UAH Arctic Temperature Profile

Guest Essay by Kip Hansen


Arctic_alone_featuredLast week Dr. Roy Spencer treated us to the latest UAH Global Temperature Update.  Overall, the ”global average lower tropospheric temperature (LT) anomaly for October, 2018 was +0.22 deg. C, up a little from +0.14 deg. C in September”.

Dr. Spencer was kind enough to include in his post, as he usually does, a chart with the actual figures from his ongoing research.  The entire post was mirrored here at WUWT.

Here’s the part that I found interesting, which only can be seen if one graphs the data from this chart:

Various regional LT departures from the 30-year (1981-2010) average for the last 22 months are:

2017 01 +0.33 +0.32 +0.34 +0.10 +0.28 +0.95 +1.22
2017 02 +0.39 +0.58 +0.20 +0.08 +2.16 +1.33 +0.21
2017 03 +0.23 +0.37 +0.09 +0.06 +1.22 +1.24 +0.98
2017 04 +0.28 +0.29 +0.26 +0.22 +0.90 +0.23 +0.40
2017 05 +0.45 +0.40 +0.49 +0.41 +0.11 +0.21 +0.06
2017 06 +0.22 +0.34 +0.10 +0.40 +0.51 +0.10 +0.34
2017 07 +0.29 +0.31 +0.28 +0.51 +0.61 -0.27 +1.03
2017 08 +0.41 +0.41 +0.42 +0.47 -0.54 +0.49 +0.78
2017 09 +0.55 +0.52 +0.57 +0.54 +0.30 +1.06 +0.60
2017 10 +0.63 +0.67 +0.60 +0.47 +1.22 +0.83 +0.86
2017 11 +0.36 +0.34 +0.38 +0.27 +1.36 +0.68 -0.12
2017 12 +0.42 +0.50 +0.33 +0.26 +0.45 +1.37 +0.36
2018 01 +0.26 +0.46 +0.06 -0.11 +0.59 +1.36 +0.42
2018 02 +0.20 +0.25 +0.16 +0.03 +0.92 +1.19 +0.18
2018 03 +0.25 +0.40 +0.10 +0.07 -0.32 -0.33 +0.59
2018 04 +0.21 +0.31 +0.11 -0.12 -0.00 +1.02 +0.69
2018 05 +0.18 +0.41 -0.05 +0.03 +1.93 +0.18 -0.39
2018 06 +0.21 +0.38 +0.04 +0.12 +1.19 +0.83 -0.55
2018 07 +0.32 +0.43 +0.21 +0.29 +0.51 +0.30 +1.37
2018 08 +0.19 +0.22 +0.17 +0.12 +0.06 +0.09 +0.26
2018 09 +0.14 +0.15 +0.14 +0.24 +0.88 +0.21 +0.19
2018 10 +0.22 +0.31 +0.12 +0.34 +0.25 +1.11 +0.38

The graph looks like this:


Sharp eyes will notice that I have not used all the data — I graph only Global, Northern Hemisphere, Southern Hemisphere, Tropics and the Arctic.  (Leaving out US Contiguous 48 States and Australia — which I did not need and they only added clutter — each being already represented in their respective hemispheres.)

It is an interesting looking graph from several viewpoints.  Here’s what I find so interesting:

  1. There are four traces that form a cluster across the graph, being very close to one another in a band about 0.5 °C in width: Global, N and S Hemispheres and the Tropics.  The temperature anomalies from the long-term means are generally in step, but not lock-step, between regions.
  1. Except one: The Arctic.   The Arctic trace is very different from the other four.

Here is the Arctic annotated with the seasons:


The Arctic shows more variability, both quantitatively and qualitatively.

The two Arctic winters are far more anomalous than the two summers.  To me, the Arctic trace looks a bit chaotic with a seasonal overlay.  Between January and February 2018, there is a shift of 1.6 °C in the anomaly.

NOTE:  UAH’s “Arctic” is really denoted “NoPol” — North Pole — and is defined as 60N – 90N. It is not DMI’s “above 80N” nor is it “within the Arctic Circle”. It is larger than both.

The Danes have been treating us to Arctic temperature comparisons for many years.  The three years covered by the UAH graphs above look like this when overlaid on one another:


We see that the Danish Meteorological Institute has calculated the average temperatures above the 80th parallel (yes, it is a model result) and we see that Arctic temperatures have been a lot less cold than the long-term average — 10-15 °C less cold.  Even at that, Winters are running 20 degrees below freezing and Falls about 15 degrees below freezing.  The Summers, however, have not been anomalously warmer.  Summers show about 100 days of temperatures above freezing — and that by only a degree or so (never breaking above 275K — 0 °C = 273.15K)

This (painstakingly created) animation shows the DMI above 80N from 1970 thru Oct 2018.  Images sourced from  DMI’s Arctic Temperatures page.


The DMI data is in agreement with the UAH data, at least in a qualitative sense, in the last three years of the animation.  It takes a good eye to see that nothing really changes much until after 2005, when there is an oddity, then after 2010 things change even more.

