« Ever sailed to 85N? | Main | New NASA videos »


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


Through the last few years, I acquired the notion that it’s not much use to focus solely on the ‘run for the yearly sea ice minimum’. My interest lies in the broad picture of change, climate-wise and ecological.

I haven’t done much homework, so haven’t been posting much through July-September.

Nevertheless, I’ve been following the Forum threads on ENSO, home brew extent and area. Of course I let my eyes wander over the daily Lance Modis pics. The same goes for OSDPD SST. Etcetera.

It is very interesting to learn how the ice pack seems to hold on, based on the parameters extent, area, volume. Because it is encircled by unusual high SST’s on the Northern Hemisphere.

It’s the Barentsz Sea and, to lesser extent, Kara Sea, that remain a tad colder than the climatology. I’ll try to express why that feature may have played an important role in keeping the parameters ‘comfortingly high’ (at least, no disaster yet).

But first, against the Barentsz/Kara ‘cold’, stands the ‘Bering heat’. I can’t remember seeing that Sea that warm through the last decade. The ‘run for the yearly minimum’ may well be not as important compared to the gradual, year-round accretion of heat and, corresponding, the degradation of ‘winter power’.
Of course, the Bering Sea cannot be held representative for the Arctic Ocean. But, an effect even stronger than last winter, there will probably not be much ice out there in the coming months.

Will that be noticeable in the peripheral Arctic Seas? I’m not sure how strong the warming in their waters is. Nevertheless, there’s a good chance winter will again prove to be not the best season possible for the ice (relatively of course; ice will form).

What role did the Barentsz/Kara sector play this melt season? The main atmospheric feature over the North Atlantic has been high pressure. Both on 1000 and 500 Mb. Attached to the Greenland Ridge, it effectively blocked the ‘Fram train’.
Wind steering over Frantsa Yosefa has also been dominated by a persistent Low near Severnaya Zemlya. It kept the Barentsz cold. It held extent high in Victoria Strait. There was the usual fragmentation in the whole sector Pole/Svalbard/Fr.Yosefa. But relatively little melt. And much more extent than we were used to all past decade.

The whole atmospheric pattern didn’t support compression into the North Atlantic sector and strong melt when ice would even have had a chance to pass into the Greenland and Barentsz Seas.
Though the Alaskan and Kolyma coasts were warm enough, winds usually blew clockwise, not penetrating deep into the Central Arctic Basin. In that config, the ‘Laptev Bite’ was the only interesting feature this summer, where winds did transport warmth from the South.

As a detail, late this season Foxe Basin and, though to lesser extent, Gulf of Boothia are nearly icefree.

These details, the Bering Sea, the SST’s, the possible ENSO driven ocean-atmosphere coupling next season, all combine to stay alert. Even in winter.


Hi Jay,

I just read your post. I think your observations are interesting. That sort of pattern captured my attention too. But on the suggestion of 'geo-engineering'... I see the pattern(s) more in the sphere of 'teleconnections', natural responses to the growing heat imbalance.
Why would the Californian economy be sacrificed imminently for an attempt that could, at best, provide uncertain results?


I've said this before, and I'm not quite sure why it hasn't sunk in, since you seem to be generally rather sharp. Your "Summer Pattern" is an almost exact match for the NAO principal(sp.) component as calculated solely from 2000+ data. Please either post your "Summer Pattern Correlation" numerical data so I can do the calculation myself, or calculate the correlation function between the JJA AO and the JJA NAO and your "Summer Pattern Correlation" for 2000-2014 and 1979-2014. The fact that the correlation is huge between the NAO and the "Summer Pattern Correlation" should provide conclusive evidence that they are really different versions of the same principal component.

In fact, the AO and NAO are also different versions of the same principal component. The difference is that the AO, being defined to be a monthly correlation, shows mostly the monthly atmospheric noise component, but the NAO, being defined to be a phase-dependent correlation, picks out the low-frequency component, and shows primarily the effect of sea ice variations and North Atlantic SSTs on surface pressure. Taking the JJA average produces a similar low-pass effect, and and your "Summer Pattern" is an extremely close match for the NAO as calculated using data only for the 2000+ period.

With continuing global warming, the varying SST region has migrated northwards, and warmer SSTs are now seen off the entirety of Siberia, resulting in the low-pressure region along the Siberian coast now seen in your "Summer Pattern", but not seen in the NAO using older data.

