« PIOMAS February 2014 (upgrade to Version 2.1) | Main | Another ice extreme »


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


Here's the animation made by Wipneus showing how Svalbard became circumnavigable in the past two weeks:


Thanks Neven for posting this informative, yet very alarming article. Also thanks to Chris Reynolds for providing supporting evidence as to how these changes in the Arctic are already influencing global weather patterns.

If the recent decline in the Arctic's Albedo, due to Sea Ice loss, is already responsible for 25% of recent global temperature increases, what will happen when we reach the point of ice free summers in the Arctic??

I can appreciate the difficulty in developing models for predicting future changes in the global weather patterns. Most climate models use a 30 year average as a baseline. However, events "on the ground" are changing the baseline every few years? I shudder at the thought of what the Arctic Albedo will be after a summer that sees 2-3 months of an ice-free state.

Colorado Bob

It was also unusually mild in Greenland where the Summit site failed to register a temperature below -50�C (-58�F) for perhaps its first time for a January.

Iceland was exceptionally warm and for the first time on record some sites (two to be exact) never fell to freezing or below for the entire month: Vattarnes where the minimum was 0.4�C (32.7�F), and St�f�n Vestmannaeyjar where the minimum was 0.3�C (32.5�F).


Colorado Bob

This is no accident either , and I find it rather off the charts that wild land fires can burn in Norway in January.

Arctic Wildfires In Winter: Norway Experiences Freakish Historic Wildfires In January

Colorado Bob

Canada’s Arctic ice caps melting rapidly since 2005, according to documents-

David Burgess, research scientist and glaciologist with Natural Resources Canada, explained that since 2005 there has been a persistent high-pressure system over Greenland, which has acted to draw in more warm air from southerly latitudes and contributed to a warming High Arctic.

This same kind of system didn’t develop in 2013, which may explain the cooler temperatures last summer, he said.

Federal data show the Devon ice cap’s northwest sector has lost 1.6 per cent of its mass since the 1960s, the Meighen approximately 11 per cent of its mass and the Melville about 13 per cent, he said. However, approximately 30 to 40 per cent of the ice mass lost has happened since 2005.

“Since 2005 it has enhanced quite significantly,” Burgess said in an interview.



I think there is a significant negative feedback which should also be acknowledged and investigated; increasing autumn winter and spring snowcover in the northern hemisphere should be significantly decreasing the NH albedo.

At this time of year, in the polar night, Arctic albedo is entirely irrelevent.

Atlanta, Georgia, albedo is relevent.

The other researcher,besides Dr Francis, whose work looks very smart currently is Dr Cohen. As the Arctic warms, the continents cool. See the WACCY Science thread on the forum:


As continental snow cover increases, then local albedo decreases, with 2 significant effects that I can see:

1. It should cause further local cooling, due to albedo, and contribute further to the erosion of the temperature differential between the Arctic and lower latitudes - so the pattern is self-reinforcing;

2. It should decrease the overall accumulation of energy within the atmosphere, as more energy is reflected directly back into space. A significant negative feedback on global temperature rise...


OTOH, in the polar night, the effect of Arctic albedo is precisely zero.

Whereas,increased continental snowcover, due to WACCY weather causes decreased overall NH albedo.

This should be self-reinforcing as local snowcover causes local cooling;

also a significant negative feedback on global temps.



Whereas,increased continental snowcover, due to WACCY weather causes decreased overall NH albedo.


Whereas,increased continental snowcover, due to WACCY weather causes INcreased overall NH albedo.

It gets more white and reflects more light.

Christoffer Ladstein

Great summary of the soon to end winter of 2013/2014, Neven! With the situation going on at Svalbard, a very fragile ice at eh Eastern coast of Greenland, a Naresstrait waking up early and such a weak seaice in the Beringstrait area, this melting season already is bound to be an exciting one!

Colorado Bob: Wildfires midwinter at the coastal part of Norway is SO rare! At the same time, further east and southward, much of the inland areas, the most popular cabinareas, are "drowning" in 3 meter of snow, people have "lost" their precious cabins... so we truly live in a world with contrasts. And most scary is they're likely to strenghten in the years ahead, making normal planning impossible maybe?!

Ac A

Hi Folks,

here are couple of graphs regarding the PNAS paper:



Ac A

Neven, and here is the full link for the paper:




Idunno. I recall that last time that issue came up for discussion, the conclusion was that the effects created by lack of snow cover in May/June by far outweigh the effect of increased snow cover during winter months. Both because the sheer extent of spring deficits have been larger and because the sun is higher in the sky. And I strongly doubt that local cooling from winter snow will ever be sufficient to retard increased spring melt due to AWG, as we have seen no clear signs of this so far.

I Ballantinegray1

I find the figures more striking when I remember that snow cover didn't begin to crash until the late 90's and sea ice didn't crash in extent/area until 07'? this represents only half the study period for snow and far less for open water Arctic ocean. What does occur when we have the full 30yr period for an ice free summer basin? What 'proportion' of CO2 warming would we be looking at then? 75%?, 100%? more?
Also ,unlike CO2 the impact from Albedo crash is near instant so changes to albedo lead to near instant responses in our climate system. Are we setting up for a sudden leap in global temp, as the ice fades, comparable to all the warming we currently experienced but over a couple of years?


So far this year it seems to have been on average about 7 degrees C warmer than normal in the Arctic:

Ice extent is at about the record minimum for this time of the year. And spring is in the air. At least here in Holland it is, where so far this winter we've had no full day below zero degrees C, for the first time ever. Average Holland temperature this winter (extended fall) has been second highest so far (only 2007 had a higher average).

Let's see how fast the Northern hemisphere snow cover will melt this spring.

Andy Lee Robinson

Winter didn't really happen here in Budapest either.

I saw a snowflake once this year, or perhaps it was the snowflake!

Seriously, only had about a dusting of 3cm once that lasted a couple of days. Now spring is thinking about springing and then the transition to summer can be pretty short.


Is certainly true that decreasing arctic ice cover has no impact on the *insolation* dynamics of albedo where the sun don't shine.

But, the thermodynamic impact of changes to *insulation* has to be significant when increasing amounts of open water at or around 0C are exposed to colder air, without cover of ice and snow. And logically more in winter, when the differential is greater?

Imagining the thermodynamic interplay of decreased ice cover in dark arctic waters makes me want a nap.

Albedo is relatively one-dimensional, as reflected radiation of incident light; how do intrinsic energy radiation and convection, latent-heat-of-fusion, etc, complicate the story? How significant is this year's much warmer arctic air temps in this context?


Just considering some comments about areas being above normal temperature. Just questioning if everyone is looking at temperature change the same way. I recently with some numbers from some wildfire weather stations in Canada. First we looked at 40 years of data for average daytime high temperatures from a selection of weather stations. We could not see any change in daytime high. Then we did the same calculation but looking at the overnight low temperatures, and we found that the average of overnight low seems to be trending 3 to 4 degrees warmer than it was 40 years ago, even though daytime temperature did not seem to change. This has an effect on fuel drying rates, cured grass, overnight condensation, number of consecutive frost free days, length of fire season and growing season, lots of things us wildfire guys care about.
So my question – when we read comments that a certain area of the arctic is 4 or 5 degrees warmer than normal, are they looking only at daytime highs being warmer, or are they also considering overnight lows, or hourly temperature? Is there consistency in how this is looked at? Maybe the overnight low is 6 or 7 degrees warmer than normal. Need to look at both ends of that diurnal fluctuation.

Chris Reynolds


Good point.

It depends on what one is looking at. Using systems like NCEP/NCAR I must confess, I've tacitly assumed they're giving daily average, and working anomlay out from that - but I've not digged deep enough to confirm it. The DMI data uses another reanalysis system like NCEP/NCAR.

When using actual temperature data whether I look at Tmax Tmin or DTR is governed by what I am studying. For example the cold winter of 2009/10 in the UK affected Tmax more than Tmin in the Central England Temperature series (IIRC).

Colardo Bob,

You've reminded me of a comment I made about Greenland mass balance loss in 2013, a comment from 2 Jul 2013. "Unless a geopotential height ridge forms over Greenland very soon, we will not see a record melt on the Greenland ice sheet..."

Since then 2013 has not seen record melt on the Greenland ice sheet. I raise this merely to comment that the power of a 'theory' is in its predictive power. What I have isn't a proper theory, merely an amateur's idea. But using the observation of the Greenland ridging summer pattern I was able to call it right on Greenland in 2013.


@1rover1, if you are talking about current temps, I think the answer is they mostly all "overnight" temps currently ;)

But otherwise, yes, a good question. For a lot of those measurements, it does seem to be tied to a daily average of min and max, which implies assessment of anomalies would factor in and be relative to the higher minimums you saw.