The Northern Hemisphere, in the UAH data, taken as a whole, does not show this type of variation.

Now, north of the 80th parallel is a very small portion of the planet but “the Arctic”–  defined as the area inside the Arctic Circle at 66.5N — is quite a bit larger. UAH’s “NoPol” is defined as 60N-90N, is larger yet.


Both are part of the Northern Hemisphere.

For comparison, here is the Arctic Sea Ice extent long term average laid over the Arctic Circle.  The 1981-2010 average is outlined in red (I think).  You see that the 2016 and 2017 Maximum extents just about fill the long term average, with some empty space around the edges in the Bering Sea and the area north of Scandinavia.


Why show you Sea Ice Extent? — just to show that the UAH temperature high winter anomaly (north of the 80th parallel) isn’t caused by a lack of sea ice — almost all of that area is covered in sea ice in the dead of Winter.

Here is one last set of graphs, again from the DMI:


These graphs start in 1960 — about the middle of the 1945-1975 cooling period.  Globally, temperatures start to rise again in about 1980 but NOT Arctic temperatures.  DMI’s Arctic temperature (above 80 ° North) are steady, if variable, right around the long term (1958-2002) mean.  It is not until 2005 that anything exceptional begins to be seen.

So, what does that leave?  That’s what I’d like to know.  Here’s Dr. Spencer’s UAH Lower Troposphere Global (top — marked with a blue line at 2005) and that regional three year graph (bottom):


One last one, really this time.  This is UAH Arctic (UAH’s “NoPol 60N-90N” which is spatially considerably more than DMI’s “above 80N” and more area than “above the Arctic circle”):


Since the beginning of 2016 (highest spike in Arctic blue on the left), Global seems to follow the Arctic signal (which is 60N-90N) and has the same profile.  The last data point is Oct 2018. [Note:  The reference period for this graph is 1981-2010]

Here is my list of questions:  (I have no answers — and I hope the readers here can shed some light on the matter)

  1. Why is the Lower Troposphere Temperature in that circle at the top of the world, 60N-90N, behaving so differently than the rest of the world ?
  1. How much does that odd behavior affect the global record?
  1. The DMI modelled Arctic Temperature, for north of the 80th parallel, also shows anomalously warm winters and springs, seemingly confirming that there is something going on, but only since 2005.  Why is that?
  1. How is it that the DMI above 80N seasonal graphs show seasonal anomalies from 5-8 degrees, but UAH Arctic Lower Trop shows less than 1.5 as an extreme? Is there some physical measurement error in the DMI figures since 2005?  Was there some change in the measurement or model?  Or is there something physical happening (sea ice doesn’t change in the same period)?

# # # # #


Author’s Comment Policy:

Readers can consider this an Open Thread for comments on ALL MATTERS ARCTIC.

I thought I would find something interesting — and I did, but not answers.  Not to worry, good questions are always more valuable than good answers.

I know Arctic warming is predicted by the AGW Hypothesis — but it only substantially appears after 2005.

All collegial comments are welcome  — please discuss, not argue.

# # # # #

237 thoughts on “UAH Arctic Temperature Profile

  1. Well, Kip, looks like you’re a Detective in your spare time. Arctic temperatures change after 2005 but sea ice doesn’t? Looks like the Butler (Professor Mann et al?) did it, changed the temperature reporting system for the Arctic, that is. The Arctic warming up does not strike me as anything negative, especially since our friends the Polar Bears are doing very well.

    • Ron ==> I’m sure convinced that less dangerously cold winters above 80N is a bad thing either. But it is odd — the timing is odd.

    • (1)
      Is there any good reason
      to trust temperature numbers
      “above 80N”, even from
      weather satellite data ?

      What difference does it make
      if the Arctic temperature averages
      a degree warmer or colder
      in one year, versus another year?

      It’s always cold there !

      I suggest the sea ice extent
      is a more useful indicator
      of the Arctic temperature,
      and ocean tide changes,
      than temperature numbers
      of questionable accuracy.

      Since very few people
      live in the Arctic,
      why is the temperature there
      important to know?

      Since Arctic ice is floating ice,
      that won’t raise sea level
      when it melts,
      why is the sea ice extent
      important to know?

      A degree or two of warming
      of the average GLOBAL temperature
      over a century or two,
      is also not important.

      What’s important
      is knowing when
      we are going to exit
      our very pleasant interglacial,
      and face a colder
      and colder climate.

      Perhaps humans
      are making the planet
      a little warmer
      than it would otherwise be,
      with greenhouse gases.

      If that’s true,
      then we are delaying
      the very offensive
      cold climate ahead,
      that will be caused
      by planetary geometry.

      And delaying the
      inevitable colder climate,
      seems like a smart thing to do
      for our children !

      A little warming in the Arctic,
      especially warmer nights,
      is not bad news at all
      — it is unimportant news.