Strictly speaking, in order to do a proper principal component decomposition, the multiplicative coefficients used to determine the principal component should be the projection of the observed pressure onto the derived principal component. However, the coefficients for summer during 2005-2012 are all very close to the same, hence your 2005-2012 average is extremely close to the result of a numerically proper principal component decomposition.

I am in fact rather mystified as to why you plot your "Summer Pattern Correlation" on the same graph as the AO but omit the more relevant NAO.


Geh, 2007-2012.


Chris, Re: 2014 500mb JJA pressure:

In general summer 2014, whether looking at surface pressure or 500mb pressure, was much more like the 2007-2012 average than different from it. Consequently the relatively high final sea ice extent and area minima of 2014 are surprising. During 2013, the summer weather patterns were obviously not conducive to melting, and the resutling relatively high sea ice extent and minima of 2013 can reasonably be considered a predicitble result of the "Summer Pattern"- and NAO+ nature of the 2013 pressure patterns. In contrast, the pressure patterns of 2014 do not appear to explain the resulting high sea ice extent and area minima.

Given the general similarity of the pressure patterns of 2014 and the 2007-2012 average, the extremely different results in sea ice area and extent, are hard to explain and the details of the differences between 2014 and the the 2007-2012 average pressure pattern are quite important.

It is worth noting the seriously borked color scale of the linked 500mb height graphic, in which zero is given a deep blue color. This graphic can generally be described as a planteary 3-wave, that is, it shows a hemispheric pattern with three negative 500mb pressure anomalies, and also three positive anomalies. It should be noted that the planetary 3-wave is a normal feature of summer weather, and any local reason for a matching fixed pressure in any of these 6 locations will also tend to result in the corresponding pressure correlation in 5 remaining locations.

During summer 2014, the PNA Pacific Ocean SST pattern has been near record levels. This tends to result in a PNA+ atmospheric low over the Pacific Ocean PNA locus, which has unsurprisingly been observed during summer 2014. The corresponding PNA+ high pressure pattern over western Canada has also unsurprisingly been observed.

During summer 2014, sea ice levels over the Laptev Sea have continuously been at record low levels. Unsurprisingly, over the Laptev Sea, surface low and 500mb low pressures expected in the event of an SST high tempreatures have also been observed.

During 2014, positve pressure anomalies both at 500mb and at the surface have been observed both in far eastern Siberia, and in western Siberia and Europe. The positive pressure anomalies off the coast of Norway have been particularly extreme.

Negative pressure anomalies over Greenland (NAO+ and "Summer Pattern"-) both at the surface and at 500mb have been also observed in 2014. It should be noted that these anomalies are relative to the 2007-2012 climatology, which is extremely NAO- and "Summer Pattern"+.

Reltive to other 2007-2012 years, 2014 saw very little open water in May and June in eastern Siberia as a result of southern advection of sea ice away from the Eastern Siberian coast.. 201. May-September can in general be characterized as a year of very low heat transport from high-latiltude land into the Arctic Ocean.

Projection of observed cloud cover on to a disc comprised of the area above 77 degrees North shows that sunshine over the high Arctic during 2014 was most probably near record levels in 2014, although likely very slightly behind that of 2012. However, given observed albedo differences between 2012 and 2014, absorbed incident sunshine in 2014 was likely far below that of 2012.

Despite abundant arctic sunshine during 2014, which was in fact above the 2007-2012 average, observed albedo decrease lagged far behind other recent years.

In comparing the summer 2014 pressure patterns to the 2007-2012 climatology, the most obvious difference is that the low-pressure ring around the Siberian coast (relative to earlier climatologies), which is present in the 2007-2012 climatology is now entirely absent, except for in the Laptev Sea. Instead of being an area of predictable storm formation, the Russian coast has become an area of predictabale doldurms, where it is expected that no significant transfer of heat between the Arctic Ocean and coastal Arctic land. Indeed, it is this lack of transport of heat, which should it happen would presumably precipitate melt ponds and albedo-feedback, which most signicificantly characterizes the differences between 2014 and the 2007-2012 climatology.


After fighting through multiple error messages to get the previous comment to actually post, let me just say, for the record, "I hate typepad", "I hate typepad", "I hate typepad", "I hate typepad", "I hate typepad", "I hate typepad."