In case anyone is interested:

You recently registered for a Sea Ice Prediction Network (SIPN) webinar to be held Thursday, 20 February at 11:00AM Alaska Standard Time (UTC/GMT -9 hours). Our guest speakers are Julienne Stroeve from the National Snow and Ice Data Center, and Cecilia Bitz and Ed Blanchard-Wrigglesworth both from the University of Washington. You can read more about the webinar on the SIPN webpage here: http://www.arcus.org/sipn. If you are not in Alaska, please check a world clock (http://www.timeanddate.com/worldclock/) to determine what time the event will be hosted in your location. Attached to this email is a PDF, which contains instructions on how to join this webinar. **Please give yourself at least one hour to download the necessary software to run this webinar.** You can also find instructions on the SIPN website (http://www.arcus.org/sipn).
Colorado Bob

A very good animation -
Watch the ice in the arctic disappear over the last 25 years


Colorado Bob

Chris Reynolds -
Yep, Greenland and what happens next are linked together. In this post Arctic Ice world.

The Sixth Extinction: A Conversation With Elizabeth Kolbert

Humanity's "most enduring legacy" will be our effect on the rest of life on Earth.

"We are the asteroid now."

Rick Aster

Thanks for the Svalbard animation, Wipneus and Neven. It is quite stunning to see.

Bob Bingham

Jennifer Francis puts up a very convincing argument concerning the lack of ice affecting the jet stream and I am convinced she is right. However this science was not on the horizon five years ago and it makes me wonder what else could be about to happen that we not know about.
Even after this dreadful period,when we were hoping that the jet stream would release its hold and move on it would appear that it is returning. Could it become a semi permanent feature? It would be devastating if it was.


@Bob Bingham - it is changes like this, extremes like this which are exactly what we need to start preparing for. Storm, flood and drought are going to redraw borders, and define the conflicts of this century.

Loss of the ice will totally re-order rainfall and climate across the northern hemisphere, even if it were to stop *here*.

John Christensen

Sorry, but I will need to completely disagree: I do not see any data supporting Dr. Jennifer Francis' claim that jet stream swings are increasing and causing more extreme weather events.

I had a look at NOAA data for January for the contiguous US states, and as you see January 2014 has been even extremely normal.. ;-)


Most extreme cold Januaries have been in this order: 1979, 1977, 1930, 1940 and 1918, where US January temperatures were 4-6C below the 1901-2000 average.

Is there an explanation how the extremes above can be categorized as 'less extreme' or more normal than the month of January 2014, which had 'only' regional extremes in the US, while ending up with average overall temperature?

John Christensen

From the same NOAA data you will notice that since 1990, there has not been a single year, where the US January negative temp anomaly (48 states) has exceeded 1C, and that the warmest January was in 2006.

Temperature anomalies are therefore moving to the positive side - consistent with a warming climate - but interannual temp swings or deviations from 1901-2000 average do not show any increase, but are reduced e.g. compared to the 1910-1940 period.


Morning John,

Have you been going over 'any data' or just those for the CONUS? I don't see how anything definitive can be said on Francis' work based on just the CONUS...

John Christensen

So does the warming climate (or changes in the jet stream) cause more or stronger extreme weather events like tropical cyclones globally?

No, it does not seem so:


The ACE (accumulated cyclone energy) from all tropical cyclone activity globally is on a downward trend, and while the long-term trend for major hurricanes is going up sligthly, you see that we have seen steadily less major tropical hurricanes since 2005.

So while a warmer, more humid climate will increase precipitation and cause a number of changes compared to prior decades, it seems the decline in temperature gradient between the arctic and tropic regions may have the effect to reduce tropical hurricane activity - or some other factor plays in here.

John Christensen

Good morning Werther,

I have not compiled a list of data to provide the counter-argument to Dr. Francis work, but I have not been able to find a data confirming that we experience an increase in weather extremes at our latitudes that should be caused by persistent changes in the jet stream.

Yes, my link was for NOAA data for CONUS, but UK Met also shows that the wettest January was in 1928, second wettest was in 2008, while this year was a close tie in third place with 1948:


Driest UK Januaries according to Met were in: 1997, 1963, 1964, 1987 and 1929 in that order, which must have been caused by blocking patterns as well.

January '14 average temperature for the UK came in on a 15th place since 1910 (same Met data), tied with 1938 and 1993, so the temperature did not deviate to any extreme extend.

Is there data available to support Dr. Francis' argument?


Science provides insight in climate change based on data. Projections on the future can be partially made on this insight. Partially, because we are witnesses of a global experiment that hasn’t happened before.
At least, not to our knowledge.

Part of the alarming interpretations are severe anomalies. They spread progressively over a larger surface of the globe. They tend to coincide more often. They get more amplitude.
These anomalies bring an urgent aspect to the known discomfort of climate change.
The known discomfort is what can be expected from a relatively long, smooth rise in temperature and ocean acidification. Humanity and part of the biosphere could, maybe, adapt to this rise in a manageable way.

That adaptation is bound to be much more difficult when this smooth rise is accompanied by accumulating disruptive anomalies. That is part of why the alarmed shout out ‘don’t take the risk…act now’.

John, you’re welcome here. But there’s a lot disputable on your opinions. If you are happy with a sense of less urgency, thats OK with me. On hurricanes you’re generalizing on supposed interrelations. On Francis’ work you’re, oppositely, specifying and diluting the case.

I invite you to figure out why a lot of other blog members do take all the signs very serious.
Or am I interpreting your posts wrong?


BTW this doesn't seem appropriate in a thread on arctic albedo.
I'll answer on hurricanes and the jet on a Forum-thread instead (if priorities allow me).

John Christensen


I follow this blog due to the focus on Arctic Sea Ice, and how this is being impacted by our warming climate, caused by increase in atmospheric CO2.

I have been trying to make sense of the arguments around the 'weather weirdness' and 'extremes', but cannot find scienfic reason or data supporting this, and will refrain from any further comment on that topic.


Neven, looking at the animation in your first comment, It looks like the franz joseph land islands could soon be surrounded by open waters if the current pattern holds, ant the sun is rising after four months of darkness, amazing and scary.


John, I repeat the quote from the second BBC article I used for this blog post:

With the UK, the US and Australia experiencing prolonged, extreme weather, the question has been raised as to whether recent patterns are due to simple natural variations or the result of manmade climate change? According to Prof Francis, it is too soon to tell.

"The Arctic has been warming rapidly only for the past 15 years," she says.

"Our data to look at this effect is very short and so it is hard to get a very clear signal.

"But as we have more data I do think we will start to see the influence of climate change."

As you can see, Francis herself says that an increase in blocking patterns isn't visible in the data yet, which of course doesn't mean it's not there.

More interesting perhaps than saying that the data doesn't show anything (because the period is too short as of yet), is to try and explain how Arctic sea ice (and NH snow cover) loss cannot possibly influence atmospheric patterns.

How can it not have an influence?

Kevin O'Neill

JC:" I do not see any data supporting Dr. Jennifer Francis' claim that jet stream swings are increasing and causing more extreme weather events."

John, then you haven't looked very hard - starting with the papers that Dr Francis has published and online seminars and the many video presentations available on the web.

It should be remembered that we are looking at very recent changes; it's almost impossible for the data to be *significant* over such a short period of time - that does not mean the data isn't there - just that we'll need more of it to confirm what we see now is not a natural variation.

It would be far more fanciful to assume that the jet stream is *not* affectted by global temperature changes that are unevenly distributed.

Chris Reynolds


I'm sceptical that Francis/Vavrus 2012 has found the claimed increase of waviness in the jetstream, from my reading Barnes has raised substantial doubts about that research.

However the UK Met Office have just announced that with a week to go the UK has already had its wettest winter on record.
And their earlier report, whilst concluding that models would be needed to assign an AGW linkage, cite warm SST's off Indonesia, and in the mid Atlantic (increased SST -> AGW), and the observed increase of heavy rainfall and linkage to warming (again - AGW). Note that the heavy rainfall events do not necessarily impact monthly/seasonal averages. And as the Met Office state - the signature is only starting to come out of the range of natural variability.

I've been seeing the fingerprints of AGW in this winter, and it was a surprise to see someone as conservative as Prof Slingo come out and say what I had been thinking.

But moving to summer, here in the UK we have had a string of exceptionally wet summers since 2007 (the year of a major Arctic sea ice crash). My recent blog post is not out of the blue. I've been wrestling with summer rainfall and the anomalous atmospheric behavior since late 2011. I have been unable to reject the hypothesis that our (UK) weather is changing, probably due to sea ice loss, despite hacking away at the problem since then! I won't bore you with details you may not want - ask if you want me to expand and I will make th time.


RE the JC vs Werther exchange above, I just want to note my support for JC.

JC is raising questions based on data. Even Francis, who FWIW I think is probably correct, has said that her work is in need of further research for confirmation, and that the evidence from the data is at the limits of detectability.

Please all feel free to express opinions or highlight data which raise questions about any of our apparent shared assumptions.

One of my own current beliefs is that Cohen...


is at least as relevent as Francis.

Fire away! Prove me wrong! (It will come as no surprise, and it's why I come here - to benefit from the rest of y-all's education.)


WRT Albedo...

I still think that the prediction of increasingly cold continents as the Arctic warms deserves further consideration here, for two reasons...

1. It will be advanced, elsewhere, as aneddotal evidence that AGW is FOS;

2. If it causes an increase in NH snowcover, this does cause an increase in NH albedo, reducing temperature increase due to CO2. See Cohen et al, 2013.

