« ASI 2016 update 1: both sides | Main | Persistent Arctic and sub-Arctic warmth »


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

Susan Anderson

Wow. That's all about it. Thanks for links.

Remko Kampen

To be sure, Arctic sea ice cannot but exist as seasonal ice at present climate conditions and we are witnessing the transition to this situation now.

Joe Wentrup

Jet Streams and ice shield guaranteed climate stability in the northern hemisphere. Wonder where's the tipping point of the jet streams. When will they collapse? There were those strange events this winter, with above zero temperatures at the north pole ... And I remember scientists found fossile rests of palm trees in Greenland from a time with similar CO2 concentration in the atmosphere as today.


@ Remko,

Why not exactly?


Part 1 seems to be mostly about the permafrost and, although not explicitly mentioned, Methane.

RealClimate published a good explanation of the permafrost-methane issue about 4 years ago which I think should be re-read to provide some balance to the tone of this article.



-Fish here - I don't think that a 4 year old blog post is a good basis for evaluating the climate risk posed by methane. Archer's layer cake model, discussed in that post is not realistic. Moreover, numerous papers in the last 4 years on the degradation of Siberian and permafrost have found that the problem is larger than we thought 4 years ago.

The Real Climate argument that stored methane would have been released in previous interglacials is not convincing. This interglacial peaked about 8000 years ago and is undergoing a late resurgence thanks to GHG emissions. This interglacial is not like the others.

We are entering unknown territory. The Arctic Methane Emergency Group is clearly at the alarmist end of the spectrum, and evidence I've reviewed does not support their extreme position, but I am not comfortable with the 4 year old analysis based on obvious geological oversimplification.

Several recent (2015 & 2016) papers show that the carbon burden of the Arctic seas is going way up as permafrost erodes. This is causing rapid acidification and deoxygination of the shelf waters in the Arctic seas. These waters will have little to no ability to oxidize methane before it reaches the atmosphere.


"Why not exactly?"

Let's have a look at that.

Right now we have legacy pack ice which has a minimum size that protects some of the Arctic all year round.

That ice is diminishing, year on year and the arctic is heating, year on year.

It takes more energy to melt all the ice that is there, than it will take to melt the ice when it is seasonal.

The second that the ice goes seasonal, that energy which is currently being used to convert the ice to water will go into heating the seas instead.

This will shorten the freezing season and lengthen the melting/warming season. Both of which will increase the heat budget hugely as every year goes by.

At the same time CO2 is increasing at an aggregate 2.2ppm per year, decadal average and increasing, decade on decade, by 0.3 to 0.4ppm decadal average growth.

So, to reiterate.

It's getting warmer
The ice is melting
Which will make it even warmer and trap even more of that heat
CO2 is growing, not falling
CO2 is trapping ever more heat

In that scenario, when we hit a minimal ice cover, the transition to seasonal ice is going to accelerate rapidly and then the heat which is trapped is going to lock in the seasonal ice situation.

The change may be very slow. But the flip over is probably going to be very quick.

Back in the 90's I likened what we were doing with CO2 as pushing a huge wheel up a hill. It's uphill all the way and takes huge amounts of energy to get there. Some time you hit the top and things start to really move. Then you start downhill again....

The downhill, as I see it, is the inability of the planet to absorb any more CO2 as it is doing today. At least doubling the growth and the impact.

I see absolutely nothing wrong with what Remko said.

Chris Reynolds


"To be sure, Arctic sea ice cannot but exist as seasonal ice at present climate conditions..."

Is this the case?

Obviously when one factors in the turn-around time from the current trajectory of fossil fuel dependence the sea ice is toast. But I suspect this is not what you mean. So...

Is the ice response lagging or leading the driving of climate forcings? I genuinely do not know. If you don't - is there proper research on this issue? I am not aware of any. If there is no research to throw light upon the situation do we have enough data to provide an answer?

And following that line of thought:

1) Is it informative that (in line with Mahstein and Knutti 2012) the relationship between global average temperature and NSIDC September Extent is linear and the extrapolation approaches 0 extent at around 2degC global average temperature anomaly (in line with Mahstein and Knutti 2012)?

2) To what degree do 2013 and 2014 and the recent volume pulse inform us of the lag or lead relationship with forcings? Here we might view a factor such a regional Arctic temperature as providing a 'wobble' on the driving signal of global temperature.

1 & 2 would suggest to me that the sea ice is in phase with climate forcings, with little lead or lag.

Have I got the wrong end of the stick entirely?

One way of going about this would be to calculate lead / lag correlations, but amongst a short period of data ~40 years with strong trends this is tricky.

I promise not to sea-lion you on this. I've let Neil's assertion ('probably very quick') go, because I am aware I am rather bore on this. But if 'your side' (the fast crashers) are right I want to be in on the excitement.

To explain - Sea Lioning...


Colorado Bob

Investigating Climate Change the Hard Way at Earth's Icy "Third Pole"

A glaciologist doesn’t let a heart transplant keep him from braving dizzying altitudes to gather crucial ice core samples from retreating tropical and subtropical glaciers

What have you learned from Tibetan ice cores?
They have given us a glimpse of Tibet’s climate history going back to more than half a million years. We learned that the extent of glaciation is related to how far monsoonal rains penetrate the Tibetan Plateau. This is in step with the slow wobbling of Earth’s rotational axis, which drives tropical rainfall in 21,000-year cycles. We also identified periods when average temperatures in Tibet went up and down by several degrees Celsius in roughly 200-year cycles. It’s still a mystery why that was the case, but we suspect this may be related to the 205-year cycle of solar activity.

In more recent times an interesting discovery is that the higher the elevation, the greater warming we have. This is in line with the observation that the vast majority of glaciers in Tibet and the Himalayas are retreating. In some extreme cases, as ice cores from the Naimona’nyi Glacier in southern Tibet show, all the snow and ice that accumulated since 1950 has melted or sublimated away at altitudes as high as 6,000 meters above sea level.



There is one topic that is missed entirely related to permafrost and carbon release and that is wildfires in the Arctic.
These create weather systems, darken ice, pump tons of carbon into the air, and farther degrades the permafrost and releasing carbon much much faster then normal release would take. I do not believe there are very many models that deal with the release of that carbon that take wildfires into consideration.

Rob Dekker

I'm with you that I don't know if Remco's statement that "Arctic sea ice cannot but exist as seasonal ice at present climate conditions" is true or not. I don't know if there is enough lag (a buffer of some sort) in the system to validate that statement.

But even if there is no lag, I question your statement of linearity :

the relationship between global average temperature and NSIDC September Extent is linear

Is it ?
If I look at global temperatures (here is NASA GISS) :
then you can certainly argue that at least since 1980 the temps are increasing somewhat linear.

However, if we look at NSIDC September Extent since 1980 :
that 'linear' line through the data seems just not right at all.
Tamino notes that a quadratic or lowess fit leaves much smaller residuals :

And when we look at PIOMAS volume, some sort of non-linear fit also seems much more appropriate than linear :

So it seems that while global temperatures are increasing linearly, sea ice volume and extent both are accelerating.

If there are no buffers in the system, then I think the most likely explanation of that non-linear behavior of Arctic sea ice is caused by albedo effect being more and more significant, now that the ice edge at summer solstice is intruding on the Arctic ocean itself.

And to test that hypothesis, 2016 is a very interesting year...

Rob Dekker

Let me add that I am open to any other explanation on why Arctic sea ice extent and volume losses seem to accelerate, while temperatures go up linearly...


"But half of the end-of-summer summer sea ice found in the Arctic in 1950 is already gone."

In terms of the albedo feedback the September sea-ice extent is really not very consequential, as it is then around the time of the Autumn Equinox, and the Arctic isn't receiving that much solar radiation compared to the peak in June at the time of the Summer solstice.

That's why this melting season is so very interesting/scary. It has set large records for May, and could do the same for June - peak months for solar insolation at the Arctic. Thus far the rate of decline for these months has been much slower than for later in the melt season. If that is now changing with this year then we will really see the albedo feedback take effect.

If I had time, what I would want to do would be to calculate the theoretical maximum albedo feedback each melting season by integrating the sea-ice concentration observations with solar insolation (under a clear-sky assumption).

I suspect that this would show that, for all of the dramatic decreases in September sea-ice extent, we ain't seen nothing yet. When we do that will really heat the Arctic up - with obvious implications for the Greenland Ice Sheet.


There is an increasing energy budget and a decreasing amount of ice.