Thanks a lot for your comments, Blaine, and sorry to hear about your problems with TypePad. Don't forget: TypePad loves you. ;-)

So, to summarize, 2014 in principle was similar to 2007-2012, atmospherically speaking, but because of the way the atmosphere was set up - ie distribution of highs and lows - less heat entered the Arctic. This caused a start of the melting season that saw very little melt ponding, which, if the work by CPOM scientists is correct, has consequences for the minimum.

Although the Sun reached the ice due to large high pressure areas over the Pacific American side of the Arctic, the lack of low pressure on the Siberian side caused a more or less static situation where the ice was melting in situ, instead of being sloshed around, compacted and transported towards lower latitudes. The result was very marked (compared to previous years) on the Atlantic side of the Arctic. The ice over on the Pacific side was strengthened compared to 2013, and thus we saw the same phenomenon as in 2010 and 2011 where a barrier of multi-year ice protects the rest of the ice pack behind it (from warmer SSTs etc).

Because of all this (lack of melt ponds and transport) the ice in the Central Arctic thickened some more, and the rate of PIOMAS modelled volume decrease slowed down considerably.

Is this summary about right?


Morning, Neven,

Your summary is pretty right. It's just that your interpretation 'the ice...thickened some more' could be better imaged as 'didn't thin as much as other post '07 years'.
We'll have to wait and see whether next winter will indeed induce good thickening. A priori, I'm skeptical that that thickening will be enough to produce much sustained volume growth.


Neven, since you ask..

There may be another explanation, but it is difficult to stick the arguments together, because some of the diagrams I need to explain are hidden in various threads on the Forum. The main story goes like this (details may come later):

One diagram showed that in recent years, the volume loss was most conspicuous in Jun, May and possibly Apr. Losses in Jul, Aug & Sep were pretty linear (boring). At the same time we have not seen any melt ponds in recent years, which to some have been a kind of a mystery. 2013 may have an obvious reason (persistent lows), whereas 2014 may require another explanation. The only way to lose extensive amounts of sea ice without forming melt ponds , is – as far as know – sublimation.

Could it be that the atmosphere in spring (after spring equinox) is now warming so rapidly (due to build-up of GHGs), that ice and snow has simply started to sublimate instead of melting?. We have seen the most dramatic cryospheric changes in the NH snow cover in May & Jun (without the usual spring floodings). This development is now paralleled by massive losses of sea ice during the same months (without the melt ponds).

Once we cross the summer solstice, the atmosphere in the NH starts cooling and consequently we should see the opposite of sublimation (i.e. condensation) taking place large-scale. I myself have been living mostly in mid-latitude climates, and I have been used to this effect (called dewfall which has been most prominent in Aug and Sep) all my life. Suddenly this year in spring and early summer, I have for the first time experienced substantial dewfall in May, Jun & Jul. I was wondering how this came about, but could not find an immediate explanation. (This change came after a winter season, where I also experienced drizzle in a so far unprecedented manner, but that is another story)..

However, after reading the thought-provoking paper on the melt vs sublimation process, I came to realize that this change from melting to sublimation large-scale should also have implications for the dewfall (what goes up – must come down). So if this dewfall-effect is not picked up by models, we will not be able to project when the remaining sea ice will be gone.

I do realize that dewfall does not really melt any ice, because the heat energy will stay up in the air, whereas the dew will fall onto the ice or the adjacent ocean area (with a temperature close to 0 deg C). The same issue is with drizzle, which is also a way of removing excess moisture from maritime air masses (leaving the heat energy above the clouds), when these air masses are advected over land areas.

So to sum up: As long as we don’t measure dewfall, we will not be able to understand what happened with the sea ice this year.

John Christensen

Neven, you wrote:

"2014 in principle was similar to 2007-2012, atmospherically speaking"

I really do not think so - let me again provide the NAO JJA perspective:

2007: -1 at max/short period
2008: -2.5 at max
2009: -1.8 at max/short period
2010: -2.3 at max/extensive period
2011: -2.1 at max/short period
2012: -2.7 at max/extensive period
2013: +1.1 at max
2014: -1.3 at max/short period

From this:
1) 2007, 2009, 2013, and 2014 had weak NAO signals, where 3 of 4 (2009, 2013, and 2014) were rebound years.
2) 2010 and 2012 had the combination of strong negative NAO and extensive period and had the strongest ice losses of these years.

Prior to 2008, strong negative NAO during summer months was quite uncommon, and since 1979 only occured in 1980, 1993 and 1998..