"We suggest that the simulated widespread warming may be due to incorrect sea ice-atmosphere coupling, including an incorrect triggering of positive feedback between low sea ice and atmospheric convection, resulting in significant model errors that are evident in seasonal predictions and that potentially impact future climate change projections."

If this science is correct, better that it is discussed honestly here first...

Chris Reynolds


Cohen is bang on - he's been chasing the winter/Arctic linkage down for over a decade. And he's shown he's got it right.

I've only just read the Pistone paper on albedo today (hat tip Ac A). The Cohen paper is for tomorrow.


The met office have declared that winter 2013/14 is the wettest in the UK since records began.
But there's still a week of February to go...

That makes the top ten records for UK winter (DJF) rain look like this:

2014 486.8
1995 485.1
1990 470.9
1915 463.1
2007 437.9
1916 433.0
1937 423.8
1994 422.9
2000 417.7
1920 402.3

It's worth pointing out that 6/10 of the most extreme wet winters come from roughly the last quarter of the available data (since 1990). By monte-carlo (not got my head around hypergeometric formula yet), I calculate the probability of that happening by random chance as 0.0085, or nearly 1%. There are three years from the first quarter (~25yrs) of the time series (1911 to 2014) the probability of that being by random chance is 0.25, i.e. not surprising.

As I've not had chance to dig into this further, unlike the summer pattern, I don't find the bare probabilities as persuasive as perhaps I should.


idunno, I'm not good with numbers, but:

1) I think that summer sea ice and snow cover decline overwhelms any negative feedback in winter, be it wrt influence on temp or albedo.
2) The winter negative feedback will not last forever.

And wrt "increasingly cold continents": Is this happening? Or aren't we used to cold snaps anymore?

Allen W. McDonnell

While the winter snow cover would be an increase in albedo during sunlight during the dark winter it acts as a nice fluffy insulation blanket that keeps the cold from penetrating the ground as easily. That should mean that once the sun come over the horizon and melts that fluffy snow the ground and ice under the snow will already be closer to the melting point than it would have been if the ground were bare.
Here in Ohio we have been in cycling states switching back and forth between the warm rain we have today with +7 C temps and the two to four week periods of -15 C or colder. We have had huge piles of snow piled up and then they all melted in December and again in January. Today they are melting once again and the snow season here usually lasts at least another month. The ten day forecast says we will be getting more snow in about 5 days with temps from -5 C to - 20C. Our normal for this time of year highs of +2 C with cold nights allowing the snow pack to gradually fade. Instead we are under our third flood warning of this winter.


@iduuno - I suspect "increasingly cold continents" is an illusion. I'll have to look at data, bi I suspect it looks that way simply because they are not warming as fast as the Arctic Ocean.


My former post about using the hypergeometric distribution was an example using the problem at hand. Here is a more general description of how it's used for similar problems:

A = [ (# of years in recent period)! ] / [ (# of years in recent period that are top 10)!(# of years in recent period that are not top 10)! ]

B = [ (# of years in older period)! ] / [ (# of years in older period that are top 10)!(# of years in older period that are not top 10)! ]

C = [ (# of years in entire period)! ] / [ (# of years in entire period that are top 10)!(# of years in older period that are not top 10)! ]

Probability = A * B / C

Of course "!" means "factorial". For those who don't know about factorial,

0! = 1
1! = 1
2! = 2 * 1 = 2
3! = 3 * 2 * 1 = 6
4! = 4 * 3 * 2 * 1 = 24
5! = 5 * 4 * 3 * 2 * 1 = 120

In the current problem, the number of years in the recent period (1990 - 2014) is 25, the number of those that are top ten is 6, and the number of those that are not top ten is 19. So A = 25! / (6! 19!) = 177100

In the current problem, the number of years in the older period (1910 - 1989) is 80, the number of those that are top ten is 4, and the number of those that are not top ten is 76. So B = 80! / (4! 76!) = 1581580

In the current problem, the number of years in the entire period (1910 - 2014) is 105, the number of those that are top ten is 10, and the number of those that are not top ten is 95. So C = 105! / (10! 95!) = 28848458598960

Probability = A * B / C = 177100 * 1581580 / 28848458598960 = about 0.00970928

Similar to the earlier problem, I would sum the probability of 6 of the top ten being after 1989, 7 of the top ten being after 1989, 8 of the top ten being after 1989, 9 of the top ten being after 1989, and 10 of the top ten being being after 1989, to get the probability of 6 or more of the top ten being in the period 1990 - 2014.

For 7 of the top ten being in 1990 - 2014, A = 25! / (7! 18!), B = 80! / (3! 77!), and C is the same as before, C = 105! / (10! 95!).

For 8 of the top ten being in 1990 - 2014, A = 25! / (8! 17!), B = 80! / (2! 78!), again C = 105! / (10! 95!).

For 9 of the top ten being in 1990 - 2014, A = 25! / (9! 16!), B = 80! / (1! 79!), again C = 105! / (10! 95!).

For 10 of the top ten being in 1990 - 2014, A = 25! / (10! 15!), B = 80! / (0! 80!), again C = 105! / (10! 95!).

Recall that 0! = 1.

Calculating all of those (from 6 to 10 of the top ten being in the period 1990 - 2014) and then adding them up, the total is about 0.0112026.

We could add the numerators of each first and then divide by the common denominator (the common denominator being C).

[ (25! / (6! 19!))(80! / (4! 76!)) + (25! / (7! 18!))(80! / (3! 77!)) + (25! / (8! 17!))(80! / (2! 78!)) + (25! / (9! 16!))(80! / (1! 79!)) + (25! / (10! 15!))(80! / (0! 80!)) ] / [ 105! / (10! 95!) ] = about 0.112026

If Chris has a way that works for him, he might decide to stick with it, but once one has the hang of this method it is quicker and more accurate.


Obviously I made a typing mistake: it should read "about 0.0112026" both times instead of "about 0.112026" the second time.


PS I used 1910 to 2014 for the whole period in the above calculation. For 1911 to 2014, I get 0.0101947 for the probability of exactly 6 top ten years in 1990 - 2014, and 0.0117851 for the probability of 6 or more top ten years in 1990 - 2014.


One more typing mistake, about three comments ago:

"C = [ (# of years in entire period)! ] / [ (# of years in entire period that are top 10)!(# of years in older period that are not top 10)! ]"

should read

"C = [ (# of years in entire period)! ] / [ (# of years in entire period that are top 10)!(# of years in entire period that are not top 10)! ]"


To state the obvious, this study reaffirms why the Artic is in many ways ground zero in the effort to understand how green house gases will alter our climate systems in the near and long term. If ice retreat continues where 2012 left off I think the next decade will see worsening and unexpected weather patterns in a stepwise transition to a new climate.

Chris Reynolds


Thanks for taking the time to help me, that's he most clear explanation I've seen on the 'net. Even though I'm tired (just one more day before the weekend - phew) I can follow it.

I will be using it when needed in future.


Hi all,

I have been investigating snowcover, without finding as much as I would like.

NSIDC latest update touches on it, here...


I get the vague impression that NH snowcover is declining in Spring and Summer, fairly level in Winter, and increasing in Autumn/Fall.

jdallen, I suspect that you are right that the continents are not warming in absolute terms, but only in relaton to the Arctic; though absolute cooling has been the pattern of the last 30 days, as per the anomaly maps overleaf on the ASIGraphs page;
a massive hot anomaly over the Arctic dividing massive cold anomalies over NA and Eurasia.

By the by, the Central Arctic is getting very close to a record low winter area:



idunno, there's one blog post on snow cover that Al Rodger did here as a guest blog in 2012.

You can find monthly anomalies here, and seasonal anomalies are at the bottom of the Long-term graphs page on ASIG.


Fantastic, Neven.

The huge Summer and Spring negative anomalies do seem to overwhelm the slight growth of Fall and Winter; but Summer has fallen off the longterm graphs page.

Thanks again.

Shared Humanity

In support of Allen W. McDonnell remarks regarding the impact of snow cover, we've had a very cold winter in Chicago although I would like to be clear this cold winter is not record cold. In fact, these kinds of winters occurred frequently in the 70's and 80's.

What was very welcome was we have also had a great deal of snow. This is welcome when compared to the relatively dry winters we have had recently. This heavy snow began falling prior to the cold air masses moving in and, despite having 12" of snow still on the ground, the ground itself is completely thawed. The insulative blanket of snow has prevented any freeze and the time for freezing is passed.

Fall positive anomalies can only have the effect of preventing ground from freezing and warm ground is probably contributing, in part, to rapid spring melt.

It should surprise no one that fall snow anomalies are climbing. The lower ice cover in the Arctic at the end of the melt season, exposes open water to evaporation which, with the early cold fronts pushing into Canada and Siberia, can only mean more snow as the atmosphere has more moisture. I expect these positive fall snow anomalies to persist and likely grow as we lose more Arctic Sea ice. This is not a good thing.