As the ice decreases, the extra energy, not required to melt ice (given that half the summer sea ice since the 1950's is gone), goes into making the environment warmer.

Which increases ice loss.

Naturally weather makes variability because it affects the energy budget. Therefore there are peaks on the decline graph which we see each season as surprises.

But, overall, I can't see how the end can be any different. Before ice will stop declining, the energy being input must fall. Not just seasonal weather blocking the input, but a long term fall like a century long maunder minimum.

In an environment where solar input is relatively stable and CO2 concentration continues to rise, then I can't see that energy budget falling until CO2 concentration falls unless there is some sudden influx of reflectants, multiple massive Volcano eruptions, continuously, over a 2 year period.

This is all without carbon sinks unfreezing, methane releases, soot accelerants in ice loss, land warming due to snow loss, cyclone intrusion and heat transport..... et al.

So that's my contention. The heat budget rises and the melting impact of the heat creates a self reinforcing reaction which accelerates the ice loss. So we see a constant heat budget increase but an accelerating loss of ice. In that scenario I cannot see any way that the Arctic won't transition to a seasonally ice free state with the current climate.

And let's be clear. We keep on pushing the climate further off the cliff as each year goes by.

I'm not saying next year or next decade. I'm just saying it's inevitable with the current climate. We don't need any more to transition it.

The fact that I believe it will happen in the next decade rather than in the next 5 decades is just a result of my observations.

Kevin McKinney

It's hard to disagree with Neil's points. One slight addition, though; there is a significant negative feedback which we've previously discussed here: that is that more open water in the fall also means more heat loss from the ocean.

That ties in with the comment about early melt, though--the implication would seem to be that seasons like this one are more damaging than what has bwen typical. Front-load area loss and you maximize the effect of the albedo feedback. Maybe you also increase ocean mixing in the CAB and get more warming at depth, too.


While I agree with you that using Archer's post from four years ago is flawed, there is much recent research to indicate that marine shelf methane is probably less of a problem than you think it is. The Center for Arctic Gas Hydrate has a good deal of expertise on the subject, as the following links indicate: 1. https://cage.uit.no/news/retreat-of-the-ice-followed-by-millennia-of-methane-release/
2. https://cage.uit.no/news/methane-not-escaping-into-the-atmosphere-arctic-ocean/
There are also atmospheric constraints: http://www.atmos-chem-phys.net/16/4147/2016/
Even so, you are definitely corrrect to say that we have to keep ch4 emissions in mind.

Chris Reynolds


Sorry for the delay in replying, work got in the way.

All of this is as a result of my looking at things from a temperature driving perspective following a recent discussion with you.

I do not think that a single function applied to September NSIDC Extent is informative. If I take the interannual difference of Sept Extent then make the cumulative sum of that, the behaviour becomes apparent.


Every year after 1997 is a negative cumulative sum figure, so I use that year as a break point, and there are two periods of different behaviour. Actually there are three, the post 2007 increase in variability (flat) is the third period after the 1997 to 2006 period of drop (linear).

This appears in the temperature extent relationship making the plot bow around the trend line.

That bowing might seem to support a non linear trend. However the early flat region is for the years before 1997,and it is certain years that pull the trend down from 2007 onwards. The presence of many mid points above the trend line is a result of the change in trend around 1997, not supportive of the application of one trend.

Mahlstein & Knutti 2012 look at a large set of models.
They adjust the models for known biasses and find that both the limited set of observational data and the models reach around 0 extent at approximately 2degC global warming.

I am very surprised that you (and many others) persist in claiming that 2016 will be a useful year to determine what will come in the loss of sea ice. From my perspective the highly anomalous warmth over the winter invalidates any conclusions drawn from what happens this year. April Arctic Ocean sea ice volume this year was within the range I said it would return to, but I have cautioned that this winter invalidates strong assertions that it backs my case regards winter growth of ice.

When such winter temperatures happen regularly, the pack will be in a far worse state than it has been following recent years (the volume increase), so what happens this summer doesn't even inform us as to the summers we can expect in future years when such winter temperatures are normal.

So if we have a record summer beating 2012 I have no doubt that it will be added to data and used in exponential fits to claim support for a rapid crash. But I'm sorry, I will not find such slack reasoning convincing at all.

Chris Reynolds

Sorry, I forgot to add the graph of extrapolation of temperature/extent.

Jim Williams

I still say we have baked in the Industrial Revolution and these numbers are all bogus -- and not in a good way. What we are seeing now are the results from what we did 150+ years ago, and what we are doing now won't be meaningful for at least another 150 years.

Jim Williams


"Is the ice response lagging or leading the driving of climate forcings?"

All the evidence from the failings of previous models and over-optimistic projections flowing from them is that not only is ice response a lagging indicator, it is a strongly lagging indicator.

I'm going off how bad the projections have been, as that is my best evidence in predicting the future.

Rob Dekker

Thanks Chris,
This graph you showed :

with this comment :

Every year after 1997 is a negative cumulative sum figure, so I use that year as a break point, and there are two periods of different behaviour. Actually there are three, the post 2007 increase in variability (flat) is the third period after the 1997 to 2006 period of drop (linear).

You starts to sound like one from Dr. Curry, about "regime change" at certain points in time.

To me, your graph shows quite clearly that ice losses are accelerating, which was my point to begin with.

So are we in agreement after all ?


Rob Dekker: I agree with Chris Reynolds. I mostly looked at September arctic sea ice volume loss, which I maintain is now about linear, if not decelerating. Here is my comment on this blog from March of 2014:


Being about halfway between last September and next September, I'm taking a look at how the game is going. The game is Match the September Arctic Sea Ice Volumes (in 1000's of km^3). The players are various types of curves. The scores are rounded standard deviations of the volume data from the curves. The lower the score the better. The players try to adapt to new data as it becomes available. Scores can go up or down over time.

Each player has its own score, but the players are on two teams of 4 players each: the Steepers and the Shallowers. The Steepers are projecting a progressively steeper decline into the future. The Shallowers have a progressively steeper decline at first, but then switch at some point (specific to each player) to a progressively shallower (less steep) decline.


Below are the players by team, along with their current scores.The estimated year of switch (a whole number year would correspond to September of that year) from going steeper to going shallower is also given for the members of the Shallowers, but if and when this switch has been made can be better determined in the post-game analysis (after September volumes are consistently 0).

Steepers and their Scores

Exponential _ 1.0089
Quadratic ___ 1.0040
Cubic _______ 0.9968
Hyperbolic ___ 0.9880

Shallowers, their Scores, and Estimated Time of Switch

HyperbolicTangent ___ 0.9841 __ 2009.31
HyperbolaDerivative _ 0.9783 __ 2007.87
ArcTangent _________ 0.9731 __ 2007.97
SidewaysRational ____ 0.9708 __ 2007.09

Recalling that lower scores are better, it looks like the Shallowers are ahead for now, but this could change. The Shallowers are betting that the rate of decline has already slowed.

There was another team, the Higher Degree Polynomials. They had better scores (in fact a sufficiently high degree polynomial had a score of 0, which is the best possible), but they were expelled for cheating. There is a rule against being too wobbly or having a turn that is too sharp, and members of this team broke that rule. Members of this team extrapolate poorly anyways, which is frowned upon.

Another player, Linear, didn't make a team during latest tryouts because of poor performance. Linear wasn't consistent with the theme of any of the teams anyways.


Exponential has a horizontal asymptote to the left. Exponential and Cubic are insisting that the slope should become steeper at an accelerating rate, especially Exponential. This could be detrimental to Exponential. It's in the nature of Quadratic to project that the slope should become steeper at a constant rate, and this game is no exception. Hyperbolic is the most Shallowers-like of the Steepers. Hyperbolic is counting on the slope becoming steeper at a diminishing rate, so that the slope approaches but never reaches a constant steepness. Hyperbolic has a horizontal asymptote to the left and an oblique asymptote to the right.

Each of the Shallowers' members has two horizontal asymptotes: an upper one to the left, and a lower one to the right. Each of these asymptotes can be at any position (so the lower asymptote can be at a negative volume). SidewaysRational is based on the curve (x / (x^2 - 1)), -1 < x < 1, rotated clockwise by a quarter turn. HyperbolaDerivative is based on the curve (x / SquareRoot(x^2 + 1)). ArcTangent and HyperbolicTangent are based on the curves that their respective names suggest. From the curves that these four are based on, both shifting and uniform stretching / compaction in the directions of each axis are allowed.