Susan Anderson

With some hesitation I'd like to address the Arctic incursions we've been experiencing in parts of the US and Europe. They affect me when I'm in New Jersey (my parents' house, up for sale, has lost value due to multiple storm events and damage, and cleanup work has been extensive with ice downed trees and new deaths in the garden from the deep freezes), but are much worse to the west and south, particularly where freezing is not a regular occurrence (Georgia, for example). Other regions are quite hot, it's that central bulge coming down through Canada that has been repeating and has not stopped even over the summer. This is very simplistic and may be quite wrong, in which case if anyone can address that in language I can understand I'd appreciate it.

I have stopped calling it the "polar vortex" as that is subject to misinterpretation, but it does seem that Dr. Jennifer Francis's hypothesis applies. The circulation that used to largely contain the Arctic has broken, resulting in a reversal of cold and warm, almost as if the Arctic came south. It continued to happen over the summer, and a meteorologist friend warns me we may have more of the same for the foreseeable future, until the Arctic is more melted out. It's almost like an exhalation of cold, and it's all too easy to think the cold is leaving the Arctic, but I'm guessing it's more complicated than that.

I know phony skeptics like to say, ha ha cold, but this pattern makes very good practical sense to me and has penetrated the general incomprehension of how weather and climate interact with the general population as well, making it easier to talk with them about how climate change, global warming, and greenhouse gas heat trapping are affecting our lives.

Jai Mitchell's work, of course, is detailed and skilled, but from an amateur point of view, this all seems quite logical to me and I would not be ready to conclude that the Arctic will build up because of it. I'll have to take a closer look.

Jai Mitchell


No, I don't see it as the reduction of heat entering the Arctic, I see it as an increase in mid-latitude water vapor entering the arctic, leading to increased relative humidity, cloud cover and cooler temperatures.

At least, the correlation between temperatures and the observation says so.

Susan Anderson

I see that much of what I'm trying to explain/ask about in my amateur way is covered in middle of the final video in the "New NASA Videos" article; no need to repeat, just take a look if the subject interests you ("Arctic Emergency: Scientists Speak").

Jai Mitchell

As a side note, it seems to me that to blame the "Arctic Incursion" solely on arctic temperatures is derived from an "east coast bias" provincialism. In which all things that happen to the east coast of the U.S. are isolated from events happening on the west coast.

Without the Ridiculously Resilient Ridge semi-permanent blocking pattern in place in the North East Pacific there would be normal (colder) temperatures in Alaska and no returning pulse of the jet stream southward through the Yukon and into Alberta and the Eastern Half of the U.S.

To disregard the pacific blocking pattern as the root cause (and then to determine if this pattern is associated with arctic conditions, or not, is isolationist and not inclusive of the entire system.


Chris Reynolds


The problem I am having is as follows:

1) I have read multiple references in papers, granted mainly winter, if I recall correctly, to AO/NAO like patterns. The formation of the AO/NAO patterns by land/sea masses is likely to apply to other patterns driven by various factors related to Arctic climate change. Yes the 2007 to 2012 summer pattern is like the AO/NAO, but I view it as significantly different enough to justify viewing it differently.

2) The AO/NAO has not in the past driven strong Dipole like behaviour within the Arctic.

3) The AO/NAO has not in the past been associated with the unusually high geopotential heights seen over Greenland.

4) 2007 to 2012 all fall within the eleven lowest NAO indices for June July August. Excluding 2014 from the 2007 and 2012 due to its complication.

Here is average SLP for 2007 to 2012.
Here is average SLP for the ten lowest NAO years excluding 2007 to 2012.
I've presented these as averages, not anomalies as some have complained about the 1980 to 2010 anomaly period partly covering 2007 to 2012. The shift to a more Greenland centred behaviour after 2007 seems clear to me. Needless to say, similarity in the Pacific and Atlantic is irrelevant as this is topographically driven, and topography has not changed.

Yes the AO/NAO are essentially the same thing. I use the AO because it reflects Arctic Ocean dominant pressure.

Summer Pattern Correlation. I've resorted to using correlation because my maths aren't up to PCA, which would have been a preferred method.