Chris (and others):

Given that you are an engineer and that you do programming I'm aware that you could write a program for implementing a hypergeometric probability calculation for the problem at hand. Nevertheless, I just wrote my own version of pseudo-code for this that you can use in case you are busy or have other priorities. I hope that you aren't insulted because I did something that you could have done for yourself, but I didn't mind writing this, and maybe it's easier for someone who's already familiar with the hypergeometric distribution. Maybe you don't expect to have a use for this type of calculation in the future, in which case you cam ignore this. Regardless, I think that it's likely that some other reader might want to use it.

Anyone can feel free to point out and / or correct a bug.

Comments are in parentheses.

There are cases where two "if" statements could be replaced by an "if" statement with an "else".

Note that the number 10 only appears explicitly once in the pseudo-code (and I use it again in remarks after the code in Min(10, (la - fi + 1))), so it could be readily changed to another number in case one is interested in the top 12 or whatever.


be = beginning year of record (in this case 1911)

en = ending year of record (in this case 2014)

fi = first year in which we are counting how many top ten years there are (in this case 1990)

la = last year in which we are counting how many top ten years there are (in this case 2014)

ho = how many of the top ten years we want to be in the years in which we are counting the top ten (in this case 6)


{cnum is the numerator of C in earlier comment, cden1 is one denominator, cden2 is the other denominator, cbig will be the larger of (cden1,cden2), csma will be the smaller of (cden1,cden2), if they are equal then it doesn't matter}

cnum = en - be + 1

cden1 = 10

cden2 = cnum - cden1

{One could do a check here to make sure that none of cnum, cden1, cden2 are negative. One of them being negative would imply a problem with the input values, such as trying to have 6 top ten years within a period of only 4 years.}

if cden2 >= cden1 then cbig = cden2, csma = cden1

if cden1 > cden2 then cbig = cden1, csma = cden2

{Instead of doing the full factorial on the numerator, we'll cancel that factorial with the larger factorial on the bottom, to keep the number from getting as big, for example 7! / (4! 3!) = (7!/4!) / 3! = (7 X 6 X 5) / (3 X 2 X 1). Then we'll do something a little tricky: divide by the denominator numbers (smallest to largest) as we are multiplying the numerator numbers as follows: ((((7 / 1) X 6) / 2) X 5) / 3, being careful to multiply by the next numerator number before dividing by the next denominator number, this way we always have a whole number as we go (this can be proven mathematically). If csma = 0 (in the case at hand this would be the case if our period of record is only ten years), then the result of cnum! / (csma! cbig!) is 1.}

cpro = 1

if csma > 0 then for kc = 1 to csma step 1, cpro = (cpro * (cnum + 1 - kc)) / kc

{The outer brackets of (cpro * (cnum + 1 - kc)) are to emphasize that we want to do the multiplication before doing the division to ensure that we always have a whole number for cpro}

{cpro is our value for C}

{Now we work out A & B in a similar way}

anum = la - fi + 1

aden1 = ho

aden2 = anum - aden1

{One could do a check here to make sure that none of anum, aden1, aden2 are negative.}

if aden2 >= aden1 then abig = aden2, asma = aden1

if aden1 > aden2 then abig = aden1, asma = aden2

apro = 1

if asma > 0 then for ka = 1 to asma step 1, apro = (apro * (anum + 1 - ka)) / ka

bnum = cnum - anum

bden1 = cden1 - ho

bden2 = bnum - bden1

{One could do a check here to make sure that none of bnum, bden1, bden2 are negative.}

if bden2 >= bden1 then bbig = bden2, bsma = bden1

if bden1 > bden2 then bbig = bden1, bsma = bden2

bpro = 1

if bsma > 0 then for kb = 1 to bsma step 1, bpro = (bpro * (bnum + 1 - kb)) / kb

{so far all of our variables have been whole numbers}

prob = apro bpro / cpro


To calculate the probability of 6 or more top ten years in the interval (instead of exactly 6 top ten years in the interval), we would sum this for ho = 6 to Min(10, (la - fi + 1)). So instead of inputting the value of ho at the beginning, one could input minho = 6, and then ho runs from minho to maxho = Min(10, (la - fi + 1)), In this case the calculation of cpro could be moved outside the "ho" loop, as it wouldn't change, and then the division by cpro could be done at the end, after summing the products of the apro's and the bpro's. Better still, the calculation of cpro could be done before the "ho" loop, then the program could print out the values of apro * bpro / cpro for the various values of ho as it does the loop each time, and then still dividing the sum of the (apro * bpro) values by cpro at the end. Alternatively, to calculate the probability of 6 or more top ten years in the interval (instead of exactly 6 top ten years in the interval), one could run the program for each value of ho and then add the results manually later, with the possibility of more rounding error.

Chris Reynolds


Thanks for all your work explaining this. I did (I'm sure) post an appreciative reply this morning, but this blog eats up my comments into the spam filter, which is why I rarely comment here.

I've a weekend to myself, and have managed to finish work at a reasonable time today, so will apply myself tomorrow.

"...in case you are busy or have other priorities."

Yep, I'm behind on my reading again (got over 60 papers I've not yet properly read). I normally don't need to look at the issues the UK rain has raised because in the Arctic data trends are so strong that ranking approaches aren't normally needed.

But I will be keeping an eye on the rankings if summer 2014 in the UK is exceptionally wet, as 2007 to 2012 suggest is likely (though not certain).


A 'PS' to my post above about UK winter rainfall - during the 1990s there was a strongly positive Arctic Oscillation, this would be expected to lead to low pressure dominated weather over the UK in winter. This is a major reason I am not placing as much weight on the "6/10 of the most extreme wet winters come from roughly the last quarter of the available data (since 1990)". It doesn't rule out a role for increased rain due to GW, but it does complicate the matter and mean that a causal inference is not necessarily correct.



You have several times stated your conviction, that recent wet summers in the UK are due to loss of Arctic sea ice. I was just wondering, whether this paper is on your “to do” list, when it comes to reading during the weekend:

http://www.ametsoc.org/2012extremeeventsclimate.pdf see Tett et al. p. 532-535 in particular.

Simon is a distinguished professor at the university of Edinburgh, and a former top scientist at the UK Hadley Centre. He has so far not been able to reproduce a connection between wet UK summers this century and dwindling sea ice in the Arctic. Apparently, either his model experiment is flawed or your assumptions are wrong. I would appreciate, if you could comment on this discrepancy at your convenience.

Hubert Bułgajewski


Again "below the line".

Colorado Bob

“The amount of ice being lost from Pine Island glacier is equivalent to every person on our planet pouring 10 pints of water (4.7L) into the ocean every day,” Professor Andrew Shepherd, an expert at the University of Leeds who was not involved in the study, told Reuters. “That’s the last thing our flood defences need right now.”

Colorado Bob

The PIG -

10 pints of water (4.7L) into the ocean every day, times 7 billion .

Now that's a number. Here's hoping Prof Shepherd comes up with the same analogy from the Greenland ice sheet.


Jennifer Francis and Kevin Trenberth on Inquiring Minds today:



Thanx Tenney,

Noted intrigung reference by Francis @ 24.30 and following to new research, as yet unpublshed, showing post-2007 persistent high pressure ridges over openwater areas of the Arctic.

Also interested that, in discussing Cohen, Francis notes that the Eurasian side has melted much more than the NA side. Kinda sobleedingobvious that it's somethng I've tended to overlook/take for granted.

George Phillies

It is perhaps noteworthy that the Bremen map is showing incomplete ice cover almost up to the pole, in late February. That strikes me as being a bit radical relative to years past, but perhaps I misremember.

Chris Reynolds

P Maker,

You just have to follow my reasoning, which doesn't merely apply to 2012, but to the years post 2007. I've gone over this matter in detail in my blog posts, perhaps the best summary is here.

However, let's look at the evidence in that summary of research.

"The Atlantic/
European summer of 2012 should not be viewed in isolation but as the latest
in a succession of very similar summers...

...Overall, the model results suggest that the atmospheric circulation over the North Atlantic and European region in summer 2012 (negative phase of the SNAO), which was largely responsible for the observed extreme anomalies in European precipitation, was influenced by global SST and sea ice extent anomalies, and that it is likely that SST anomalies in the North Atlantic played a particularly important role"


"...the level of internal variability in the model simulations is insufficient
to account, with significant likelihood, for the magnitude of the observed anomalies (not shown)."

They conclude that:

"The degree of decadal variability in these features suggests a role for forcing from outside the dynamical atmosphere, and preliminary numerical experiments suggest that the global SST and low Arctic sea ice extent anomalies are likely to have played a role and that warm North Atlantic SSTs were a particular contributing factor."


"Among recent summers, only 2007 is outside
of the 1961–90 climatology (Fig. 11.1). This illustrates the difficulty in analyzing individual events when there is a great deal of internal variability and so we
focus on longer timescales. On five-year timescales, the recent run of wet summers is extremely unusual with observed values at about 120% of normal, corresponding to a deviation of more than 2.5 standard

So the precipitation during summer 2007 to 2012 is unusual. What of their model results?