The players were polled with respect to their best guesses for the September 2014 volume (their guesses are shown below). None of the players seemed confident in their guess. "Our job is to try to look for the trend, not to predict the September volumes of individual years", remarked one of the players.

Steepers' Guesses for September 2014 Volume in 1000's of km^3

Exponential _ 2.520
Quadratic ___ 3.166
Cubic _______ 2.788
Hyperbolic ___ 3.087

Shallowers' Guesses for September 2014 Volume in 1000's of km^3

HyperbolicTangent __ 3.371
HyperbolaDerivative _ 3.492
ArcTangent _________ 3.586
SidewaysRational ____ 3.622

Note that each of these guesses is below the record low of 3.787 from September 2012, and well below the 5.479 from September 2013.

If the September 2014 volume (in 1000's of km^3) is between about 2.50 and about 10.15 then the order of the players from best (lowest) score to worst (highest) score will remain the same as the current order from best score to worst score, which is {SidewaysRational (Shallowers), ArcTangent (Shallowers), HyperbolaDerivative (Shallowers), HyperbolicTangent (Shallowers), Hyperbolic (Steepers), Cubic (Steepers), Quadratic (Steepers), and Exponential (Steepers)}.


Each of the current top five players believes that the beginning of when September arctic sea ice volumes are more likely than not to be virtually 0 (no arctic sea ice except for icebergs, ice shelves next to land, etc) will probably be after 2018. If this is the case then it is too early to predict, based on the game so far, whether it would be a year later than 2018 or decades later than 2018.

Rob Dekker

D_C_S produced a lot of text in the 30 min after my post, but the essence is captured here :

I mostly looked at September arctic sea ice volume loss, which I maintain is now about linear, if not decelerating.

September Arctic sea ice loss is quite well captured in this graph from PIOMAS that I showed earlier :

Curious where the "linear" and "decelerating" statements come from given this evidence.


All of this of course presumes that PIOMAS is correct and a reliable indicator. We have several ice thickness models. They tend toward grouping into two sets. However there is broad variation among them in predicting or estimating the thickness of the ice at various locations across the Arctic. Some are at times directly at odds with visual evidence.

I do not believe it is a safe bet to presume the accuracy or reliableness of any of them. Each does provide some insights that are useful for asking questions. In the large course scale they do track the general trend. But I certainly wouldn't bet the future on their accuracy.

More than that, I would encourage everyone to always remember the limitations of these models and the assumptions and presumptions that they are based upon.

As we get closer to the first ice free Arctic summer, it seems highly likely that we will see increasing variability as the ice thins, both in reality and in the models behaviors. These variations are vastly different from one another. The physical variations occur for all the many reasons talked about here. The model variations add to that dependence on the assumptions upon which they are based.

It is perhaps interesting that we are indeed seeing greater variability from the trend lines for all of the fit models. That is partly real and partly possibly artifactual. However, that increased variation taken as increasing random or pseudorandom variation also affects the estimated curves of the predictors, both means and standard deviations. Until we reach zero for several periods we will not be able to know what the actual curve was that these variations vary from. Until then, the predictor equations will have changing parameters even if we are only seeing random variation at work.

Trying then to judge which is right before the end may leave us chasing Aluce's rabbits down so many holes, and in a backward foot race with the Red Queen.



Or, perhaps as a more apt analogy, this may leave us trying to determine the best and most accurate counting technique for counting angels dancing on the heads of pins. That might involve all sorts of interesting math and speculation while serving no real meaningful purpose.

The ice is going. The ice is going faster, and faster, and faster yet. And here we are transfixed and unable to look away. All the while, the catastrophe that lies just ahead is perilously close, and closer at an increasing pace as we continue forward at breakneck speed into our peri less future. Perhaps someone might just apply the brakes. Please. Pretty please. Now! Please?


Rob Dekker

Quite interesting.

Just like D_C_S, Sam produced a lot of text within 30 minutes of my post. This time, without producing ANY counter to my claim that Arctic sea ice extent is accelerating, both in volume and extent. Just arguments about why volume estimation cannot be trusted.

I made my point.

I'm looking forward to Chris' response.


The september ice loss will be approaching zero when there are little ice to melt after having open water in august.



I agree with you. I suspect that PIOMAS is under representing the melt, and it is melting faster and faster. The positive feedbacks have kicked in. Still, the annual variation is large enough that even after we hit zero for several consecutive years, we may still not be sure what curve best represents the data.

Either way though it doesn't matter. In the most extreme case we might be talking about a ten year difference. In even the time since humans began growing crops circa 10-11k years ago, this is only 0.001% of that time. In the time since we began burning fossil fuels in a big way (circa 1800) this us only ~5% of that period. The point is, this is an incredibly rapid shift and the difference in when we hit zero whether estimated one way of another is completely unimportant. We -are- going to hit zero very quickly now. That is huge. That is the important story.

And rather than that causing humanity to change course, we are doubling down on burning it all faster. That is insane.



0.1%, 0.001 fraction.

Rob Dekker

Sam said :

I agree with you.I suspect that PIOMAS is under representing the melt,

No Sam. PIOMAS is not under representing the melt and neither did I claim that.

I claimed that both ice extent and ice volume are accelerating, and so far neither Chris, nor D_C_S, nor you have brought forward any argument that disputes that.

Just massive amounts of text that oddly enough seem to be presented within 30 minutes of my posts.

Rob Dekker

I smell a rat in this thread, so let me summarize :
Even using Chris's own graph :

it is quite clear that Arctic sea ice losses are accelerating, and this is even more clear in the PIOMAS volume graph I showed above.

Even without going into the cause (which I think is albedo feedback, led by snow cover decline) can we just simply agree on that simple fact ?


Everyone smells rats and not only in this forum!

Coal is dying and resource bottlenecks are coming: everyone is preparing for panic!

Did renewable energy increase 10% globally last year?

(More people are reading these blogs every day and some don't appreciate them one little bit!)

Chris Reynolds

Jim Williams,

Sorry, but none of that tells us anything about the relationship between forcings and the response of the sea ice.


There was a regime change in the mid 1990s, that is what that treatment of September extent is picking up. The regime change is IMO best explained by Lindsay & Zhang.

L&Z find ice/ocean albedo feedback as the driver. Not snow cover. They also find a substantial increase in upward longwave associated with winter ice growth - which might explain why a linear hypothesis with a break point at 1997 seems to work rather than the (1-exp) form of exponential crash equation that might simplistically be assumed to apply.

As I said above, the trend since 2007 in Sept extent is level if it's a rise or a decline (uncertainty allows either) it may not be declining at the same rate as in the preceding period. Time will increase the statistical confidence to ascertain properly what is happening post 2007.

So we have a strongly demonstrated inflection of around 1997, the year after which all years of the cumulative sum of interannual differences are negative, and which ties in with the mid 1990s start of an ice/ocean albedo feedback driven decline proposed by Lindsay & Zhang. This justifies the breaking of trends and application of a piecewise linear trend around that point.

That isn't the same as an exponential trend. Nor does it imply self-acceleration to zero. Eisenman has shown that previous 'toy model' studies finding a fast exponential type crash to zero lacked both seasonality (winter growth feedback) and the decline of poleward heat transport with a relaxation of the pole/equator temperature gradient.

Those factors apply in the real world, so their consequences probably do too.


Thanks for that, an interesting read.


While I appreciate your urgency, all the evidence I've seen from reading various articles on climate news sites and posts by climate scientists indicate we're not doubling down on burning everything. Quite the opposite actually


The only thing we globally have done is to reduce the rate of increase. And that is only do to recessionary forces.

Though there has been huge investment in wind and solar, these are far outstripped by the actions that would be required to be meaningful.


Rob Dekker:

My previous comment does dispute your claim.

Curves that accelerate and then decelerate fit the data better than curves that accelerate and then continue accelerating.


In this article from a few months ago, statistician Grant Foster shows that arctic sea ice extent loss accelerated, but then suggests that it stopped accelerating, and appears to have even decelerated:


In this graph, he shows a relatively slow decline, followed by a rapid decline, followed by a relatively slow decline:


His article contradicts the claim of a continued acceleration of the decline in arctic sea ice extent.


Hi Kevin.

Yes there is significant heat loss. But there is significantly more heat to lose.

This then leads to significant moisture locally in the area which can, with the right weather patterns, impact the onset of winter..

Which can tend to negate the heat loss overall.

In the end my contention was that even with the existing heat budget it is enough to finish the job. We don't need to keep pushing CO2 out, the existing CO2 is there and will do the job just fine.