      Even the polar bears
      would not care.

      My climate science blog:

    • Mike ==> Good question. Cowtan and Way (2013) fooled around with arctic data but not sure if they are responsible. The linked essay (mine) says: “In 2005, 2006, and the first part of 2007 there seems to be some discernible differences, still in the one or two hundredths of a degree range, which disappear for a year or more beginning what looks like mid-2007 and lasts until 2009 (the right edge of the yellow box). There is something definitely different in the last few years of the data.”

      So, if DMI uses Cowtan and Way’s reanalysis method, that may be what we are seeing…

      Cowtan and Way would not affect UAH NoPol.

      • Even though the climate authorities of the Danish government are to the most part trustworthy, they still are under tremendous pressure from people like Mann et al to adhere to the global warming meme. The only temperature data that both sides trust is the UAH data. That still leaves the question of why the Arctic temperature anomalies are higher than the rest of the globe?

        • Apparently the UAH satellites cannot cover the globe above 85 degrees in latitude for both poles. This could have some explanation for Arctic temperature anomalies being higher but for the area of 85 to 60 degrees it is still a puzzle.

          • No satellite covers the “Pole Hole” above 85 degrees, and there are no weather stations that far north either, except possibly a very few automatic ones on ice floes.

          • The area “above” 85ºN is only 2.84% of the area above 60ºN

            It becomes almost irrelevant within the possible measurement errors.

        • Hi Mosh

          notice they use – “The mapping step makes use of the optimal interpolation algorithm known as kriging”

          wonder where that idea came from?

      • Kip,
        Late to the party here, but I noticed that DMI clearly indicates that the split in the data is around 2002, although you note the large divergence post 2005. I suspect the trend change post 2002 is due to the change in the “hindcasting” reanalysis they did pre-2002 vs the current ECWMP analysis done post 2002. In addition, post 2002 they ECWMP had some “updates”/step changes in how the model was run, resulting in possible step changes in the output with relation to 80N+ temperature assumptions since 2002.

        The real question here is what was done on the hindcasting that may not be done now — ie. validation and correction of erroneous observations — and what changes have been made to the ECWMP since 2002 which may have biased 80N+ temperatures. Answers to these two questions may resolve the clear change in trend since the early 2000s.


  2. And this is why the flat line for the post 1960 arctic tree ring data on warming was correct when it was flat while the instrumental data showed a sharp rise in temperature and was also correct. The growing season for the trees is just the summers (which are not warming) while the temperature measurements are year round. The instrumental data should never have been tacked onto the paleo data.

    • Sean ==> “instrumental data should never have been tacked onto the paleo data.” — and it never should be tacked on. Ever….

  3. It might be worth looking at oil spill into the Arctic ocean. I know a lot of spilled oil comes down the big rivers, but when it started I have no idea.


  4. Arctic area has few weather stations. So a lot of temperatures are ‘estimated by extrapolation’. This fact may account for the ‘high’ temperatures. Common sense says that a temperature of -30 instead of -40 is no big deal. Both are well below freezing. So an ‘anomaly’ of 10 degrees should be taken with a sackful of salt, considering the method of ‘estimation’ when direct measurements are lacking. Possibly these have been juiced up since 2005. Remember satellite measurements, RSS, are only upto 60 degrees N.

    • “Remember satellite measurements, RSS, are only up to 60 degrees N.”

      Why is that, since the satellites are in polar orbits?

      • Tom, I am not sure it is the same reason, but a lot of imaging satellites in polar-offset orbits also don’t report data for polar regions. Their problem is a lack of receiving dishes such that they can’t store the digital data in memory long enough to come into range of a receiving station, so they are programmed to not collect the data.

      • Weather satellites are not in true polar orbits. They are in Earth-synchronous orbits so as to pass over a given site at the same time every day. Data would be extremely difficult to use otherwise.
        The “best” possible earth-synchronous orbit has an inclination of 98 degrees which means that it never comes north or south of 82 degrees latitude. By making off-nadir (slant) observations it is possible to get data up to about 85 degrees latitude, but no further.

        • That’s good enough for government-funded climate science work: They’ll just krig their way to the Pole from there. If only Scott of the Antarctic had known that trick in 1911 then he could have turned back weeks earlier and claimed to have beaten Amundsen to the South Pole.

        • “earth-synchronous orbit has an inclination of 98 degrees which means that it never comes north or south of 82 degrees latitude”: This is unclear; my first reading was that you meant the satellite would always be north of 82 degrees north. But of course it’s not physically possible for an orbit around the Earth to not cross the equator. Perhaps you meant it never goes north of 82 degrees north, nor south of 82 degrees south latitude? The second sentence of that para seems to imply this is what you meant.

          • mscwell ==> I think he is just pointing out that these satellites, that UAH depends on, don’t really measure temps at the very tip top of the globe, above 85N or so.

    • The summer cleaves very closely to the long term average. I think it’s because of all the melting ice. The heat that it takes to melt ice is about eighty times as much as it takes to warm water up 1°C. As long as there’s ice to melt, the temperature can’t change much.