"For each of the factual, counterfactual, and counterfactual-NH ensembles the average NW Europe precipitation for 2007–12 is close to 100% (Fig.11.1). These are all inconsistent with observations.
The maximum percentage of normal precipitation from any single simulation in the three ensembles is 112%. This suggests that the recent changes cannot be explained by sea ice changes or internal variability, as simulated by HadAM3P."

Yes, as simulated by HAD AM3P sea ice seems to have no role, but if you check out the first link in this reply, Screen 2013 (using coupled HADGEM and ACCESS - CMIP5 model) finds that the simulated and observed 300mb meridional wind patterns for the difference between wettest and dryest summers, and the low ice and high ice simulations are virtually identical, as shown in this graphic.

So the earlier model HAD AM3P is unable to produce the link that a later


This study also uses HAD AM3P with data from the ClimatePrediction.net experiment.

They conclude:
"Although the summer UK rainfall in 2012 was unusually large, the model distributions studied suggest that any anthropogenic influence on these patterns was minimal."

However the 'all forcings' run shows (fig 12.1) substantially less rainfall (~300mm) than the observations (ECWMF)over the UK. In respect of this they note:

"This implies one of two possibilities: either (1) our findings are correct for the prescribed atmospheric CO2 concentrations, SST, and sea ice fractions appropriate to 2012, then the chances of the observed summer UK rainfall patterns occurring are in fact incredibly small, or (2) there are systematic biases in our atmospheric modeling structure."

They note that problems with this model have been noted in paper number 10 of that collection.


Chris Reynolds

Part 2.


The authors use analogues of atmospheric flow with previous years. They select the 20 days in the past that are most similar to each day in 2012, limiting the study to the summer by using a window of 60 days tied to the day studied. They study the North Atlantic Ocean, and bordering land masses.

They find that "...spatial patterns of analogue precipitation follow the observed ones, albeit with lower amplitudes."

They conclude:

"The trend in summer precipitation over northern Europe is significantly positive but is marginally reproduced by analogues of circulation (with no statistical significance)...

Hence, we cannot attribute this upward summer precipitation trend since 1971 to changes in the atmospheric patterns themselves, because the frequency of cyclonic patterns has not significantly increased nor have the analogue reconstructions of precipitation. This suggests a contribution of climate change to precipitation rate in northern Europe. We conjecture that such a trend could be due to precipitation rates within the cyclonic patterns, which convey more moisture because of increased temperatures."

So this is more in line with Trenberth's concentration on the effects of increased atmospheric humidity.

But this does not address the sea ice impact on Greenland ridging and the post 2007 summer pattern.


Chris Reynolds

Part 3.

These papers are part of a set chosen to examine various weather extremes over 2012. At the end of the document is a summary.

The summary finds that:

"...there is a danger in drawing too strong
a conclusion from a small sample of 19 analyses of 12 events that were not chosen at random. That said, approximately half of the analyses found some evidence that anthropogenic climate change was a contributing factor to the extreme event examined, though the effects of natural fluctuations of weather and climate on the evolution of many of the extreme events played key roles as well."


"the attribution conclusions for such relatively complex events remain somewhat

...Other studies, using different and perhaps improved tools (e.g. new models) will undoubtedly come forth and will further test the efficacy of surmised and plausible causal factors....

...With the increasing sophistication of event attribution studies comes a greater focus on assessing the capability of the various current approaches to provide robust answers to such difficult questions as whether there is a link between anthropogenic climate change and the extremely wet summer in northern Europe in 2012, which was not only an extreme event in itself but also the sixth summer in a row in which UK rainfall was higher than the 1981–2010 average."

It seems to me that none of these papers that addressed European/UK rainfall in the summer of 2012 support the idea that Arctic sea ice has played a role. However as I have argued in this long (sorry!) series of responses to your question - I am not convinced that any of these studies refute the linkage with Arctic sea ice and the wider issue of the 2007 to 2012 wet summers in the UK.

The study concludes:

"As the science underpinning the attribution of extreme events matures, it will nurture and make possible the creation of operational climate attribution
systems (Stott et al. 2013). Then, as attribution of extreme events become part of routine climate services and part of the GFCS, scientists will need to continue outreach into our stakeholder communities. Climate risk reduction is a key driver of many climate adaptation activities (van Aalst et al. 2008). Therefore, many communities will need to better understand exactly what attribution of extreme
event science can and cannot say. To return to the opening analogy, this means answering the question of how the change in the driver’s speed was responsible for changing the odds of colliding with a texting driver on a wet road, which would be the extreme event we are trying to attribute."

In that sense I should point out that from the studies I have read I am not convinced that outside the Arctic itself any part of the globe is showing such strong anthropogenic impacts as the UK itself (mediated by the loss of sea ice), and then only in summer!

I had cause to read the Greene paper on Hurricane Sandy last night.
Yet again I'm left interested but not convinced of a climate change impact. Their conclusion:

"Although a direct causal link has not been established between the atmospheric phenomena observed in late October 2012 and the record breaking sea-ice loss observed during the preceding summer months, all of the observations are consistent with such an interpretation."

Which I think is reasonable.

I just think that the evidence with respect to recent UK summer rainfall is far stronger. But there is a danger that having studied this since 2011 I am biassed, I try not to be but it's hard to see it in oneself. You'll have to make your own mind up on that, and on whether my responses to the points raised in the BAMS paper you presented hold water, or are just evasions.

Anyway, P Maker, thanks for showing me those papers.

Rather than just write a new blog post I may well re-post this string of responses on my blog. But I'd welcome people's critical comments on those papers and my response. Or indeed any other part of my reasoning.

If needed I can precis what I've blogged on about the UK Summers/Arctic issue to save people time reading my blog. What I've posted above is already massive enough.


Spam comments released, trilogy now complete.

Chris Reynolds

Ouch, there is a lot!

Sorry, P Maker, but you did ask...

Thanks, Neven.

Chris Reynolds

Spotted a major typo, and need for explanation.

In comment 1/3 I wrote:

"Yes, as simulated by HAD AM3P sea ice seems to have no role, but if you check out the first link in this reply, Screen 2013 (using coupled HADGEM and ACCESS - CMIP5 model) finds that the simulated and observed 300mb meridional wind patterns for the difference between wettest and dryest summers, and the low ice and high ice simulations are virtually identical, as shown in this graphic.

So the earlier model HAD AM3P is unable to produce the link that a later "

should end:

So the earlier model HAD AM3P is unable to produce the link that a later model is able to.

Looking at the meridional wind plot - this needs explaining. Zonal winds go around the planet, meridional go north and south. The plot I linked to shows activity at 300mb pressure, at this height in the atmosphere is found the jet stream. So the similarity of those plots indicates the meanders (north south) in the jetstream, and shows that the difference between observed wettest an dryest years is virtually identical to the modelled difference between high ice years and low ice years. Put simply the loss of sea ice in the model seems to be forcing the same pattern of jetstream meanders as are observed between wet and dry years.



Thanks for sharing your thoughts with us. As I read through your wise reflections, I also had a chance to listen to the podcast with Jennifer and Kevin sorting out the giant Jet Stream issue. Earlier this Boreal winter, the “stuck meander” over the North Pacific/North American continent did spur the discussions in another thread. More recently, the “stuck meander” over the UK has been the key issue.

I took the time to re-read my old textbook on the subject of meanders to look for an analogy, which may help to explain the phenomenon.

Meanders are – according to Gregory & Walling, 1973: 247-257 – a common feature of many natural systems. A number of factors influence the wave length of these features – including the amount of material transported in the stream. When the amount of material in the stream increases, the wavelengths get shorter and the meanders get wider.

A few examples serve to illustrate the huge natural variability of these phenomena:

1) meanders on glacier surfaces transport melted ice (H2O) (see http://onlinelibrary.wiley.com/doi/10.1002/jgrf.20135/abstract ),
2) meanders on the surface of karst landscapes transport dissolved CaCO3,
3) meanders in ocean currents transport dissolved sea salt (NaCl),
4) meanders in tidal streams transport flocculated clay minerals,
5) meanders in river valleys transport bottom load and suspended sediments
6) meanders in “atmospheric rivers” transport mainly evaporated H2O.

By analogy it is inferred that an increase in atmospheric water vapor content due to anthropogenic global warming should lead to shorter wavelengths, or - if you wish - larger meanders.

Hence the hypothesis put forward by Jennifer (http://thinkprogress.org/climate/2013/04/16/1867371/video-explains-how-loss-of-arctic-ice-weakens-jet-stream-amplifies-extreme-weather/) could be tested using a simple measure of the wave length of the Jet Stream meanders (e.g. duration of flow with sinuosity > 1.5). If the wavelength has been systematically shortened over the past few decades, this should not necessarily indicate, that the temperature difference between the Tropics and the Arctic has decreased. Alternatively, a link from a wavier Jet Stream to increased evaporation from tropical and subtropical oceans areas may be substantiated.

End of part 1.


Part 2

Again, according to my old textbook, meanders in rivers deposit material on the convex bar and erosion takes place on the concave bank. By analogy this would mean that the Jet Stream would deposit energy over the Pacific Hot Spot and “erode” Arctic air masses over the Bering Sea. Similarly over the Atlantic: High SSTs are being reinforced south of the UK and Arctic air is being “eroded” north of Scandinavia, as we see with the dwindling sea ice around Svalbard.