However the variability of weather and solar is just enough to keep an unreasonable doubt alive. Which blocks action on further CO2 increases.

Those further CO2 increases just mean the difference between 2030 and 2060.

The end it not in doubt. Just the timeframe.

Also the onset of another all season arctic ice cap will heavily depend on just how much more CO2 we push up there. Really the difference between hundreds and thousands of years.


With all due respect, emissions stalled in 2014 and then dropped slightly in 2015. Even Michael Mann said that in his latest interview with Thom Hartmann. I'm pretty sure no recession was going on then



To the contrary. See below.

And this is without accounting for man made methane releases, or any of the natural releases and feedbacks, or destruction of the natural sources and sinks.

Even then, what is required is not slowing of the increase, or even stable emissions. What is required are dramatic year on year reductions.

Add to this the huge positive impact from the loss of sulfate aerosols and particulates as reductions occur, and even more dramatic reductions are required. To even make a dent, these likely need to exceed 10% per year sustained until emissions are less than a few percent of current emissions. Deciding to do anything less means that we return to Eocene conditions. Even these levels of commitment are likely not enough.


Global Emissions
2014. 9.795 GtC
2013 9.735 GtC
2012 9.575 GtC
2011 9.449 GtC
2010 9.995 Gtc



2010 above includes 0.855 Gtc of natural sources and needs to be reduced by that amount for an apples to apples comparison.

For 2015 EIAs preliminary estimate (likely somewhat low) is for a minor increase.


Year GT CO2 (not GTC as above)
1978 17.44
1979 17.96
1980 17.78
1981 17.54
1982 17.36
1983 17.47
1984 18.05
1985 18.32
1986 18.65
1987 19.28
1988 19.94
1989 20.33
1990 20.62
1991 20.75
1992 20.70
1993 20.81
1994 20.93
1995 21.48
1996 21.99
1997 22.23
1998 22.38
1999 22.52
2000 23.32
2001 23.58
2002 23.94
2003 24.99
2004 26.18
2005 27.05
2006 27.86
2007 28.78
2008 28.87
2009 28.32
2010 29.84
2011 31.29
2012 31.49
2013 32.07
2014 32.13
2015 32.14

The reduction in increase comes mostly from the economic conditions driven in large part by the US attempt to destabilize Russia which spiraled out of control with the Saudis increasing production leading to falling prices as they attempted first to shutdown Russia (a major competitor), then Iran (a major adversary), then the US fracking industry (viewed as another competitor) and finally Venezuela.

Even with such massive changes in global supply and demand as economies stall, and as wind and solar ramp up, the best we could do globally from anthropogenic estimated releases was to nearly but not quite stop increasing emissions.

And this is only anthropogenic CO2 releases. It does not include huge releases of methane from Aliso canyon, from the fracking boom, or natural sink losses from the massive fires all over the place. These both serve to destroy the sinks, and to convert them into additional sources.

Add to that the effects of global warming intensified El Niño, and the conditions are much worse - the the highest year on year increase of atmospheric CO2 (using Mauna Loa as a surrogate) that we have seen.

It is now very likely that no weekly average CO2 measurement at Mauna Loa will fall below 400 ppm ever again in our lifetimes, or even in the next millennium. We are now at nearly 408 ppm CO2 there, 128 ppm above the peak baseline coming out of the last ice age.

We should be on the slow decent back to ice age conditions with falling CO2 over the next 90,000 years. Instead we are in a rocket to the sky.

Add the several hundred ppm of CO2 equivalent from methane and other warming gases and you might begin to get the idea that we are in some sort of trouble.

The consequences of all this northern hemisphere heating is devastation of the arctic ice and a high probability that we will rapidly convert to a mixed mode globally with the northern hemisphere experiencing an equable climate with a single Hadley cell extending all the way to the North Pole while the Southern Hemisphere maintains a three cell circulation. Should we have pushed far enough and hard enough to melt Antarctica over the next many millennia, the whole earth might then shift to equable climate conditions.

Fortunately for us, Antartica is a large heat sink and it will take an immense amount of heating to destabilize the southern atmosphere.

As a result, the Southern Hemisphere may become a refuge. Or perhaps not. The oceans may play the critical role there.


What you're saying is simply untrue: https://www.technologyreview.com/s/601415/carbon-dioxide-emissions-keep-falling-in-the-us/
Emissions slowing is not happening due to recession at all. You have provided no evidence to support that claim.


You also seem to be making a lot of projections that aren't necessarily inevitable. I would go so far as to say they might be extreme. And there was a study done that posited that emissions did fall: https://www.greenbiz.com/article/low-carbon-economic-growth-really-possible


And I don't know where you're getting the hundreds of additional Co2 equivalent PPM. That's simply hogwash. Those gases are already accounted for, and even then it's a bit misleading: http://www.climatecentral.org/news/scientists-seek-new-measure-for-methane-20413



First - begin with geography. The United States is -not- the world, much though some might believe it is.

Second - do simple accounting. The global warming gases include more than co2 and methane. They also include water vapor, the largest impacter of all. The changes in the others move massive changes in water.

Third - look to geologic history. Back up through time to higher GWG levels and watch as first the north polar ice goes away and becomes an equable climate about 3 M years ago. (Advances and retreats 2-5 M years, and all gone prior to 5 M years). Continue back to 25 M years when Antartica was iced over and the period from 25-38 M years when it formed. Then finally go back to 40 M years before present and before when there was no ice. The whole world was in an equable climate.

It takes very little effort to compare current and trended GWG levels and durations and positive feedbacks to conclude that barring massive immediate changes that we will achieve GWG levels that will eliminate the Arctic Ocean ice and rapidly melt Greenland. These levels will remain high enough long enough that Greenland will fully melt out. Voile - northern hemisphere equable climate.

Is there enough there long enough to melt Antarctica? Good question. I suspect not. But the release of 1,650+ GTC of tundra carbon and massive amounts of oceanic methane as the oceans warm in response to the warming atmosphere is going to make a huge dent in the Antarctic ice.

Already we are seeing the major glaciers that block and protect the ice failing. We should fully expect to see West Antartica go ice free. Rising sea levels will aid in releasing Thwaites and others to dump an additional huge amount of ice into the oceans as water further raising ocean levels.

The transitional period will be 'challenging'.

Next look at the rate of change through this whole period. The challenge to ecosystems and biomes will be extreme. We already are in the beginnings of one of the half dozen greatest extinction events in all of world history. We are rising far faster than the PETM. Far faster is an absurd underestimation of the rate of change.

And with that my friend you tire me. Enough.

Ron Mignery


I have some questions. You suggest that the northern hemisphere would go to an equitable climate with a single Hadley cell circulation when the Arctic ice and Greenland ice disappears. Is it known how rapidly that would develop? Would the northern tropics then cool to the new average and if so how much? Is Greenland ice relevant since the average elevation of the ice sheet is over a mile?


Our collective knowledge about how the atmosphere and ocean work in an equable (not equitable) climate is quite limited. We can surmise things from sediment and soil cores, diatom and related counts, isotope ratios and a few other inferences.

Even with all of that, I am unaware of an adequate complete model to explain the details.

What we do know is that the far north becomes quite warm. The tropics don't cool. And the mid latitudes apparently are semitropical. What it looks like is that the pole goes very warm which removes the major atmospheric heat engine. Coriolis forces still apply.

Lacking jet streams which occur at the cell boundaries (absent in an equable climate), storms likely behave very differently from what we are accustomed to. Our accustomed climate bands (rain, desert ...) likewise go out the window. What that means for biomes is an open question.

Can we reliably grow crops in such conditions? Who knows? Growing crops requires the right mix of temperature, rain, soil and sun in the right times (for which the plants have become adapted). Will those be consistent year to year? Again, who knows?

What will this do to animals, insects and diseases? Who knows?

The transitional phase is likely to be the worst though, as the system vassilates between states and through unstable intermediate conditions. How long will that last? Again, who knows? I suspect that Greenland will play a large role in that until it completely melts. After that, the oceanic currents and the atmosphere can stabilize into a new operating mode. Will that take a millennia or a century? Good question.

Andy Lee Robinson

Bobcobb - the rate of change of emissions are decreasing, but emissions are still happening and increasing the concentration of CO₂ in the atmosphere.

Even if it were possible to emit nothing, the consequences of what we have already emitted are still enough to postpone at least the next glacial period, and yet to play out.