      • Bernie and Commie Bob ==> We are talking air temperature here. I have walked over snowy frozen rivers in the spring with air temps in the 60s F.

        Sea ice might affect air temp at 2 meters, but probably not Lower Troposphere.

        • OK, so what’s your explanation of why the arctic temperature sticks so close to the long term average during the melt season? You can go all the way back to the 1950s and there’s not much difference.

          There’s a huge difference between a snowy frozen river and a whole ocean.

          • The ice works as an enormous thermostate. It takes a .lot of heat to melt ice (333 KJ/kg to be exact), so until all ice has melted the temperature will only be just above zero.

            Water works the other way around in autumn. Until the sea has frozen temperature won’t sink much below 0 C. In ice-covered area it drops quickly once the sun is gone.

            This is well known to anyone with experience of the Arctic. It can be quite sunny and nice inland in summer, but on the coast it is always cold and often foggy.

          • you are looking at DMI, which is ECMWF, a weather model
            it would keep the temp about melting ice at the temperature you see.

          • Yes indeed. ECMWF models actual weather short-term, and fairly well as a matter of fact. Therefore it is not as trustworthy as a “true” GCM.

            Is that what you are implying?

        • I guess it’s pretty windy up there so there would be good turbulent mixing of air. So the constant temperature of the phase change would predominate every summer. 2005 changes? No idea.

      • But isn’t the melting due mainly to warmer water entering the Arctic region. If that is the case then how could melting ice have all that much effect upon air temperatures?

        • The ice melts from below long before melt water is visible on the surface. That occurs when the air temperature is well below freezing. In that case, the air temperature determines the thickness of the ice. In other words, the heat transported in by ocean currents can’t be ignored.

          Having said the above, there is much more heat that comes in via the atmosphere. My source for that is a set of illustrations for a geography course at the U of Colorado. Sadly, it isn’t always available and it is currently a dead link. Fortunately, there is a version captured by the Wayback Machine. Ocean heat transport is 3 watts per square meter. Atmospheric heat transport is 84 watts per square meter. Radiation is net negative over the year. The slide is called “Annual Energy Budget of the Arctic Ocean”. It points out that the numbers don’t balance so there’s lots we don’t know about the Arctic Energy Budget.

          The other thing to note is that ice dramatically changes the local climate. On the north shore of Lake Superior, the climate is moderated by the lake until it freezes over. That will also hold for the Arctic as well.

          It’s dramatic how close the air temperature sticks to the average in melt season. Most of the energy it takes to melt the ice comes from the atmosphere. Refer to the slide “July Energy Budget of the Arctic Ocean”. The atmosphere supplies 100 watts per square meter to the ocean and the solar contribution is a puny 10 watts per square meter. It seems pretty reasonable to conclude that melting sea ice regulates the air temperature.

          • “The ice melts from below long before melt water is visible on the surface. ”

            Not necessarily. In some parts of the Arctic it is the other way around. The meltwater on the surface seeps down and refreezes under the ice when it comes into contact with the subzero salt water. This was already shown by Nansen on the Fram Expedition in 1893-96.

    • 1. We know its warming in the winter, AS PREDICTED.
      2. Summer temps ( at 2m) will see more increase…… After the ice melts.
      the excess heat goes into melt ice limiting the increase in air temps..
      so yes, during the summer you can expect that temperatures will not be going up until
      ice melts

  5. Re-read the Svensmark hypothesis on the role of cosmic rays in cloud development. It is well established that during the night, temperatures are higher if it is cloudy. This is true everywhere since the clouds prevent the escape of heat when input from the sun is absent, especially in the polar regions where the nights are very long, i.e. 24 hours in the winter. Svensmark clearly predicts warmer winter temperatures in the polar regions with increased cosmic rays in the lower atmosphere causing increased cloudiness and with the decreasing sunspot activity, measured cosmic rays in the lower atmosphere are increasing. The finding of increased winter temperatures fits in well with the cosmoclimatology theory.

  6. I’m no expert, but couldn’t some of the apparent variability be due to the fact that there are fewer temperature recording stations, and the satellite measurements don’t work well at the poles? Couldn’t it just be an artifact?

    • This was my first thought. But why would the satellite have a hard time in the winter but not in the summer? So my effort (if I had time to look into it) would start by looking into how the satellite makes and records its measurements, and how the surface temperature data is collected (created?).

      My second thought is the air temperature is a result of the extent of ice cover, not vice versa. Isn’t there already a scatter-plot of surface air temperatures (not SST as in water temperatures, I mean air) vs. ice coverage? And not temperatures extrapolated from a reading in St. Louis, I mean real measurements. I think I just answered my own question, because we know there are almost no permanent weather stations over the oceans. And I think that may explain the difference between satellite and surface, but what about the rising non-summer anomaly measured by satellite?

      I should have switched to keyboard to type this, instead of this dumb smartphone.