June July August
1979 -0.440 -0.297 0.386
1980 -0.081 0.738 -0.262
1981 0.516 -0.389 -0.571
1982 0.724 -0.539 0.023
1983 -0.800 -0.134 -0.272
1984 -0.048 -0.114 -0.588
1985 0.249 0.265 0.446
1986 -0.670 -0.081 0.250
1987 0.721 0.209 0.486
1988 -0.557 0.485 -0.241
1989 -0.639 -0.695 -0.032
1990 0.015 -0.334 0.478
1991 0.016 -0.197 -0.862
1992 -0.232 -0.499 -0.388
1993 0.285 0.382 0.048
1994 -0.797 -0.622 -0.442
1995 -0.400 0.657 -0.157
1996 -0.654 -0.733 -0.340
1997 0.795 0.252 -0.600
1998 0.669 0.147 -0.261
1999 -0.636 0.399 0.680
2000 -0.181 0.140 -0.097
2001 -0.304 0.189 -0.166
2002 -0.576 -0.489 0.179
2003 0.254 -0.088 -0.423
2004 0.382 -0.557 0.475
2005 0.463 0.260 0.386
2006 -0.682 -0.097 0.027
2007 0.725 0.759 0.532
2008 0.612 0.774 0.567
2009 0.685 0.784 0.436
2010 0.660 -0.405 0.639
2011 0.787 0.788 0.824
2012 0.633 0.316 0.459
2013 -0.755 -0.126 -0.463
2014 0.470 0.394 0.656

Correlations with NAO by month.
Jun Jul Aug
1979 to 2014, -0.870, -0.657, -0.517
2000 to 2014, -0.837, -0.780, -0.580
1979 to 2006, -0.857, -0.574, -0.288

Note the drop in correlation through the summer when the post 2007 period is excluded.

However I will defer to your expertise and desist from mentioning the matter again. I'm busy enough with sea ice and the day job.

Pete Williamson

Given that some of the discussion here is about atmospheric circulation/dynamics I thought I'd draw peoples attention to this new paper in discussion

Recent summer Arctic atmospheric circulation anomalies in a historical perspective
A. Belleflamme, X. Fettweis, and M. Erpicum
The Cryosphere Discuss., 8, 4823-4847, 2014


it looks at recent changes in circulation over the arctic, how it may affect ice melt and compares it to the past >100years.

heres my summary of it, people can disagree with what I've highlighted as important.

The paper is looking at the frequency of daily high pressures over different parts of the arctic. It identifies two patterns (type2 and type4) that it shows have doubled in the 2007-2012 period compared to the long term average. these type are highs over Beaufort and highs over Greenland, these are both good for increases in melting sea ice and GIC. They construct an historical record of these circulation patterns and compare this recent event to past events. They find similar departures for these two patterns in the past although given we are still going through the present anomaly its not possible to say whether the present is outside 'natural variability'. They do mention though that these departures are seen ~100 years ago when presumably both the global and arctic temperatures were cooler.

What does this study prove? I like to think whats plausible as there seems to be enough uncertainty and insufficient data to make firm conclusions (as is often the way).

It seems plausible (maybe even likely) that multi-annual to decade changes in atmospheric circulation have contributed to the post-2007 melt. I think if you look at their graphs you could say that the type2 circulation has been high since 1980s with the present departure coming on the back of that and the type4 has been increasing since the late 1990's

Attribution of what cause these changes (crudely is it natural or anthro or a bit of both) is impossible from this paper but to me it seems evidence points to it (at least) in part being natural. For example similar events occurred at the end of the 19th century when presumably global and arctic temps were lower and sea ice was higher.

How the future develops might help to understand this. they note just what a huge reversal 2013 was. The daily occurance of these two circulation patterns went from way above the historical average to way below. I;m guessing 2014 is similar. presumably if this continues forward in the future then we may get a better idea how the present circulation anomalies compare to historical events and also to what extent its natural and to what extent its had an impact on declining ice.

james cobban

@Susan Anderson: With some hesitation I'd like to address the Arctic incursions we've been experiencing in parts of the US and Europe.

Calgary Alberta just got hit with a very unusual three-day summer snowstorm, which ended yesterday after dumping 35cm of snow and ice, knocking out power to large parts of the city, and downing thousands of trees:


The temperature was +25C the day before the storm hit.

I really do not think so - let me again provide the NAO JJA perspective:

John Christensen, you have a point, of course, but at the same time, like Chris Reynolds says: AO/NAO do not necessarily tell us where the highs and lows are positioned, and this is crucial.

It's the reason I tend not to look at AO/NAO that much, although it is a rough indicator.

No, I don't see it as the reduction of heat entering the Arctic, I see it as an increase in mid-latitude water vapor entering the arctic, leading to increased relative humidity, cloud cover and cooler temperatures.