Per Tzedakis et al 2012, "glacial inception would require CO₂ concentrations below preindustrial levels of 280 ppmv"

We would have to "unburn" our fossil fuels at a rate much greater than we burnt them to have any chance of restoring the atmosphere's chemical composition to one similar to what all life on Earth is currently evolved to, and to do it before consequences become burned in.

Almost everything we do generates CO₂ - imagine everything we do sequestering CO₂ instead. It's not gonna happen.


Thanks, Andy, you're right. Bob cited one very good positive article: Renewable energy can do the job. Avoiding CO2 levels above 500ppm is possible without wrecking the world's economies.

Keeping CO2 levels below 500 ppm will avoid the worst possible climate catastrophes (Google wet stratosphere) but levels between 400 and 500 ppm haven't been seen since the Miocene or earlier when the climate was much warmer than today.

What's hard to figure is the effect of the extraordinarily rapid rate of change in GHG levels. Even the PETM may have been slower. When change happens very rapidly in the geologic record very bad things tend to happen as oceanic and atmospheric chemistry go into the "red zone". Hydrogen sulfide is not your friend.


Have you not paid any attention to the studies and articles I have cited regarding oceanic methane release? I'll say it again, you make a lot of assumptions without backing it up. I seem to remember Chris Reynolds questioning them as well.
Your claims regarding emissions and renewables are simply false. There was no recession in 2014 or 2015. It's not "enough" for me because you make patently false claims and extreme predictions which you don't provide any convincing evidence for.


Yes, andy, you're right, but the whole point is to reduce emissions and avoid as much damage as possible.


Pardon me for saying so, but I'm pretty sure the Canfield ocean you're referring to isn't going to happen for a long time, if at all



I won't play - 'bring me another rock' with you. That you don't want to change and don't want to believe you (and we all) have to is clear. Keep your head in the sand all you like. Not knowing or not accepting won't save you.


Your flair for the doom amazes me. Michael Mann, who I dare say is a lot more qualified than you are, views things quite differently: http://www.rivernewsonline.com/main.asp?SectionID=6&SubSectionID=59&ArticleID=72852

Andy Lee Robinson

Well, it's good that we're all agreed on having to reduce emissions, but I still don't think it will be enough to save the summer Arctic and avoid damaging sea level rise and the other headless horsemen.

Those addicted to their power and influence derived from fossil fuels, really really don't want to do anything about it and will continue dragging their feet to the very end, by which time aggressive positive feedbacks will be kicking in and perhaps remove any ability to influence the outcome.

As it seems that the survival of most of our species is not enough of an incentive to act, we have but a short window of opportunity to make greed work in our favour, and make renewables an irresistible investment - there are encouraging signs, such as the 3¢/kWh deal for a solar plant in Dubai last week.

I'm fairly optimistic that the energy transition and revolution will gather pace and overwhelm the entrenched efforts to suppress it - the scope for growth and profits in renewables are far higher than what can be expected in a mature but dwindling FF sector.

As we bump up against the walls of our Pale Blue Petri Dish, our future can only be renewable, or not worth having.

If the world is run by solar and wind barons, at least the air will be easier to breathe!

There was no recession in 2014 or 2015.

Not in the books, no. :-B


I cannot agree with you more. In my humble opinion, whether the renewables surge continues to gain steam or not, particularly in the States, really depends on the election here. I realize some environmental proponents aren't excited about Clinton, but her policies are actually pretty good. I think the chances of a Trump win are slim to none. Not even the electorate here, stupid as they can be sometimes (or pardon my language, but the idiots who don't bother to vote in the midterms that left us with Repugnant majorities in Congress), are going to make that bad a mistake.

Lawrence Martin

June 5

June 6

Can you spot the subtle change?

Susan Anderson

re Tamino and ice loss. There is a curious misstatement above. I keep up with his articles and at no time could anyone credibly indicate he's promoting the idea of overall recovery of Arctic ice. He's a statistician who does top level work, including in-depth analysis of trends. Here's a more recent article; he appears, like most of us, to be watching in bemused worry what is going on now.

As for Mike Mann, he too has been misrepresented. Check the subtitle in the link:
"Michael Mann: Energy markets moving toward renewables; Pace not fast enough to avoid climate-change consequences"

It is typical but irritating to see experts some of us know well cherry picked or even mildly misrepresented. The superior tone is not justified. He doesn't encourage negative attitudes, but he is also not interested in ignoring the scale of the problem.

Susan Anderson

For more from Tamino, he's written more recently. Here's one more:



Bob Cobb is a concern troll who has posted here and on the forum under different names such as BryantFinlay, Bryant Morganelli and Bryman. I'm fine with people changing names, or posting under different ones, but let people know, so they know who they are talking to. This guy has never done that.

I guess his goal is to sow discord and spoil the atmosphere on blogs and forums, setting up strawmen and provoking. I don't know if it is because he's a cornutopian or a (paid) climate risk denier, and I don't care.

Sorry for not catching this earlier (someone mailed me about it), especially to you Susan, who took the effort to engage and keep the peace. I suspected Bob Cobb was a concern troll, but didn't know it was the same guy I had blocked/banned previously.



Susan Anderson: Neither Grant Foster nor any of the people commenting here claimed that there was a recovery in arctic sea ice extent.

Chris Reynolds


To add to DCS, just because some of us think what happens next isn't a fast crash, that should never be taken as saying there is a recovery.

Arctic sea ice is in terminal decline and anyone who says it is recovering immediately writes themselves out of informed discussion of the matter. Claims of a recovery can only be made by people with a total lack of grasp of the evidence.

Bill Fothergill

@ D_C_S & Chris Reynolds & Susan Anderson

There may be a bit of mutual misinterpretation/misunderstanding at play here.

I think you guys (and most others on Neven's site) are singing from the same hymn sheet, but have subtly misread each other's comments in this instance.

May I suggest re-reading the D_C_S comment concerning Tamino (5th June 22:31), then Susan's responses (7th Jun 08:11 - 08:16), and then the rejoinders by D_C_S and Chris immediately following Neven's "bobcobb exposure".

You are actually agreeing with each other - without realising it.



Bill Fothergill: I read those comments again. I noticed that there is some content about arctic sea ice and some content about what was or wasn't claimed here about arctic sea ice. I suppose that one would need to practically be a bit of a logician to sort it out with respect to those two topics. This comment is neither about arctic sea ice nor about what was or wasn't claimed here about arctic sea ice. How's that for logic? Also, this comment self-references, and refers to its own self-referencing. See "Gödel's incompleteness theorems", if you wish. I've gotten way off topic. I think that it's time to end.

Jim Williams

What people who haven't been watching don't understand is that we who have been watching are holding our breath. The best way of seeing why is by looking at the NSIDC extent graph. 2012 had started turning down by now; which hasn't happened so far this year. At the same time, we are currently just about at the point where 2012 was after it turned down.

Does this year's extent stay "flat" and run into 2012's? Or does it soon do the same thing 2012 did?????

(Ignore the divergence and increase in extent in the central ocean over the last few days....that is just to confusing to contemplate.)

Susan Anderson

Bill Fothergill is right. Being an amateur and aware of my ready tendency to step in where angels fear to tread, I took the trouble to provide links; I keep up with Tamino and knew he had more recent material. I think we should call him Tamino and not Grant Foster, since that "outing" was a product of WattsUpWithThat; when I see that it makes me suspicious. That's why I was so ready to look carefully at the source and make sure it's not coming from fake skeptic central.

Separately, I didn't like the misrepresentation of Mike Mann's careful efforts to maintain a positive outlook and approach. Some time ago I had a bit of a wakeup call when an Australian artist pointed out that apathy and despair were forms of laziness. I fear we are in trouble, but that's no excuse not to go on trying. Fact is, it's life and death, but many people don't see it that way.

D_C_S I'm sorry if I misunderstood; I thought you were claiming that there had been a recovery, and citing Tamino as evidence. If that was not the case, this was my bad. Sincere apologies.

Neven, I am wary for the above reasons of weighing in - I've been accused of being a kind of "hall monitor" by denial advocates. I rather like the exercise, but I hadn't seen him as wrong, just hotheaded. Glad you cleared that up. He has also shown up at aTTP's.

Susan Anderson

Chris Reynolds, I can't find it, but I know I said just that recently. Neven is such a gentle person, without appearing to criticize, he redirects, but you all shouldn't hesitate to correct me when I overrun the boundaries of understanding. I'm happy to be able to tag along, mostly.