      • Bruce and Red94 ==> Well, it could be an artifact, but that is less likely as it (something) shows both in the DMI modelled “above 80-N” data set and in the UAH satellite data set.

        • Kip, I can’t tell from what’s been posted so far, doesn’t the Northern Hemisphere data set and line in the graph include the data points that make up the North Pole line and graph? If that’s true, then would a graph of the Northern Hemisphere data points outside the North Pole appear even more flatline? Of, if it has already been subtracted out, then a graph of ALL Northern Hemisphere would have a bit more variability? I think that would be an interesting comparison.

          • that is a good point red viper. we would need to see the northern hemisphere data with the arctic circle component removed.

      • In addition it is winter when the troposphere is at its thinnest. Does the measurement process process this gradient or zone of measurement consistently.
        The Tropics conversely are much thicker and summer winter consistent.

  7. So what we are left with is winters always below freezing, last few years Summers same or colder than 1958 on the DMI site.

    Hmm, why the melt-

    “Winds from Siberia reduce Arctic sea ice cover, Norwegian researchers find”

    The researchers began to analyze the circulation patterns over the Arctic.

    “We found that these patterns can explain in large part why the ice cover decreased so much more rapidly after 2000. Wind patterns depend on the position of major high-pressure and low-pressure systems. We discovered that months with very little ice cover and high temperatures corresponded with crucial variations in the wind patterns,” explains Mr Sorteberg”

    • richard ==> Thanks for the link. Here we are not talking about sea ice but the temperature profile in the UAH North Pole data — oddly anomalous compared to the rest of the Northern Hemisphere and rest of world.

      • Kip, as I noted already, isn’t the air temperature related to – a result of – the amount of ice coverage? I would like to see a scatter plot of average air temperature for this “North Pole” region vs the amount (maybe normalized to percent coverage?) of sea ice for the same region. And done at different time lags, up to a year, what the heck, maybe up to 2 years? Covering up the ~0° C (isn’t the freezing point of sea water a bit below that?) open water with ice that can be any temperature below 0° C could allow a lot more variability of the air temperature, but it would probably have a time lag, don’t you think?

        I did already see you comment about frozen rivers in 60° F air temperatures, but I think that’s a lot different, when you have land mass all around that river, than in the open ocean where you have nothing but water around that temperature reading.

    • As a matter of fact the sea-ice has not decreased at all for more than a decade since 2007, so that analysis had better bet taken with a grain of salt.

  8. Is there any way to seek an appraisal from DMI? Their opinion would be valuable. Does Dr. Spencer have additional thoughts? It would be nice to hear from those directly involved.

  9. My first answer to why the Arctic is different is that the Arctic (north of 66.5N) is a mere 4% of the globe. As a result, it will be subject to much wider variations than say the Tropics (23.5N – 23.5S), which is about 40% of the globe.

    My second answer is that unlike the rest of the world including Antarctica, the Arctic is mostly ice-covered water. As a result, it is subject to very different conditions than the rest of the planet. For example, warmer water can come up from the Pacific and push under the ice, leading to melting from below.

    My third answer is that the MSU satellite doesn’t cover the area from 85N to the pole. And they define the “Arctic”, curiously, as 60N – 85N (6.5% of the global surface).

    As to the DMI, it’s models all the way down, so there’s no telling what they’ll come up with.

    Thanks for an interesting post,


    • Willis – regarding Suzanne Morstad’s question above – is there CERES data going for enough back to allow us to compare cloudiness in the Arctic with winter-spring Arctic temperatures?

    • w. ==> Yes, the DMI area is above 80N (very tiny — and modeled). UAH uses 60N-90N for North Pole, larger than the Arctic circle by a bit — it it is this data set that shows the hinky behavior of larger swings and dips.

      All your ideas are sound — but no smoking gun (I didn’t find one either…).

  10. Hans Ahlmann has the same observation:

    The winter temperature at Svalbard on the rim of 80 Deg N is very unstable.
    It goes from -3,5 (1909) to +4,5 (1932).
    That is not CO2 but H2O.

    The rest of the season he does not see the same variation.
    The green curve is due to no sun over the horizon and the “heating” (red curve) is due to water in the air hiding the black sky!

    • Lasse ==> Terrific link, thank you. The data for West Spitsbergen shows similar temperature anomalies in the Winters in the 1930s.

      • Kip, it also includes the meandering Northern Jet Stream and the wandering Polar Vortex.
        As we have seen recenyly when the Polar Vortex slips “South” warm air moves in to replace it.
        That leads to quite large swings in Polar Temperatures.