I remember you (and others) writing about that on the ASIF at the time. It's a very interesting explanation of why temps were lower this year, despite periods of high pressure and thus insolation.


@Jai Mitchell

Not to disagree with any of your discrete observations, which I'm sure are accurate and well-informed. And "... an increase in mid-latitude water vapor entering the arctic, leading to increased relative humidity, cloud cover and cooler temperatures" makes sense. I'm curious whether it also relates to fog cover (if in fact spring or summer 2014 was unusually high) and if the timing would have affected melt pond formation.
However, my lay impression is that for this melt season as a whole, atmospheric patterns and attendant pressure gradients resulted in less-than-normal heat advection into the Arctic from lower latitudes. Do you know of any aggregate data that show the opposite?


james cobban:

Of course the damage would have been far lesser if the deciduous trees weren't still in full leaf.

Insurance costs due to weather-related disasters in Alberta have skyrocketed in recent years. The insurance industry is taking note.

"Severe weather in Alberta is not an anomaly," said Heather Mack with the Insurance Bureau of Canada. "It is just the way things are now. And it has become a trend."

The last paragraph is from this article:



Couple of astounding things going about, first the NW passage is experiencing a great NW wind blow event, suddenly opening the main channel a greater deal, but there is still melting in the mix, look at archipelago ice trapped in Straits and Channels melting daily. Second, JAXA sea ice towards East Siberian sea is remarkably disappearing fast very late in the season:


Third Cryosphere Today and NSIDC are reporting ice where I cant prove its existence, especially next to Brodeur peninsula, on EOSDIS there is nothing in Admirality Inlet, nothing significant visible flying over by air as well. Yet there is ice reported. Same
thing next to Bothia peninsula if you compare with JAXA and EOSDIS.

Finally Susan, if the Arctic temperatures change greatly, expect radically new weather to the South. The entire circulation system rearranges itself according to where the heat increases.

John Christensen

Chris and Neven,

I still find that you are both off track with regards to the influence and importance of the NAO.

Chris wrote:

"3) The AO/NAO has not in the past been associated with the unusually high geopotential heights seen over Greenland.

This is not correct.

See this article from DMI explaining the exceptional high in 2012:


Key note from this article:

"Our analysis allows us to assess the relative contributions of these two key influences to both the extreme melt event and ongoing climate change. In 2012, as in recent warm summers since 2007, a blocking high pressure feature, associated with negative NAO conditions, was present in the mid-troposphere over Greenland for much of the summer."

And see this article (NOAA) comparing the conditions in Greenland in 2013 to the 2007-12 period:


From this article:

"Meteorological Conditions

In contrast to the previous six summers, summer 2013 was characterized by a positive North Atlantic Oscillation (NAO) and persistently lower-than-normal 500 hPa geopotential heights. Consequently, warm, southerly air masses were diverted eastward away from Greenland and cool northerly airflow in west Greenland (see Fig. 4 in the essay on Air Temperature) promoted cooler, wetter and cloudier weather than normal, and less melting than in recent years, as reported above. This is reflected in the NSAT data (Table 7), which show that during the summer months (June, July, August) NSAT values were generally near or below one standard deviation of anomalies relative to the 1981-2010 baseline period, indicating that summer 2013 NSATs were "normal" with respect to that period.

Are you sure you want to hold on to that argument above?

John Christensen

"Yes the AO/NAO are essentially the same thing."

No. The AO is very broadly defined across the northern hemisphere, while the NAO specifically shows the relation between Iceland lows and Azores highs - which is an important index.


"The Arctic oscillation (AO) or Northern Annular Mode/Northern Hemisphere Annular Mode (NAM) is an index (which varies over time with no particular periodicity) of the dominant pattern of non-seasonal sea-level pressure variations north of 20N latitude, and it is characterized by pressure anomalies of one sign in the Arctic with the opposite anomalies centered about 37–45N."

Negative AO therefore typically causes the Arctic jet stream to become more wobbly, while a positive NAO tends to have the opposite effect.

Last winter we had a positive AO, but wobbly jet stream, so Arctic weather was clearly impacted from elsewhere, which Jai has pointed out.


"Through fluctuations in the strength of the Icelandic low and the Azores high, it controls the strength and direction of westerly winds and storm tracks across the North Atlantic. It is part of the Arctic oscillation, and varies over time with no particular periodicity."