Apparently, the Polar Jet Stream above 250 hPa has been stronger this winter, whereas the Jet Stream below this level has been weaker. This weakening of the Jet Stream may be an extension of the general weakening of the observed Hadley Cell circulation since May/June last year. In the absence of a strong weather-forming Jet, I will again postulate, that basic climatology takes over. This includes the important influence of topographical barriers, such as the Rocky Mountains, which Jennifer also mentions.

Acknowledging that the Rockies is a solid topographical barrier to the lower parts of the Jet Stream, it has been fascinating to see the persistence of the ridge and through pattern over the North Pacific/American continent these past 18 months. Various Jet Stream “films” exposed on the ASI blog show, that the wavelength is shortened (waves are compressed) and the waves get “stuck” on the western side of the Rockies, whereas the wave patterns on the eastern side have been more variable over the course of the past year and a half.

As the atmosphere picks up moisture from the warmer Tropical and Subtropical oceans, this high “load” in combination with a weaker Jet Stream may have helped to maintain a “stuck and convoluted meander” west of the Rockies. The jet bends south over North America and helps to pull down Arctic air for months on end as we have seen. When the dry and cold air mass leaves the continent, the turbulent air picks up moisture again from the warm Gulf Stream and new storm systems intensify, as they cross the Atlantic and hammer into the British Isles.

Previously, I have alluded to the “lack of weather” under the “new normal”. Similarly, it may now be the case, that the Rockies have come to play a more dominant role because the traditional strong temperature gradient between the Tropics and the Arctic has diminished – eventually leading to a more permanently weaker Jet Stream. It is also quite likely, that this “stuck pattern” has now finally been demolished by the onset of early spring in the Northern Hemisphere. Now the warming has started in the Tropics and Subtropics in earnest, and the temperature difference relative to the Arctic has just about exceeded the threshold for a viable Hadley Cell. Come March/April and - if the temperature gradient again goes below the threshold - we are back to “the new normal”.

If Kevin and Jennifer could agree on a common indicator (such as duration of Jet Stream flow with sinuosity > 1.5) it would be extremely interesting to see whether this indicator shows an increasing trend over the past 15 years. On the other hand, if Kevin could “tweak” his model to include atmospheric water vapor as a driving force behind a more meandering Jet Stream, then the two of them might be able to resolve the issue, and we would have more reliable climate models.

Shared Humanity

Thanks for this post P-maker. I have always felt that we have underestimated the impact of northern hemisphere topography on climate and the manner in which the jet stream behaves.

We fully understand how mountain ranges affect microclimate with heavy precipitation on the windward side and generally drier weather on the leeward side but I think the Rockies, the northernmost mountain range in the hemisphere has always had a hemispheric wide impact on climate. Other large topographical features (Atlantic and Pacific Oceans, Asian land mass) surely have similar impacts.

75% of the world's tornadoes occur in the U.S. The Rocky Mountains interaction and influence on the jet stream are, I believe, the reason. The jet stream tends to meander north west of the Rockies and then heads south across central North America where it clashes with gulf moisture streaming north.

Your suggestions about how these topographical features might interact with a weakening jet stream in a warming world is fascinating. If we accept that a slowing jet stream will result in more stuck weather and you recognize the impact of northern hemisphere topography on the jet stream, my biggest concern is that these stuck weather patterns will begin to develop a stickiness or persistent pattern seasonally. It won't simply be that weather in a region persists for extended periods but that specific types of weather will begin to be persistent features for regions in the northern hemisphere. Perhaps Great Britain will need to get use to far more rainfall than ever recorded on a regular basis. Perhaps the southwest in the U.S. will be thrust into a permanent drought status as the weakened jet stream more regularly meanders north in the Pacific, carrying needed rains further north up the coast.

If the influence of topography on a weakened jet stream increases, we may be closer to severe climate impacts than we realize.

Shared Humanity

Here is research on the possible causes of a mega-drought period in the American Southwest which persisted for several centuries and wiped out native American cultures in the four corners region of the Southwest.


They research the Pacific SST and conclude that persistent La Ninas were a contributing factor. The research does not look directly at the jet stream but I have to believe that if we could study this, we would find a shift in the behavior. When we have a La Nina, what is the influence on the jet stream? What influences, if any, do the Rocky Mountains have on the jet stream during a La Nina?


FishOutofWater here responding to John Christianson concerning tropical cyclone energy:

The intensity of the Asian summer monsoon has increased with the warming of the Indian ocean associated with the strong trade wind phase of the Pacific Decadal Oscillation (PDO). The stronger than normal uplift over Asia, especially the Tibetan plateau creates a huge plume of hot relatively dry air that subsequently sinks over the ocean basins. That subsidence inhibits tropical storm formation in both the western Pacific and north Atlantic basins.

However, when the monsoon season is over in November, all hell has been breaking loose in the Philippines and the western Pacific as some of the most intense typhoons ever witnessed have spun up over the growing warm pool in the western Pacific.

Understanding what's happening with tropical cyclones and climate change requires studying what's happening globally with the heat in the oceans how that heat interacts with the atmospheric circulation patterns. Ryan Maue's charts provide good information, but Ryan Maue has failed to analyze what's driving the changes he sees in his charts. The Weather Bell folks may have allowed their conservative ideology to affect their analysis. Weather Bell has good weather information, but they are out in right field on the climate.

Chris Reynolds

P Maker,

Perhaps I am misunderstanding you. In all of the examples you gave mass transport is due to flow. In the case of rivers flow is constrained by the channel of the river. Where flow is not so constrained intertial waves may develop.

My maths isn't good enough to get a really good grasp of the physics, but the Rossby waves (meanders) on the jetstream are constrained to a degree by the fact that the jet forms at the northerly extent of the Ferrel (mid latitude) cell, where it meets the Arctic cell. However like the river, but much faster, the Rossby waves form the channel through which the jet flows, by forming the boundary region between air masses.

Rossby waves are inertial waves, like waves on the sea surface. The restoring force is the force that seeks to 'level things out' and in doing so propagates the wave, in the case of the ocean's waves the restoring force is gravity. In the case of the Rossby waves the restoring force is the variation of the corriolis 'force' with latitude. The coriolis 'force' is highest at the pole, lowest around the equator. So the waves are across the surface in Rossby waves, not up and down like water waves.

Wavelength of a Rossby wave is related to the phase speed - the speed of propagation of the wave, not the jet stream. You can see a graph of this here:
Figure 4.3 around the middle of the document.

In that example for a Rossby wavelength of 5000km, phase speed is zero and the Rossby waves stalls, the weather under it gets 'stuck'. For waves of longer wavelength flow is from west to east shorter waves propagate from east to west. The graph I directed you to is derived from the equation:

C = U - Beta/k^2

Note that U is the zonal velocity of air flow in the jet, this is important for Dr Francis' argument. C is phase speed, Beta is a factor describing the change of the coriolis effect with latitude, k is the zonal wavenumber (don't worry about them).

You state that:

"By analogy it is inferred that an increase in atmospheric water vapor content due to anthropogenic global warming should lead to shorter wavelengths, or - if you wish - larger meanders."

I'm sorry but the analogy is not correct. What Francis is positing is that a slowing of U in the equation above is changing jetstream phase. Simply put, as the temperature gradient between mid latitudes and the Arctic reduces, the speed of air flow caused at the boundary between the mid lat and Arctic air masses reduces, and as a result phase speed is affected such that for certain numbers of Rossby waves events where the progression of the wave along the jet stream ceases or slows massively, and the weather gets 'stuck'.

Actually the UK weather doesn't seem to be an example of this. As I'll show in an upcoming blog post it seems to be due to the anomalously large ridging in the atmosphere over Greenland steering the Jet. For the years with over 20% above average precipitation over the UK I've broken them down into pre and post 2007, before 2007 those wet summers were characterised by the average 300mb flow aiming at the English Channel, for the four years after 2007 the jet is slightly stronger on average (only by about 1.5m/s) and aims at the bay of Biscay. However over the eastern seaboard of the US and through much of the Atlantic jet velocity at 300mb height is ubstantially greater over a larger area post 2007.

Where water vapour comes in may be in increasing activity of storms and total rainfall, but the impact of this is not (IMHO) detectable, and the steering factor is greater.

The complicating factor with any analysis is, as Barnes found to be a critical flaw in Francis & Vavrus 2012, the general rise in the atmospheric height due to AGW. The Rockies are a factor, but my previous analysis of waves in the atmosphere falls foul of the rise of the atmosphere, and I've not had the time to see if I can factor out that issue from my analysis.


SH (Scott?)

Fascinating read at the LDEO site you referred me to: http://www.ldeo.columbia.edu/res/div/ocp/drought/medieval.shtml

I took note of the very thorough analysis of past mega-droughts you presented.

It stroke me that, in the years of those Medieval mega-droughts in your place (~AD 936, 1021-51, 1130-70, 1240-65, 1360-82), we had Nordic people living comfortably in southern Greenland and no multi-year sea ice off the east coast of Greenland.