  11. Oh, I forgot to mention one last oddity. The UAH category “Extratropics” includes the Arctic and the Antarctic. From their “ReadMe” file:

    Sectional definitions for uahncdc.XX files

    Global -85 to +85 latitude
    Hemispheric 0 to +/- 85 latitude
    Extratropics +/- 20 to +/- 85 latitude
    Polar +/- 60 to +/- 85 latitude


    • w. ==> Quite right. However, the data set for UAH version 6, 1978 to present, is at

      and has a definitional note at the end:

      “GL 90S-90N, NH 0-90N, SH 90S-0, TRPCS 20S-20N
      NoExt 20N-90N, SoExt 90S-20S, NoPol 60N-90N, SoPol 90S-60S”

      It is correct the the MSU satellites cover only up (and down) to 85 degrees, leaving the tiny circle 85N-90N uncovered. For UAH NoPol (nominally 60N-90N) it probably doesn’t make much difference. The UAH NoPol figure probably does not really go all the way to 90N.

      • above 85ºN is only 2.8% of above 60ºN,

        It will make very little difference within the range of measurement errors.

      • What does make a difference is including Greenland which because of its high altitude is known to cause the MSU/AMSU data to be noisy and should be excluded from any analysis. Spencer has discussed this.

  12. Kip, it was pointed out acouple of years after that time that we were in a “Pause”which was admitted to by Trenberth and Jones in climategate emails and after the email release. IN the fillowing years In 2007, Hansen “revised” the 40yr cooling period after the mid 40s, first in the United States and then afterwards globally to give it a slightly rising temperature (pauses were forboten!).

    The Dreaded Pause was the most catastrophic event warming proponents had faced. They were already engaged in a campaign of “hottest climate ever”, hockey sticks, etc. when nature threw her curve ball. An unknown number of climate scientists were afflicted by the well publicized, so-called “Climate Blues” which was terminal to their careers. The more political types undertook egregious revision of the climate record, tellingly, both Hansen and Karl, the worst, did it on the eve of their retirements!

    The deep cooling before 1980 was removed and the 1930s 40s pushed down because, otherwise, almost all of the warming would have occurred by the late 30s and the two bigest ‘pauses’ occurred since 1950, when CO2 was deemed to have only begun to have an effect. More recently they shoved the 1950 starting point back to 1850 to bankroll half their 1.5C dangerous rise by 2100 (to double the warmibg of the past century). They didnt want pauses longer than the warming they were worrying about.

    • That is why the UAH data set is the only one that both sides trust and will be the final nail in the CO2 scam.

  13. Almost all “global warming” has been in the Norther Hemisphere, mostly in the Arctic, mostly in Arctic winters. In other words, “global warming” isn’t global. This fact, alone, is enough to show that CO2 is not the primary cause because CO2 is well mixed in the global atmosphere. UAH data show almost no warming at the South Pole but CO2 concentrations are essentially the same as the North Pole.

    I suspect Arctic warming is mostly due to cyclical movement of warm air and water from the North Atlantic into the Arctic. I suspect the Atlantic ocean rather than the Pacific ocean because the Norwegian sea is wide and deep while the Bering Strait is narrow and shallow.

    There is a near-perfect correlation between the Atlantic Multi decadal Oscillation and global temperature but the record is too short to draw any firm conclusions.

  14. The Arctic is what? 4% of the total world surface and it seems that both sides of the Climate Change/Global Warming discussion obsess over it. My general question is why? Or so what? Why should I care what the sea ice is doing?

    I think a lot of it has to do with perception drawn from the distortion of world maps produced by the Mercator projection of the world map.

    We are told the Greenland ice cap is losing mass. Considering that calving of icebergs is a temporally delayed function of snow fall I wonder how much is due to temperature. Isn’t Greenland way below freezing nearly everywhere nearly all of the time? And since the IPCC tells us:

    Globally averaged mean water vapour, evaporation
    and precipitation are projected to increase.

    IPCC AR4 Chapter ten Executive Summary

    Wouldn’t warmer winters produce more snow, not less?

  15. Until we have 100-200 years of instrumental data for the Arctic (and globally) it’s impossible to say what is normal and what is not normal. Even then, it would not be straightforward. Apart from just measuring it’s fractal dimension maybe, to confirm the obvious – that climate variation is nonlinear-chaotic and thus has the log-log fractal signature.

  16. Kip
    I am not sure if I explained this to you before. I had a similar problem with my own results trying to understand how it could be that by my own estimates [from my own results], temperatures here in SA, and the SH in general, were actually going down [especially looking at Tmin] whereas in the NH they were still rising.
    [although on average it seems Tmin global is dropping]/
    If you click on my name you can read some of my results in my final report on this.
    It did take me some time but I think I finally figured it out. Come down 1 or 2 km into a goldmine here and notice the sweat soon pouring from your face. Meet the elephant in the room: earth’s inner core and he has been moving, north east to be exact, much faster this past 100 years than the century before.
    Apart from the solar variation in irradiation [as per the relevant solar cycles] it seems there is also a variation in the alignments of earth’s inner core with that of the sun’s inner core, both mostly consisting of [molten] iron: the Eddy cycle. This explains the settlements in Greenland by the Vikings that only now are becoming visible. One of my forefathers, Willem Barentz, was convinced of a passage to the east via the north. He must have heard this from stories in Norway. Unfortunately he and most crew died trying to find it. Hence, the “Barentz” sea.