As I see it, the AO is too broadly defined to be useful, while the NAO and the PNA provides an improved level of insight into how atmospheric conditions in the Northern Pacific and Northern Atlantic are influencing Arctic weather and ice conditions.

John Christensen

Hi Susan Anderson,

You wrote:

"With some hesitation I'd like to address the Arctic incursions we've been experiencing in parts of the US and Europe. They affect me when I'm in New Jersey (my parents' house, up for sale, has lost value due to multiple storm events and damage, and cleanup work has been extensive with ice downed trees and new deaths in the garden from the deep freezes), but are much worse to the west and south, particularly where freezing is not a regular occurrence (Georgia, for example). Other regions are quite hot, it's that central bulge coming down through Canada that has been repeating and has not stopped even over the summer."

You need to look at the US temperature record:


When you do that, you see these temperature deviations and rankings (Coldest year since 1895 being '1') for the winter months of 2014 for NJ and Georgia:

New Jersey:
Jan: -3.6, 22nd
Feb: -1.0, 44th
Mar: -3.9, 18th

Jan: -6.0, 6th
Feb: +1.3, 76th
Mar: -2.5, 29th

The largest deviation of these two states for the winter months of 2014 was January in Georgia, ranking 6th coldest.
Please note here that the five Januaries with stronger Arctic outbreaks all were prior to 1980..

Indiana was near the center of the Arctic outbreaks and had Jan, Feb, and March ranking 10th, 13th, and 13th coldest.

Please do some more research on the actual temperatures yourself, but all you will find is that the past winter was cold in the eastern US, but it can only be considered really cold on the backdrop of many warm winters in the past 10-15 years..

John Christensen

Michigan March 2014 was fifth coldest on record and not far from record cold. You find the four most severe Arctic outbreaks during the month of March in Michigan in the period of 1899-1960, where severe Arctic outbreaks in general were much more frequent.

The argument from Dr. Francis as has previously been discussed, seems still to be theoretical and not yet reflected in actual climate data, but if you believe so please show this.

Susan Anderson

I've put my response on the NASA video article, sorry, hope some of the kind instructors and relevant commentators will take a look at it, rather than me cluttering by copying it here.

Chris Reynolds

John Christensen,

Hurrell, 2000, Climate: North Atlantic and Arctic Oscillation (NAO/AO). Extract from the Encyclopedia of Atmospheric Science - used in place of the peer reviewed paper I can't find at present.

"That the NAO and AO reflect essentially the same mode of tropospheric variability is emphasized by the fact that their time series are nearly identical, with differences depending mostly on the details of the analysis procedure."

Or checking in my Favourites I find a link to this page:
"It follows that the NAO and AO are synonyms: they are different names for the same variability, not different patterns of variability. The difference between the terms is in whether that variability is interpreted as a regional pattern controlled by Atlantic sector processes or as an annular mode whose strongest teleconnections lie in the Atlantic sector."

Figure 1 on that page is an amusing summary of the relationship.

I still find that you are both off track with regards to the influence and importance of the NAO.

I'm not negating the influence and importance of the NAO. I'm just saying it's a rough indicator that helps compare melting seasons, but the devil is in the details.

Susan Anderson

slightly OT, but fascinating stuff about Bandarbunga here:




Perhaps one of you clever people can post the animation above (penultimate).

John Christensen

Hi Chris,

It is correct that the NAO is a subset of the AO - the NAO is simply the Atlantic component of the AO.

I am interested in the NAO, since this oscillation shows in one mode (negative) an increase in Atlantic moisture moving north into the Arctic region, while in the opposite mode (positive) less moisture moving north from the Atlantic.

It is interesting to analyze if this transport of Atlantic moisture impacts the Arctic sea ice and SST during summer months.

Since I specifically am interested in analyzing a hypothesis on the transport of Atlantic moisture and the NAO is just a subcomponent of the overall AO, which includes the entire Northern Hemisphere, why on Earth would it have relevance to use AO numbers and not NAO numbers for the analysis of Atlantic moisture transport???

Chris Reynolds


If you're right (the 'if' is not meant to cast doubt, it just means I don't know), then you use the appropriate tool for the job. And in your case the NAO seems appropriate.


Sorry for the late reply. I meant to post a follow-up earlier, but I've been quite sick for the past few days and hadn't logged on to my computer at all until today. Yes, I'm mostly better now.

Neven: Yes, your summary agrees closely with my own view of the situation.