On the other hand, we have often wondered, why our Viking ancestors took their long boats south in those years, raping, plundering and drinking their way through Paris, Rome and Cairo. If your hypothesis is right, and if the summer weather was miserable in Scandinavia (and the British Isles for that matter), no wonder they fled south.


I will come back to your response in the morning.

Cheers P

Chris Reynolds

Another one in the spam folder....

For those interested I have uploaded some graphics comparing record wet years, before and after 2007, and record dry years, onto my Google Drive here:

Let me know if there are problems getting those images.

David Miller

JC, I'd like to point out that taking the average continental US temperature anomaly isn't going to tell us much about a wandering jet stream.

Think like the statistician with a foot in boiling water and the other in ice water saying 'on average I feel fine'.

To do an analysis like that you'll have to at least break the temperature down into regions and look for both high and low anomalies by region. I think the real mark of a meandering jet stream is simultaneous high low anomalies in different regions.

FWIW, my $.02, YMMV, etc.

Chris Reynolds

David Miller,

I agree.

FWIW here's a timeseries of Arctic (90 to 65degN) to Mid Latitude (65 to 25degN) temperature anomalies (difference from long term mean). It's from an old post of mine so only goes up to 2012.
The downward deviation in all seasons but summer is due to Arctic warming above the rate of warming of mid latitudes.



Coming from either end of a black tunnel and try to meet in the middle, may still be a bit difficult. I will however give it a try.

As I read Jennifer’s hypothesis, she claims that the gradient between the tropics (highest GPH) and the Arctic (lowest GPH) at the top of the dense troposphere (~250-300 hPa) is now less steep than in the old days, because the Arctic has warmed more than the tropics. Hence the Jet Stream is slower now and the phase speed of the Rossby waves has gone down and the meanders (horizontally) are wider now. So far so good.

You then try to introduce the Coriolis force, which – to my knowledge - has not changed in recent years. It will presumably affect all the stream flows I mentioned. Could we please leave this apparent force aside for a moment as an explanatory factor. I hope we can agree, that gravity is driving all the flows I mentioned? I agree that the Coriolis Force is influencing the meandering pattern, but hope you also agree, that inertia is not driving the process we are discussing?

I really like your diagrams at: https://drive.google.com/folderview?id=0B3pB-kdzoLU3WkloQVg1Z09uNUE&usp=sharing – the 1st one clearly illustrates the spiraling pattern of the Jet Stream. Presumably it is driven by evaporation over the warmer parts of the Atlantic and the Pacific oceans, then circles the NH once before it fiddles out over the Arctic. Either the moisture in the Jet Stream precipitates out or the flow descends to a lower pressure level. Maybe somewhat similar to the Colorado River, which runs out of moisture before it hits the Pacific Ocean.

Your 5th diagram clearly shows the changing “atmospheric topography” through which the Jet Stream has been flowing in recent wet UK summers – that means the Jet Stream has been running east on the southern side of the Greenland high and then straight into the “valley” over the UK and then all the way up to the Russian rim of the Arctic Ocean. I think this leaves open the possibility, that advection aloft via the Jet Stream may both explain the recent wet summers in the UK as well as the vanishing sea ice to some extent.

On re-reading Francis & Vavrus, 2012: (http://marine.rutgers.edu/~francis/pres/Francis_Vavrus_2012GL051000_pub.pdf ), I could not help noticing in Fig 4, that over the past 15 years, the 500 hPa troughs over the western Atlantic (during Summer) and the eastern US (during Autumn) have become more narrow. This is an indication, that the jet stream waves have become more compressed. It will be interesting to see, whether the shortening of the Jet Stream wave length is now also taking place during the winter months. If this is the case, we may have found an explanation for your very wet Winter weather this year.

Chris Reynolds

P Maker,

I'm not introducing the coriolis 'force' as a changing factor. Unless the earth's spin changes that will not change. But it is a factor in the basic equations. Francis' concentration on geopotential heights is due to the framing of the issue in those terms in the paper "Winter Northern Hemisphere weather patterns remember summer
Arctic sea-ice extent" by Francis et al 2009. The geoptential height changes are due to tempeature changes, thus the two views are complementary.

"Your 5th diagram clearly shows the changing “atmospheric topography” through which the Jet Stream has been flowing in recent wet UK summers – that means the Jet Stream has been running east on the southern side of the Greenland high and then straight into the “valley” over the UK and then all the way up to the Russian rim of the Arctic Ocean. I think this leaves open the possibility, that advection aloft via the Jet Stream may both explain the recent wet summers in the UK as well as the vanishing sea ice to some extent."

Sorry, I don't understand any of this. The 300mb zonal wind plots merely show the jet's average position in the years stated. There are no valleys and rims in the sense of the jet. Aside from the Greenland ridge (in GPH) other features positions and intensity - notably the UK low pressure are the result of the jet. In the case of record dry years, the break in the jet progression over the UK is due to the strong block (red mass on dry years' GPH).

However I do not consider it sound to rely on Francis & Vavrus 2012.

Here is figure 3 from Barnes 2013.

You'll note that the vertical axis shows 500mb GPH (the height at which 500mb pressure is found), while the wave extents are on the horizontal axis. If F&V2012 were correct then the difference between the periods 1980-1995 and 1996-2011 should be apparent in a shift along the horizontal axis.

In fact wave extent (horizontal) shows little shift, with the greatest shift in the GPH's.

Furthermore Barnes uses seasonal and daily extents of the rossby waves. While seasonal extents show an increasing trend, daily do not!

Thus F&V2012 is finding the general rising of the atmosphere with warming, their method is not finding the north-south extention of rossby waves on the jet.



I told you, it would be difficult to find each other in a dark tunnel...

However, I was referring to your diagram #5 with the label: "Summer Jet Record Wet after 2007 - GPH.png". Hope this helps.

I will need more time to dig into Francis et al. (2009) and other links you provide.

Janne Tuukkanen

There might be El Niño in the making:


John Christensen


It has been a pleasure following all of the blog entries in the last few days, encouraging me to add a few comments:

Great discussion, very helpful to those of us with less formal studies in this field.
I intuitively understand Jennifer Francis' argument that a strong jet stream will plough through a local low or high, but since most of our winter time highs and lows of significance are supported by considerable topographical features (Greenland, the Rockies, Norwegian west coast and the Central Siberian Plateau), I could also see the jet stream working like a water hose being sprayed against a rock in its path: If you increase the water speed of the hose, the water will deviate more from its original path, once it hits the rock, not less.

@Neven/idunno on Continents:
I would tend to agree with idunno that we need to observe the continents more closely, as the Arctic is warming. It has received very little publicity, but as I have referenced a couple of times the unusual Central Siberian Plateau high of this winter has been impressive, causing significant negative deviation from average temps. (http://www.ncdc.noaa.gov/sotc/service/global/map-blended-mntp/201401.gif). So did the lack of Arctic ice influence/enforce the Siberian high, or was the Siberian high so forceful that it impacted weather around the NH? Or both ways to some degree?

David wrote on temps:

JC, I'd like to point out that taking the average continental US temperature anomaly isn't going to tell us much about a wandering jet stream.

Think like the statistician with a foot in boiling water and the other in ice water saying 'on average I feel fine'.

To do an analysis like that you'll have to at least break the temperature down into regions and look for both high and low anomalies by region. I think the real mark of a meandering jet stream is simultaneous high low anomalies in different regions.

You are right in principle, and I did point out that I had not performed any detailed analysis. However, I would argue that if CONUS temps are 3-4C below average, you will always find even higher regional deviations, as the US rarely would observe similar weather pattern coast-to-coast.

I did check the regional temperatures for California, and West Virginia (WV was near the center of the January cold spell this year) for the coldest Januaries across CONUS( 1901-2000 average = -0.9C):

1912 (-4.0C): Cali: Warm, WV: Cold, supporting standard jet stream 'dip' across mid/eastern US, while west coast stays warm.

1918 (-4.3C): Cali: Cold, WV: Cold, supporting very wide/strong jet stream dip, impacting across the US.

1930 (-4.4C): Cali: Sligthly cold, WV: Average. Major cold outbreaks this year across Colorado, Texas deep into the middle of the continent. Narrow and very strong jet stream dip.

1937 (-3.7C): Cali: Coldest on record, WV: Warm. This was an usual month with the weast coast being very cold, but near average to the east.

1940 (-4.4C): Cali: Warm, WV: Cold. Standard jet stream dip across central/eastern regions, leaving west warm.

1963 (-4.1C): Cali: Sligthly cold, Wv: Sligthly cold. Very wide jet stream dip, slightly less intense cold.

1977 (-4.8C): Cali: Sligthly cold, WV: Coldest January on record. Very wide jet stream dip, very intense in eastern regions.

1979 (-5.5C): Cali: Slightly cold, WW: Cold. Coldest CONUS January on record. Very wide jet stream dip, with intense cold in central regions.