    Funny, to realize that he died for the benefit, of what we cry for now, not wanting it [for fear of global warming] ?

    Anyway, you will realize that that a change in the position of earth’s inner core does not affect summer T that much as it is more controlled by the sun but it will affect winter T a lot when there is no sun at all [above 60 lats.. It is simply earth warming things up from the inside to the top and that one is changing now that the inside of earth has shifted.
    I welcome all of your comments on this.

  17. When the jet streams are more meridional there are more incursions of warm air and cloud into the Arctic.
    They are more meridional during the winter months because of the increased latitudinal temperature gradient.
    They are also more meridional when the sun is less active because solar induced changes in stratospheric ozone push the height of the tropopause down over the pole so that cold surface air breaks out equatorward more readily.

  18. I’m not sure that it makes any difference, but I don’t think UAH covers North of 85N or South of 85S. That’s really two very small areas, and maybe not worth worrying about. That’s what Wikipedia says and seems consistent with usual orbit planning practices.

    My first question would be to what extent (if any) we are looking at meanderings and ice cover of the warm Gulf Stream and West Spitzbergen Currents. I expect that someone around here knows off the top of their heads

  19. Joe Bastardi has taken note of this very phenomenon for quite some time and comments upon it from time to time in his daily/weekly update on Weatherbell Analytics. ( He has proposed a working theory he discusses there that it is a water vapor phenomenon caused by moisture introduced by warmer ocean temps esp. super El Ninos such as 1998, etc. which has created slightly more moist atmosphere even in the arctic. And, so he argues, it only takes a small amount of water vapor introduction to change the temperature behavior of the lowest temperature air – that is, winter temps in the arctic.

  20. You can also see the Arctic winter warming in the CFSR data. It is most likely due to more open water. The Arctic sea ice reached it’s current level around 2006 and that is when this started.

    With open water energy from the water is released into the atmosphere which warms it. It only makes a big difference in the winter when the average temperature of the old ice covered area was well below freezing and now it has increased to near freezing. In the Summer those areas are normally near freezing and hence little change.

    Naturally, this has a big effect on the GAST for such a small area of the planet. Some of this also affects the NH extra-tropical areas. My own assumption is this is due to the +AMO bringing warm water into the Arctic and melting the ice.

    This is the reason I only look at spring/summer temperatures to try and understand what is really happening with the climate. Here is what I found when I ignore the influence of the Arctic on climate. I use April-September removing the high and low anomalies. Also try to use only ENSO neutral non volcanic years.

    1980-81 14.4 C (58.0F) -.06C
    1990….. 14.5 C (58.1F) .01C
    1995-96 14.6 C (58.2F) .08C
    2001-02 14.7 C (58.4F) .18C
    2007….. 14.7 C (58.3F) .15C
    2014….. 14.7 C (58.4F) .18C
    2018….. 14.7 C (58.4F) .20C

    Still a warming trend from 1980-2000 but nothing since then. A total of only .26 C of warming in 4 decades. This is why I expect to see a significant cooling of global temperature data sets when the AMO goes negative and the Arctic sea ice returns.

    • Richard, I agree with your analysis. I suspect that water in its various forms (water vapor, clouds, precipitation, oceans, snow, ice, glaciers, etc) is far and away the most dominant influence on climate changes on the order of years to decades to centuries and probably even out to a few millennia. Other greenhouse gases, including CO2, are probably very minor players as are human induced changes in land albedo. I expect that most of the small effect from CO2 occurs in dry cloudless air over deserts at night, including polar deserts.

      At longer time scales from tens of thousands of years to millions of years, earth’s orbital/axial mechanics appear to be a major influence (glacial cycles) and at even larger time scales, plate tectonics begin to exert a substantial influence. And of course at very long time scales, on the order of hundreds of million years, the gradual increase in output from the sun will have major effects and will eventually boil away the oceans and burn the earth to a crisp by a few billion years.

      • Exactly. It is in dry climates where CO2 could have a large effect. Interestingly, this effect diminishes at higher altitudes meaning places like Antarctica and Greenland also see little of the potential effects.

    • Richard and Brian ==> But ….The Arctic is pretty much totally ice covered by the true Winter months — very little open water outside of the long-term norm.

      Am I missing something here?

      • Kip, winter ice extents in recent years are lower than in the 1990’s and 1980’s, and down to 60N there are large areas of exposed ocean north of Europe in the winter that would have been covered in ice.

        • Stuart, you’re right and in addition more exposed ocean without ice leads to higher water vapor in the air above leading to more clouds to be spread around by the wind, adding to the warming effect in the winter night.

  21. I think the answer is simply water vapor. How much does it take to warm 245K air 5 degrees vs air at 275K? Compare that to trying to warm 300K air in the tropics just .5 degrees.