Incident sunshine was quite high in 2014, probably even over the 2007-2012 average, although below 2012. Certainly, looking at the pressure pattern over the Arctic this summer, I would have predicted much stronger melt than actually occurred. During this summer, the pressure over Greenland was slightly lower than the 2007-2012 average, and the pressure over the Laptev Sea coast also was slightly lower. The pressure over the rest of the Russian coast was, however much higher. In other words, we had much lower heat transport from land into the sea ice near the coast than in other recent years.

NH snow cover in May, although well below the longer-term average was the highest in the past 5 years, indicating lower heat transport due to lower land temperatures as well as due to weaker winds.

Despite the high amount of sunshine, albedo decrease occurred later than in most other recent years. In the Beaufort at least, CRREL buoys showed unusually deep snow cover over the ice at the start of the melt season, which would be a different reason of this. In the Beaufort and Chukchi, there was an unusual amount of multiyear ice close to land, which would lead to an initially slower melt.

Looking at Pistone, Eisenman and Ramanathan 2014, and in particular their supplemental figure 6 has significantly changed my view of the importance of open water early in the melt season. It shows an extremely close same-month correlation between albedo decrease over the Arctic Ocean and lack of sea ice, without a residual correlation with year, even though the geographic patterns do not match well. I take it as strong evidence that Arctic Ocean open water is a reliable predictor of general Arctic warmth and concurrent melt ponds.

One major difference of your summary from my own view is that on the Atlantic side, I think the biggest factor was the extremely anomalous amount of heat transported away from the Barents and Kara sectors over the winter. With persistent southerly winds over the Barents, I see these sectors as having gained more heat than the recent norm this summer, although not enough to make up for the heat the lost over the winter.

This summer, the sea ice has generally been blown into the Barents Sea as much as through the Fram strait. For example, take a look at this current drift track for buoy 971540. When have you ever seen a buoy go from near the middle of the entrance of Fram straight around the east side of Svarlbard like that?

Generally drift speed was fast over the winter, slow in June and July and fast in August, but generally slow for the summer, as we would expect for the given JJA pressure pattern. This certainly contributed to slower growth of open water away from the Atlantic early in the melt season, when much of the open water is normally caused by advection of ice.

John Christensen

Thanks Chris!

I would be interested in knowing what the DMI ice team thinks of the possible impact of a strong NAO index, so may try to reach to them.


Thank you very much for that data, Chris. It's been fascinating playing around with it.

I've been comparing the JJA averages of the AO, NAO, and your 2007-2012 average "Summer Pattern". It's striking how much the recent data is completely different from the prior period. We all know it it's been different, but it's strking how much the statistics mathematically confirm that it looks like two different data sets spliced together, especially for 1-year autocorrelation of the NAO and "Summer pattern".

"Summer Pattern" numbers for the 2007-2012 defintional period should be taken with somewhat of a grain of salt, since the 6-year-average definition guarantees that random noise would produce somewhat of a correlation with a function which is 1 only over the 2007-2012 period. Nevertheless, the increase in correlation with NAO is striking, with nearly no correlation prior to 2007 other than that which would be expected from their mutual correlation with AO, but with this switching to being a stronger correlation than with AO over the 2007+ or 2000+ periods. (NAO/"Summer Pattern" correlation 1979-2006: -0.658 2000-2014: -0.870. AO/"Summer Pattern" correlation 1979-2006: -0.851 2000-2014:-0.804)

Correlation of "Summer Pattern" correlation with the AO+NAO linear combination of maximum correlation is equally striking. (Yes, I am aware that, these being linear correlations, I really should be taking the correlation matrix and inverting it, and I may do that also and compare the results, but the results be similar.) It shows the best combination as being virtually all AO (3.5% NAO variance by CPC normalization) prior to 2007, with -0.879 covariance), but with 68% of combined variance of the best-fit function relative to CPC normalization being NAO in the 2000-2014 period (-0.924 covariance!).

Yes, this allows 1 more variable of spurious correlation, but over a 15 year time-period a >92% covariance still shows extreme predictabilty from given AO and NAO numbers.

I must admit to being somewhat surprised that it appears no increased year-to-year AO/NAO covariance is yet emerging, which would appear to agree with your statements.

Since this thread is now rather stale, and this discussion is already rather technical for this board even without a larger data deluge, I plan on posting more complete data to the Sea Ice Forum, where I currently have no account.

The comments to this entry are closed.