Susan Anderson,

He put his name in the image in this post on his blog:



-Fish aka George here. Thanks, Neven for outing the concern troll. His ad hominum attacks gave him away. That was classic trolling.

The great thing about this blog and the forum is the cooperative effort. The troll was trying to disrupt the cooperation.

As Chris Reynolds notes, our discussions and disagreements are about the rate of the decline, not the reality that Arctic sea ice is in terminal decline.

The GFS is now showing the Beaufort high rebuilding at some point after 240 hours out. I've tried to find the paper, with no success about the summer cycles of high and low pressure over the Arctic. Now I think it is 2 weeks high and 2 weeks low, not 6 weeks as I said in a previous comment. If someone knows about that paper or the cycles, it would be good to review the details.

It looks to me like a battle is developing with 2012 and the weather cycles will play a big role. This year the jet stream pattern is not like 2013 which had a very cool stormy June. There is much more heat being advected over the Arctic ocean this year.

Susan Anderson

D_C_S: It's not a savory history; I remember it well because McIntyre was vile about Shakun and Marcott on DotEarth; McIntyre's troops made it personal. Revkin covered himself with shame by supporting McIntyre and the hate doubled overnight.

"McIntyre charges Grant Foster aka “Tamino” with plagiarism in a Dot Earth discussion"

I'm not providing the WUWT link because the whole thing was disgusting. Yes, of course, Tamino had no choice after that but to make the best of it.

Susan Anderson

One more, this from Tamino, then I will cut it out.

He [Roger Pielke Jr.] has accused me of a lack of “professional courtesy” for such horrible deeds as blogging under a pseudonym. He even went to the trouble to dig up my real name and post my hometown location on his blog. How professionally courteous of you, Roger. That certainly advances our understanding of sea ice trends.

I knew that Jr. Roger Pielke was a deceitful manipulator but I didn't know he would stoop that low. Obviously Pielke wasn't able to make a successful scientific argument so he dived to the gutter.

Thanks, Susan for that bit of history. -Fish.

Rob Dekker

D_C_S said :

My previous comment does dispute your claim.
Curves that accelerate and then decelerate fit the data better than curves that accelerate and then continue accelerating.

Dear D_C_S. You are just curve-fitting with arbitrary formulas, none of which have any relation to the physics of ice melt in a warming world.

Chris said :

There was a regime change in the mid 1990s, that is what that treatment of September extent is picking up.

Chris, "regime change" is a very un-satisfying term, as it has no relation to physics. You mention Lindsay & Zhang 2004, who report that feedback due to ice losses is amplifying ice losses since the mid 1990's, which actually makes my point that albedo change causes the September losses to accelerate.

I am totally at a loss as to why you believe that there is a third "regime" that sets in after 2007, and " the trend since 2007 in Sept extent is level". Did the that feedback that Lindsay & Zhang were talking about suddenly come to a grinding halt after 2007 ?

I Ballantinegray1

Hi Rob D!

After 07' folk were clamouring for recovery b( some folk) but I had it in mind that part of the extent/area was the 'collapse and spread' of older floes.

A few years later Prof Barber confirmed that this , to some lesser or greater extent, had been occurring and brought us the new ice type of 'rotten ice'.

If they was another 'step change' in 07' it was the opening up of the basin so ass to allow this 'collapse and spread' phenomena to play out. 2010 then took a good portion of this ice when we saw the record low volume?

The basin ,today, sees the lowest volume/extent for the time of year. surely this makes for an interesting Sept even if we end up with another 'average' weather year? The fact so much dark water has been open for so much longer must build up a greater 'melt momentum' than previous years could with less 'dark water' open to the sun?


Rob Dekker:

Those curves weren't about trying to predict the future. They were about what has already happened. The article by statistician Grant Foster that I posted is also about what has already happened with respect to arctic sea ice extent decline.

Here is that article again:


Here is that graph again:


Also, physics-based models generally predict a slowing in arctic sea ice decline. So it's both predicted and observed.

John Christensen

Hi D, you said:

"Now I think it is 2 weeks high and 2 weeks low, not 6 weeks as I said in a previous comment. If someone knows about that paper or the cycles, it would be good to review the details."

While I have not seen a paper, it is correct that there is a jigsaw movement - this is what Edward Lorenz spotted many years ago.

I see the AO most typically having frequent spikes, roughly going down for one week, then up for one week, so that you have about two weeks between each negative or positive peak.
Then at times, the jigsaw slows down, and you can have about 4 weeks between each positive or negative peak.

Here you can see the normal/frequent peaks during spring, but then it slowed down, so we had a positive peak mid-May and are headed for a positive peak mid-June. After this, the forecast is for normal/frequent transition again:


I am sure you can then run statistical analysis to discover other patterns, as it seems like the more extreme peaks appear with possibly 8-10 weeks between e.g. each positive extreme peak like the one we had mid-May.

I hope this helps.


Rob Dekker: PS

The graphs in your comment of June 05, 2016 at 06:55 also use curve fitting, but graphs that accelerate and then decelerate fit the data better than graphs that accelerate and then continue accelerating.

Rob Dekker

The graphs that I showed on June 05 are these :
Of volume (from PIOMAS) :

and of extent (from Chris Reynolds) :

Neither one of these shows that "graphs that accelerate and then decelerate fit the data better than graphs that accelerate and then continue accelerating".

In fact, only an accelerating trend (as shown by Chris Reynolds' two piece linear trend lines) is significant. A "decelerate fit" at the end of the series is not statistically significant, and also there is no physical reason for deceleration.

D_C_S also said :

physics-based models generally predict a slowing in arctic sea ice decline. So it's both predicted and observed.

No, D_C_S. The physics-based modes do not predict a slowing in arctic sea ice decline, at least not until sea ice extent at the minimum gets well below 4 million km^2 (which it has been only once so far).

Instead, the models predict increased "variability", (due to... you guessed it : albedo amplification) which means that warm summers get lower extent and cooler summers higher extent. 2012 was and example of of the first, 2013/2014 were examples of the latter.

The "mean" however will continue to accelerate. No "slowing in arctic sea ice decline" there.


Rob Dekker:

Obviously a curve that only accelerates the decline is going to have a best fit that accelerates the decline! That does not imply that curves that accelerate and then decelerate don't have a better fit.

Grant Foster shows acceleration and then deceleration. My calculations suggest the same.

You just agreed with me that models may PREDICT a slowing of the decline. The data shows a slowing of the decline already, which may or may not last.

Chris Reynolds


"Chris, "regime change" is a very un-satisfying term, as it has no relation to physics. You mention Lindsay & Zhang 2004, who report that feedback due to ice losses is amplifying ice losses since the mid 1990's, which actually makes my point that albedo change causes the September losses to accelerate."

No relationship to physics?? Have you even bothered to read L&Z?? Their hypothesis is that the impacts of the AO and PDO in the mid 1990s set into play a new regime of sea ice albedo driven volume loss. If you're looking at PIOMAS volume data, the key paper to understand it is L&Z.

Your point regards albedo loss has been about snow albedo on land, has it not? L&Z find that in the PIOMAS model the decline of volume has been driven by sea ice albedo and ice/ocean albedo feedbacks.

On this subject you might want to consider Lawrence 2008, paragraphs 7 & 8, and figure 2.
That supports the interpretation that loss of sea ice drives the adjacent land warming.

"I am totally at a loss as to why you believe that there is a third "regime" that sets in after 2007, and " the trend since 2007 in Sept extent is level". Did the that feedback that Lindsay & Zhang were talking about suddenly come to a grinding halt after 2007 ?"

No. But the albedo feedback acted preferentially on multi year ice not first year ice. This is due to the differing time constants (memory) of the two classes of ice.

This is explained by Armour et al 2010, "Controls on Arctic Sea Ice from First-Year and Multiyear Ice Survivability."

"Given the strong thickness–growth feedback of sea ice (Bitz and Roe 2004), where in a warming climate we can expect the thicker MY ice to thin at a greater rate than the thinner FY ice, and the fact that the ratio of MY to FY ice entering into the MY ice category each year is decreasing, it is likely that the difference between FY and MY ice survival ratios will decrease in a warming climate. If this occurs, the Arctic sea ice system would move toward a regime of decreased memory and decreased sensitivity to climate forcing..."

Note that last sentence: "...a regime of decreased memory and decreased sensitivity to climate forcing..."

It is a vitally important point! If you don't get it, read Bitz and Roe 2004 "A Mechanism for the High Rate of Sea Ice Thinning in the Arctic Ocean."
Once you understand Bitz & Roe you will understand the statement by Armour. If you read Bitz & Roe and still think the statement by Armour is wrong - you haven't understood Bitz & Roe.

