth, given it's relative puniness compared to the sun... it being for instance 1365.1234 Watts/M^2 at any point on the sun facing side at the same moment in time. Never seriously considered the Earth Magnetic Field to work as causing different TOA at different locations. I'd not be surprised though that TSI get's influenced before it get's through to the top of the atmosphere and the earth's surface, ignoring the atmospheric impediment itself. At same time I'd consider this one of these "we don't know enough, therefore..." 8th decimal arguments. Broadly solar irradiation is constant at TOA and by the time it's getting to Earth's surface, a quarter of it is left [global average].

Rob Dekker

Seke Rob,
I think that Wipneus referred to the Milankovich cycle plot to show that the small global GHG forcing 4 million years ago was able to prevent glacial periods to occur during times when Arctic irradiance was at least an order of magnitude larger than GHG forcing.

In other words, a little bit of CO2 (280->560 ppm) can prevent 80 W/m^2 swings in regional Arctic irradiance from creating glacial/inter-glacial cycles.

Seke Rob

Rob Dekker, believe in the American language they'd say "check", a "Vinkie". The 3 MC's are one of the least included [relatively simple to understand] points made in public discussion to explain why at other epochs with lower or higher CO2 it was warmer or cooler. We should be cooling [since about AD 0, coincidence ;>), yet we're warming.]

Rob Dekker

Sorry, Seke. I do not understand.
What are you saying ?


Hi Guys, I've been following this thread with some interest. What I'd like to see is a plot of arctic seas against ice. I'd just like to see the 'other side of the coin'. I think the exponential is a natural growth curve, it is the sea that is growing. I'm happy to say that I lack the information, skill and time to do this, but I feel it would be worthwhile in the current situation.

Artful Dodger

Seke Rob wrote (October 08, 2012 at 12:29)

"Broadly solar irradiation is constant at TOA"

Hi, S/R

Top-Of-Atmosphere Total Solar Irradiance (TSI) does indeed average around 1365 w/m² when corrected to 1 AU distance to the Sun. And this is the appropriate data to use when studying the Solar cycles.

However, for terrestrial phenomena like climate we can not ignore the elliptical shape of the Earth's orbit and it's coordination with the seasonal cycle. In 2012, the Earth was furthest from the Sun (1, aphelion) on about July 5:

At aphelion, Earth is 3.28 percent farther from the Sun than at its closest approach. This means that Earth receives about 7 percent less radiant heat at its farthest point from the sun than at its closest point (Austral Summer).

The dates for aphelion and perihelion can vary by a few days. Perihelion (the Earth's closest approach to the Sun) typically occurs between Jan. 1 and 5, and aphelion ranges between July 2 and 5.

SOlar Radiation & Climate Experiment (SORCE) data shows that on June 5, 2012 TSI was 1322.68 w/m²; by July 5, 2012 TSI decreased to 1316.74 w/m². The difference at TOA is about 6 watts, or about 1.2 w/m² at the surface of the Earth near 70N.

As we know, the total energy imbalance which has caused the loss of Arctic sea ice to this point is about 0.8 w/m² so early albedo loss is potentially a significant forcing going forward.

The effect of Earth's orbital mechanics is to reinforce any early loss of albedo (sea or land) during the Northern Spring. Early sea ice or snow melt creates yet another positive feedback leading to greater melt.


Seke Rob

Lodger, I'm unfamiliar with any of these 40-45 Watts off TOA values you mention of 1322.68 and 1316.74, but take your word for it given the accompanying factoids. Multiple sites are off-line ATM, JAXA, UAH, Washington, Lasp, so cant follow your link. You cite http://lasp.colorado.edu/sorce/data/tsi_data.htm#summary_table

The daily data I use is [see my Solar graphs page] https://sites.google.com/site/allthingsclimatechange/the-sun ]

For 6 hourly: http://lasp.colorado.edu/sorce/tsi_data/six_hourly/sorce_tsi_L3_c06h_latest.txt

For daily: http://lasp.colorado.edu/sorce/tsi_data/daily/sorce_tsi_L3_c24h_latest.txt

Curious how 3.28% distance variance causes 7% more [or less] watts to arrive at TOA. For laughs, how Pluto was warming per Goddard going from 7+billion distance to 3 billion over the approach period of 125 odd years [He did not consider that fact [lying by omission is possible too]. Yes per Greg Kopp of SORCE, their numbers are corrected back to exactly 1 AU [using JPL ephemeris VSOP87 it so says] But, I'd consider, outside of Milankovitch, on human lifetime scale, the ellipse to be constant enough, that at any date point in the year, [but for the 1.3-1.4 watts variance over a solar cycle at 1 AU], the TOA for any date of the year to be constant. Computing that against albedo to determine the P/Ejoules is no doubt of major relevance.

The 0.8 Watts is already more than the solar cycle TSI variance from median at 1AU. To me discussions such as "how much does the kinetic energy of the Arctic gyres or 6000 tons per second leaking out of lakes top of GIS cause warming" is surreal. Obfuscation, bigtime, compared to the permanent 0.8 Watts/Meter square, humanity has added [well including consequential feed-backs it's more like 3 watts]


Curious how 3.28% distance variance causes 7% more [or less] watts to arrive at TOA.

Seke the reason for this is the inverse square law.


Seke Rob

Spreading thinner... makes perfect sense [all that stuff I'd forgotten about] :D

Artful Dodger

Happy to be of help, S-R. And thank-you Phil for introducing the inverse-square law.

SORCE is a NASA satellite launched in 2003, and operated by UC Boulder. The SORCE daily data file lists TSI @ 1 AU in column 5 (TSI_1AU), and TSI @ true distance in column 10 (TSI_True_Earth).

It's this column 10 we want to use for our study of Boreal Summer and Arctic sea ice:

yyyy-mm-dd TSI_True_Earth
2012-03-20  1372.4
2012-04-20  1347.8
2012-05-20  1329.2
2012-06-20  1318.4
2012-07-20  1318.5
2012-08-20  1333.7
2012-09-20  1355.6

So the timing of the melt is a feedback. Earlier melt = More heat per day.


Rob Dekker

Thank you for the TSI data ! Very interesting.

Please also note that amount of that TSI that makes it down to the surface ultimately determines the amount of heat absorbed once ice changes to open ocean.

In that regard, Tamino has made three posts now on ice, snow, and ice+snow albedo effect.

These posts are well worth reading, considering that they quantify the increased heat absorption due to decreased ice and snow cover, using a simple calculation method.

The last (snow+ice) suggests that albedo effect over the past decades is as high as a 0.45 W/m^2 global forcing, which puts it fair and square in the ballpark of GHG forcing that presumably caused it in the first place. Not to mention that the effect may be mostly regional, thus significantly adding to sea ice loss.

Moreover, the IPCC does not account for this albedo effect in AR4, which may partly explain why ice loss and polar amplification seem to be underestimated in the official records.

Even more interesting is a post from commenter "Thomas Huld", who used the ECMWF database to calculate exactly how much extra heat the Arctic absorbed due to sea ice loss :

Thomas uses a method that is comprehensive and accurate : for each month in the record since 1980, calculate how much heat was absorbed in the Arctic by combining sea ice concentration and downward solar radiation data in gridded monthly maps,
Apart from confirming earlier (Hudson 2011) estimates of the albedo effect unfolding in the Arctic as we speak, his numbers suggests that there is an increase in cloud cover over ice-free regions.

I'd sure recommend you read his post, and provide comments on it, since it is clear that this guy is on the right track in quantifying how much sea ice (and possibly snow cover) anomalies over the past decades really contribute to global warming.

The comments to this entry are closed.