2014 (-0.95C): Cali: Very warm, WV: Cold. Standard jet stream dip across central/eastern regions, leaving west warm, but not near record level cold and reduced geographical reach of jet stream dip.

And finally, top ten cold januaries for West Virginia:

1977: -8.2C
1940: -6.6C
1918: -6.6C
1978: -5.2C
1912: -5.1C
1948: -4.5C
2014: -4.3C
1994: -4.3C
1970: -4.3C
1981: -4.2C

So this is my point: Yes, we saw a US cold outbreak due to a dip in the jet stream, but the dip was both less intense in temperature deviation as well as in geographical reach, indicating that prior major jet stream dips have been either larger, colder, or of longer duration.

The jet stream of this year has been either smaller, less cold, or of shorter duration, and therefore does not merit the branding of 'extreme weather' IMHO. Again, if there is any data available pointing in a different direction, please share that.

On the Euro/Asian side of things this winter, the NAO played an insignificant role, as Greenland and the Central Siberian Plateau seemed to have had the main interaction with the jet stream, splitting it in a northernly branch pushing lots of moisture north into the Barentz region, as well as allowing a strong jet stream going across mid-Asia regions south of Siberia across China into the Pacific and then moving towards the North causing the high anomalies in Alaska and north-east Pacific.

Did lack of Sea Ice cause the Siberian high to get stronger than usual this year, or did the indecisive NAO allow the Siberian high to get well-established, most likely leaving it in place, until spring-time heat enters the area?

If there is no new data available yet to support the theory of extreme weather becoming more likely with a weakened jet stream, it should then be verified that a strong jet stream to some degree prohibits or limits the occurence or extend of these same extreme weather events.

This is what the data does not seem to support and why I find that without support in data analysis, the theory does become speculative.

Jai Mitchell

It seems that the thread has shifted from arctic albedo and NH summer snow cover anomalies to Barnes et al and northern hemisphere continental cooling. . .

FYI, there is significant indication that what we are actually seeing re: cooling continents/warming arctic isn't so much due to albedo but rather an expansion of the Hadley cell that is driving the mid-latitude cell to "pulse" into the arctic.

for example: on 2/22/14 at 02:00 a pulse of mid-latitude latent heat was pushed through the bearing straight and back up over the Chukchi and East Siberian seas.

see animation here: http://weather.utah.edu/index.php?runcode=2014022110&t=gfs004&r=NH&d=DT

This "pulsing" of the mid-latitude cell into the arctic has been successfully modelled and more recently proposed as a primary response to global warming here:


This work suggests that the expansion of the Hadley cell may be enough to cause the eddy-driven jets to shift poleward with climate change, and that the leading mode of zonal-wind variability will transition from a shift to a pulse as the jet moves poleward.

Colorado Bob

Scientists from The University of Queensland have discovered a microbe that is set to play a significant role in future global warming.

UQ’s Australian Centre for Ecogenomics researcher Ben Woodcroft said the methane-producing micro-organism, known as a ‘methanogen’, was thriving in northern Sweden’s thawing permafrost in a thick subsurface layer of soil that has previously remained frozen.
Mr Woodcroft said no one knew of the microbe’s existence or how it worked before the research discovery.

Jai Mitchell


what we are seeing is a climate change-driven transformation of the earth's Hadley cells: The Hadley cell is growing/expanding northward, pushing desertification into the Midwest, the mid-latitude is pushing into the polar cell, eventually, the Hadley cell will grow in strength to the point where it will push the mid-latitude and polar cells to collapse together.

we are just beginning to witness this transformation now.

Colorado Bob

"We found as frozen soil thaws, this microbe produces more methane," says lead researcher Associate Professor Gene Tyson of the Australian Centre for Ecogenomics at the University of Queensland.

"This causes the temperature to rise and more thawing to occur in a massive feedback loop. It's a really, really big problem,"


Steve Bloom

Jai, the Barnes paper you linked is the submitted draft. The final is http://www.atmos.washington.edu/~dennis/Barnes%26Hartmann_Scales2011.pdf. I haven't looked to see if there are significant differences.

But on the substance, this is fascinating and seems rather prescient relative to recent weather patterns. Also, it hadn't been clear to me that Barnes has what seems to be a competing hypothesis relative to Francis and colleagues.

Having not actually read the paper carefully or checked subsequent work, I don't want to overstate things, but this does sound like the sort of instability one would expect in the earliest stage of a collapse of the NH circulation (begging the question of collapse into what).

Chris, have you looked into this line of research?

Steve Bloom

It turns out there are lots of citing papers, not just from Barnes, a fair number of them conveniently with public copies:


Much reading ahead for me, just as soon as I'm done kicking myself for somehow missing this stuff.

Steve Bloom

Bob, those bacterial methane results would seem to put much more weight behind DeConto et al. (2012)'s explanation of the Eocene hyperthermals (starting with the PETM), although I suppose there would be no way of directly determining whether that species or a functionally-equivalent one existed at the time:

'ware the yedoma.

Jai Mitchell


not competing lines of research, the warming of the arctic necessarily weakens the polar cell boundary, so Francis is also correct.

These two factors work together.

Steve Bloom

Jai, Barnes certainly seems to act as if they're competing, although perhaps you're right that elements of both could co-exist. But on your specific point, isn't it the case that polar cell boundary weakening as such is uncontroversial? I had thought the dispute centered on the detectability and extent of jet amplitude lengthening.


Y'all might run the model out here:


Jai Mitchell


I may be significantly off here but I understand that Francis found that the polar jet stream slowing down and correlated it with a warming arctic. To my understanding this is not controversial.

She then said that this dynamic leads to a meandering (increased meridional) action in the jet stream, that leads to more blocking patterns that produces the weird weather we have been seeing.

Barnes simply said that there was no significant correlation of north/south Jetstream speed and blocking patterns with arctic warming.

Neither of these analyses speak to the potential expansion of the Hadley Cell and its effect on the polar cell boundary, specifically the effect of increased pulses pushing into the arctic and disrupting the polar jet.

I am not sure but I think that the barnes' paper might have a correlation issue with the time duration/intensity of individual blocking events. not sure though.

Jai Mitchell

The reason I suspect the issue in barnes' is off a bit is her apparent correlation to blocking event intensity (N/S Jet Speed) vs. blocking even duration.

both images taken from this analysis http://dosbat.blogspot.co.uk/2013/01/northern-hemisphere-blocking.html

Colorado Bob

Steve Bloom -
Woodcraft has most of that little bug's genome sequenced, I got a quarter that says that little rascal has been hanging around for a very very long time. I suspect his ancestors are sleeping deep in the yedoma as I type.

Thanks for that pdf.

Chris Reynolds


Good to hear from you again. No I've not seen that research, my forays into the atmosphere only really directly relate to sea ice and Arctic changes. But I'll have a look at it in the next few days. Having just read the conclusions and abstract I suspect I just don't have the background to evaluate it properly

Jai Mitchell,

Sorry I don't get what you're saying. Bear in mind that different blocking indices find different things and it's hard to tell which is right sometimes. The problem that Barnes outlines - that F&V2012 are detecting the thickening of the atmosphere due to AGW - still stands and Dr Francis hasn't done anything to change that problem as yet. Were it an easy problem she'd have sorted it out by now. Her recent statements that the timeseries is too short are IMO correct, as I outline above - there are sound reasons for expecting effects on Rossby wave phase velocity.

As for winter cooling, I think Cohen's nailed that one.

Colorado Bob

Steve Bloom -
What I found really interesting is that methanogen works with 2 other microbes to produce the methane . The classic compost pile , where groups of microbes move in concert to break down organic matter. This points to it being very very old.

The real question is what temperature is it's preferred range.

Colorado Bob

Steve Bloom -
This one has had me worried for years. What really bothers me is all the other microbes being unlocked from the ice. For example, what if a mammoth microbe melts out and bumps off every elephant on the planet ? It’s not like there’s a giant pool of them now.

Jai Mitchell


There seems to be an obvious correlation to blocking duration, even if there is not one to frequency. That is all that I was saying, with regard to atmosphere thickness, I believe that Francis' response provided her position that the barnes results match well with her paper,

says Francis,

"She claims that because warming is shifting a particular height contour northward, it is incorrect to conclude that wave amplitudes are increasing. In fact, it is this northward shift – in particular the larger shift in high latitudes where warming is greatest – that we hypothesized would be a factor causing the waves to elongate."

I believe that the actual controversy regarding Barnes vs. Francis is that Barnes utilized speed values at the 250mb Geopotential Height instead of the 500mb range and that this implicates her paper as an intentional hit job (in Frances' view)

Figure 4 presents measures of wave phase speed. While FV12 did not present wave speeds, we speculated that larger amplitude waves should have slower wave speeds. Her measure of phase speed for waves at 500 hPa slows with time, supporting our speculation. She then measures speeds at the 250 hPa level and finds no change in speed. This much higher level is near the tropopause, often above the jet stream, and can be affected by dynamics of the stratosphere. The stratosphere is cooling with increasing greenhouse gases, leading to very different dynamical changes. Why did she choose to analyze this level? My only guess is to deliberately cast doubt on FV12.

The comments to this entry are closed.