"Neither one of these shows that "graphs that accelerate and then decelerate fit the data better than graphs that accelerate and then continue accelerating.

In fact, only an accelerating trend (as shown by Chris Reynolds' two piece linear trend lines) is significant. A "decelerate fit" at the end of the series is not statistically significant, and also there is no physical reason for deceleration."

This is just not correct.

The first breakpoint of my two trend fit is strongly supported as outlined previously in this thread.

There is a second potential breakpoint in 2007. However the noise in September extent since then precludes firm conclusions as to the reality of a relaxed rate of decline or even levelling since then. That is not the same as saying that it is not there. Livina and Lenton 2013 find an abrupt change of statistics after 2007.
And as DCS points out both he and Tamino seperately find suggestive evidence of a recent relaxation of trend in extent loss.

Post 2007 the pack has consolidated a transition to a predominantly ice free state. First year ice has a much lower rate of thinning than does multi-year ice. Since 2010 most years have seen April PIOMAS level to around 19.3k km^3. There are physical reasons to expect a third regime of curve inflection.

At this point I should make clear that I am repeating myself. I noticed that you tried to dismiss Lawrence back in 2014 in comments here (final comments).
You also made the same mistake regards L&Z's findings back then as you seem to be making now (in L&Z the loss of ice is due to the albedo feedback in the ocean, not sensible warming from adjacent landmasses).

So, I'm sorry, but but as I am not interested in circular discourse further discussion seems rather unproductive.

I'll leave you to reply as you see fit.

Rob Dekker

We do not disagree on the conclusions of Lindsay & Zhang 2004.
Your and my statements are identical, and explain the albedo-induced acceleration of extent seen in your graph :

And regarding Lawrence 2008's conclusion that "loss of sea ice drives the adjacent land warming." Note that Lawrence was talking about AUTUMN, not spring. So your statement that :

"I think the retreat of snow in summer is being caused by the retreat of sea ice."

is not sustained by Lawrence 2008, and as I noted on your blog, I think you have it backward : Spring and early summer snow retreat facilitates Arctic ice melt. Not the other way around.

Then regarding your argument for a new regime that causes ice extent after 2007 to be flat, you use Armour et al 2010 as an argument, and specifically this statement from that paper :

it is likely that the difference between FY and MY ice survival ratios will decrease in a warming climate. If this occurs, the Arctic sea ice system would move toward a regime of decreased memory and decreased sensitivity to climate forcing

Now, I did not read that paper before, so I looked it up, and found it here :

I read the paper today, but I did not find any evidence for the claim that " the difference between FY and MY ice survival ratios will decrease in a warming climate" nor that "the Arctic sea ice system would move toward a regime of decreased memory and decreased sensitivity to climate forcing".

In fact, tracing the core of their argument, I ended up in this section on page 17, which states :

If this occurs, the Arctic sea ice system would move
toward a regime of decreased memory and decreased sensitivity to climate forcing (consider Fig. 5, or the equations in Appendix A, as β
g → α and βm → α).

So now I need to dig into Appendix A to find any evidence for your claim of a new regime after 2007.

Chris Reynolds


"Note that Lawrence was talking about AUTUMN, not spring."

Staggered. I am really staggered.

Figure 1 panel b) (graph) and more to the point - figure 2 panel a). April/May both show anomalous warming.

"I read the paper today, but I did not find any evidence for the claim that " the difference between FY and MY ice survival ratios will decrease in a warming climate""

Then you just don't get it, or more to the point don't want to.

Do you think that you'll get closer to reality by arguing like a lawyer? There really is no point in discussing anything with you.


Chris, yes that paper does cite evidence that there may be a slowdown in volume loss with the transition to first year ice and states:

"If the memory time scale and mean state sensitivity of Arctic sea ice decrease sufficiently quickly under a warming climate, a slowing in the rate of area and volume loss could occur.

This is consistent with the characteristic trajectory of September sea ice area decline in twenty- first century simulations where the rate of change of Arctic sea ice area decreases late in the simulation despite a continued increase in climate forcing (e.g., Fig. 6).

An inflection point in ice decline could also occur if during the warming the ice survival ratios decrease quickly for some time and then slow in their rate of decrease at some later time.

It is the interaction between the trends in the survival ratios (the forcing) and the memory time scale (the sensitivity to forcing) that determines the trajectory of ice decline, and further study with a FY ice tracer within fully coupled climate models is needed to determine the exact reason for the inflection point seen in climate simulations. To establish whether we should expect such a trajectory in the observed sea ice system, we should attempt to estimate the mean state and trends in Arctic sea ice survival ratios from observations and compare these with the results of coupled simulations that exhibit such behavior."

But, note early in the paper the authors state that the model does not account for changes that may be taking place in the oceans.

"However, the model does not reproduce the observed trend in winter|likely because the winter ice extent is highly sensitive to the ocean heat flux convergence beneath the ice (Bitz et al. 2005), for which the model uses annually repeating forcing. That is, our simulation does not have ocean circulation changes, which act to enhance Arctic warming in winter (Bitz et al. 2006)."

The recent PDO shift is bringing much warmer water than normal up the coast of Alaska and into the Bering strait. And there's a surge of warm water that's been rapidly melting ice north of Svalbard.

Once there is sufficient heat built up in the Arctic ocean and first year ice has less than the critical thickness value to survive the summer heat:



Kevin O'Neill

Chris and Rob, you seem to be arguing past each other. Just break it down to basics:

NH snow cover typically melts from south to north. This indicates it's a result of the annual seasonal cycle; warming temperatures synchronous with growing insolation as winter transitions to summer. Global warming introduces a trend of less and less snow. This is the same phenomena we see with NH hemisphere lakes. Later onset of ice and earlier disappearance of ice. This has been changing for decades.

That last statement is important. Lake ice data shows this effect has been occurring since long before the relatively recent declines in SIA or SIE. Thus, the premise that declines in sea ice cover are the primary driver for reductions in NH snowcover is unconvincing.

Approaching the problem from the other direction, we know that temperatures over ice and snow are going to be colder than those over land or water free of ice or snow. While this is the basis of the albedo feedback, it typically has little effect on snowcover near the newly opened seas. The two primary reasons are the temperature of the newly opened seas (near or below the melting point of snow initially) and the minimal heat capacity of air. There is a tertiary reason: even when ice cover is removed we'd still need consistent or persistent offshore winds for any effect to materialize.

Having said this, reductions in sea ice cover could change synoptic weather patterns. This is the basis for 'Warm Arctic Cold Continents' which relies on a rapid advance of snow cover in the fall/winter. Unfortunately (for this discussion and for knowledge in general) the recent decline in sea ice cover has occurred in too short a time period for us to make many statistically significant evaluations.

We *know* that reduced sea ice cover has to affect the weather -- at least locally, if not regionally and globally. What we don't know is exactly how this will manifest itself and what the range of variation is or can be.

My own surmise is that by the time we do have statistical significance it will be a post-mortem discussion.

Kevin O'Neill

In the preceding comment the sentence "Global warming introduces a trend of less and less snow." doesn't really convey what was intended.

It should read something more akin to: Global warming introduces a trend in maximum snow cover and a trend in it's disappearance.

I can even argue with that characterization, but at least it's closer.

Chris Reynolds

Fish (D),

As long as there is ice survival in the summer mechanical deformation will continue to build up thickness of the northern coast of the CAA.

The trend in decline of the Atlantic Ocean dominated losses (e.g. Barents) is far stronger than any of the seas within the Arctic Ocean Basin. This suggests a decoupling. The primary issue here is probably the depth of AW within the Arctic Ocean basin, and the maintenance of the halocline due to winter ice growth.

Winter ice growth will continue to be strong as long as the winter remains cold. This winter was a blip due to an ENSO.

Bitz 2006 associates Rapid Ice Loss Events (RILEs) with ocean influx of heat. Notably, so far we have not seen any period since 1979 meet the criteria set in that paper for a RILE.

Across much of the pack (peripheral seas), April thickness is already such that it is below or equal to summer melt. Hence those seas are virtually seasonal. However the ice edge progression to the central Arctic means losses there only start in July, and even in 2012, the strong declines in extent in that region only started in early August, at a minimum of around 2.5M km^2 this is 1.5M above a daily min of 1M and further from the monthly average for September of 1M km^2 (virtually ice free). Taking the 2012 loss rate between 1/8/12 and 1/9/12 (linear R2=0.99), one would have to start the extent loss at that rate on 13 June to get to 1 million daily by 1 September.

Since 2010 Arctic Ocean April volume has not dropped below 19k km^3.

Chris Reynolds


Lake ice, snow cover, Arctic sea ice, are all being driven by AGW, with feedbacks driving sub trends within each metric, and feedbacks between the metrics. I cannot extricate those feedbacks without a model. However I have presented evidence from a PIOMAS based study that finds sensible air temperature is not a driver for the recent loss of volume - it's an ice/ocean (and ice albedo) albedo feedback. And I have presented a study that supports modelled temperature increases over land adjacent to ocean being higher due to ocean warming, even in April and May. What I am trying to get over is that the suggestion that snow is driving ice loss, whilst currently fashionable, is far from a conclusive case.

Note that for all potential explanatory variables, the greatest losses in 2010 and 2012 are clear outliers, not explained by the explanatory variables. Using the residuals from the linear fit of ice thickness and may volume loss, a linear fit of temperature to the residuals has an R2 of 0.11, for snow that is 0.03.

With an increase in variability that may be the case post 2007 (at present the effect seems largely to be due to 2007 and 2012), it may take longer than the rest of this decade to establish any inflection post 2007. Post-mortem? Perhaps, but after Eisenmann and Wagner pulled the rug from the fast crashers I struggle to see the evidenced argument for such an expectation. FWIW I am not seeing such an indication in the underlying data.

Ultimately there is probably little between the expectations of Rob and myself. Rob seems to expect a virtually ice free state within a few years (by 2020?). I used to think that but have reconsidered and now put it sometime next decade, possibly later (there is substantial uncertainty).

But that isn't the main issue: What made me so angry this morning was what I see as a dismissal, and evasion of, the points myself and DCS have put forward.

Kevin O'Neill

Chris, I agree that there were some rather cavalier attitudes expressed. You and I discussed last year (or perhaps 2014) on your blog the possibility that we've reached a 'floor' in PIOMAS volume. That since the the negative feedbacks increase as Sept. minimum SIE decreases, we may reach (may have reached?) a point where the negative feedbacks equal the positive feedbacks and the result slows and even the linear trend is no longer appropriate -- much less non-linear trends that accelerate the volume loss.

Of course the difference between yours, mine, and Rob's positions can probably be encapsulated on a chart with a total difference of a decade or so as far as the end result. I'm rooting for the quick result, but cognizant it's just as likely the longer one. No bets here :)

Rob Dekker

I'm sorry if I stepped on your toes, Chris. I don't want to disregard your arguments, but I do check the science behind them, and will bluntly report what I find. Especially when feels wrong (like snow losses caused by ice loss) or "to good to be true" (like a slowdown in ice melt.

For that latter argument, I dug through Armour et al, found no evidence for their argument, and was directed to Appendix A.

Now, this seem awkward, but I just read Appendix A of Armour et al, and it does not say anything about "decreased memory and decreased sensitivity to climate forcing". There is not even any discussion there about memory or climate forcing. Nothing.

So awkward as it is, I found no evidence in Armour et al for any 'slowdown' in summer ice melt rate when the Arctic moves from MY to mostly FY ice.

Rob Dekker

This is serious.
When a scientific paper makes claims that are not sustained by evidence, there is a BIG problem. And the trick of directing to an Appendix to find the answer that is not there I have seen being played only by Richard Lindzen.

So now that it is clear (at least to me) that Armour et al makes claims that are not sustained by evidence, the credibility of the scientists involved is on the line.

This paper has Cecilia Bitz name on it, so it this paper actually went to press, I'd seriously question her credibility. But luckily, it seems that Armour et al never made it into a scientific journal. Or did it ?

Colin Maycock

Rob, it was published. http://journals.ametsoc.org/doi/full/10.1175/2010JCLI3823.1

Rob Dekker

Thanks Colin. That's not good.
Luckily the unsustained statement that "if that occurs .... the Arctic sea ice system would move toward a regime of decreased memory and decreased sensitivity to climate forcing" is not part of their abstract, which instead states :

The September area is found to be nearly in equilibrium with climate forcing at all times, and therefore the observed decline in summer sea ice cover is a clear indication of a changing climate.
which IS sustained by evidence, and suggests that the trend in summer ice decline will continue with climate forcing, with no apparent slowdown.
Chris Reynolds


I'm sorry, you probably got caught up in my anger over the Orlando massacre.

That said, I think you're doing it once again in the comment immediately above. Playing games to avoid the issue.

That paper was presented at the autumn AGU meeting in September 2009. http://adsabs.harvard.edu/abs/2009AGUFM.C41C0478A
It is hosted on Cecilia Bitz's website. Blanchard-Wrigglesworth, Eisenman, and Gerard Roe all offered comments (on an earlier version).

Anyway, as I see it, it is suppoorted by Bitz & Roe.
It's just an exploration of an implication of Bitz and Roe. I use the Armour paper because the phrasing of the paragraph I quoted is so perfect.

Anyway, their finding that "...it is unlikely that a “tipping point” in September ice area and volume will be reached as the climate is further warmed." has since been backed up by Wagner & Eisenman.
That was published in a peer reviewed journal.

Anyway, I have just stumbled upon a more interesting subject from a paper over at Eisenman's site.

Rob Dekker

Chris, if you are so convinced that Armour et al has evidence that "the Arctic sea ice system would move toward a regime of decreased memory and decreased sensitivity to climate forcing", then why don't you point it out in their paper ?

Chris Reynolds

I have pointed it out - the whole paper. Really your request is like asking what symbol in a quadratic equation gives rise to its behaviour. Can you see why I see you as evasive - I point to papers to support my case and you throw back daft demands. To me the issue is quite clear, as it is in Bitz & Roe.

In essence the issue is that FYI grows back in a single winter. MYI takes years to form. Therefore MYI has a longer memory than FYI.

You can see this in the PIOMAS data, the loss of multi year ice (thickest) in April aligns with the decline in September extent.

Now why do you think that would be?


As the multiyear ice component drops to a minimal level the system becomes dominated by first year ice and the effective heat capacity of the system involves the ocean, not just the atmosphere + Ice + a an approximately constant heat flow from the Atlantic layer at 300m depth.

The 2015 paper Chris just cited shows how the effective increase in heat capacity as the ice cover thins could cause a slow down in the rate of loss of sea ice.

However, as the system becomes dominated by the Arctic ocean, what's happening in the ocean becomes important. The 2015 paper neglects the effects of possible changing ocean dynamics.

The warming of the Arctic and the advance of trees into areas formerly covered by tundra could increase evaporation rates and reduce run off. The northward flowing rivers in Russia could carry less fresh water into the Arctic ocean.

Moreover, there is strong evidence that the positive winter NAO pattern from 1988 - 2000 drove a large pulse of warm water into the Arctic. The thinning ice didn't just melt from warming air temperatures. It melted from below. Part of the Barents sea has transitioned from being dominated by Arctic water to Atlantic water.

However, the effects of changes in the ocean have considerable lag times. The effects of the strongly positive winter NAO in the '90s lasted well into the following decade. Note that a strongly positive NAO has reappeared in the last few years.


The Arctic is changing from an ice dominated system to a ocean dominated one so we must do a better job of tracking the changes in the ocean.

Rob Dekker

We are not even done talking about Armour et al, and why it does not provide evidence for a slowdown in ice retreat on a warming planet, when the next paper is mis-quoted :

The 2015 paper Chris just cited shows how the effective increase in heat capacity as the ice cover thins could cause a slow down in the rate of loss of sea ice.

It really does not say that, D.
It only finds that there is no "increased rate of retreat" as the ice disappears.

Rob Dekker

Chris said :

I have pointed it out - the whole paper.

I'm sorry Chris. That is the very definition of what you accuse me of (being "evasive").

Rob Dekker

Chris said

In essence the issue is that FYI grows back in a single winter. MYI takes years to form. Therefore MYI has a longer memory than FYI.

Armour et al defines that memory differently, and either way, it does not say anything about the response to climate forcing.
Which is the key to you hypothesis that there is a flatting in ice extent after 2007.

Look at it back to basics :
Since FYI melts out faster than MYI, we would expect sensitivity to climate forcing to INCREASE, not decrease as Armour suggests, as the Arctic transitions to an ice free state.

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.


Post a comment

Your Information

(Name is required. Email address will not be displayed with the comment.)