    When you get a huge Pacific High or Low pumping water vapor into the arctic in winter, it has a much more pronounced influence in winter than during the summer. When we have primarily zonal flow winters the water vapor doesn’t make it that far north. When we have loopy jet streams (goreism), water vapor is more likely to make it there in the cold seasons.

    • rbabcock ==> But would physics just suddenly start to work in 2005? I understand the water vapor issue but that should have held for all years since at least 1980, particularly in the really low ice years.

  22. I would first check and compare the number, i.e. count, of temperature determinations per region in the UAH plots.


  23. I like looking for simple stuff first. What if all that white stuff on the sea/ground buggers up the ‘radiance’ measured by the satellite. After all its not temperatures that are being measured, they are just inferred by some BS model.

    • JimW ==> In this case we are comparing like to like — UAH to UAH and sidechecking with DMI.

      Unless UAH has done something wacky in 2005-2008 or so, there must be something other than a long-term inability to deal with snow and sea ice.

  24. The Arctic shows more variability, both quantitatively and qualitatively.

    I’ve done my thousands of hours of station data, and it occurs to me that, while the cold spots are inherently volatile, there are also the issues of much smaller area and relatively few stations. And that nearly always makes the zigs and zags bigger.

  25. Joe Bastardi has had several discussions concerning injection of water vapor into the atmosphere as a result of the super NINOs. His theory/conjecture (my words) is that the temperature dependence of the mixing ratios will affect the low temperature limit in the polar region and consequently will result in higher low temperatures. This is a non-linear function which affects higher temperatures less and is seen in near normal summer high temperatures. I’m not an expert in this field by any measure so please excuse me if I’ve somehow confused this but from a standpoint of water vapor and saturation vapor pressure this seems to make some sense. I’m sure the overall system has more complex interactions that affect the global air flow and water vapor transport. I’m not aware of any discussion of this by Joe outside Weatherbell which requires a subscription. If someone has more references to this I’d be interested to see the comments.

  26. Hi Kip……..Interesting graphs and puzzling observations. Thanks for posting this.

    Please note that there is a typo in the text/legend just below your last graph. It should read “(highest spike in Arctic Blue on the LEFT)”………not right. Cheers!

  27. Only reliable temperature in the general Arctic area are those available from just outside the Arctic circle, i.e. from Reykjavik met office. Some 6 or 7 years ago I did ten year forecast based on some other long term data and promptly forgot about it.
    I’m currently far away from my pc so unable to update and see if the forecast was a failure or otherwise. It would be great if anyone else would like to have a go, at least for the annual temps where the scale is actual temperature in degrees C, while winter one might be too, but it looks like the anomaly, so best ignored unless annual turns out to be good.

  28. I think the reason for high variability in the Arctic is storm-system-related phase change releasing latent heat, and high and -pressure centers moving warm Pacific air northward at the higher elevations, and warm/wet air masses carried northward through the gap between Greenland and Northern Europe.

    to help visualize this, below is a url for precipitable water, spherical rendering, with the north pole in the middle.,91.40,671/loc=-115.949,89.968

    The gold color is very dry air, and the blue is higher moisture.

    A few things of note:
    1) There are two strong cyclones visible, one in the Atlantic and one in the Pacific. The moisture elevated from these two storms is picked up by lesser power low pressure systems. In the Pacific, one south of China (at the top, right of center) a second just south of the Alaskan Archipelago (top, left of center). The latter of these is feeding moisture northward, and getting picked up by otherwinds and carried into the high Arctic, as seen by the blue Total Precipitable Water (TPW) content extending North of 80 degrees.

    2) The Water Vapor from Cyclone in the Atlantic is connected by winds to a low pressure system between Iceland and Greenland, which is quite close to the arctic circle.

    3) Now, put nullschool in motion. Click on the word ‘Earth’ in the lower left corner, and the controls will be displayed. Find the line labelled ‘Control’, and the double right arrow on that line. Clicking on this will advance the display by 1 day per click, and you will be able to see the daily changes in moisture pattern motion through that month.

    4) After just 1 click, the moisture from the Pacific side moves to cover the north pole, and extends all the way to the Bering Sea, shown as a top-down motion in from this perspective.

    5) With a few clicks… the blue color dissipates in this area, replaced with gold, representing very dry air. What happened to the precipitable water? It phase changed into water (rain) or snow. Surface temp at the North pole is below freezing.

    6) But, it is not below freezing at 850 hectoPascals of pressure. Using the controls, find the line marked “Height’, and click on the 850 to the right of that. Also find the line Labeled ‘Overlay’, and select ‘Temp’ See the green to the north of Alaska, the Berents and Eastern Siberia? Almost 9C.

    7) In this configuration, click a few times to advance the date. The green area (above freezing at 850 hPa pressure) extends all the way to the North Pole. What is less obvious is that the temperature is decreasing. If you click in the green area, the temp is displayed in the upper left of the screen.

    8) Advancing the date to October 8, the temp at 850 hPa is -3C.

    Through the same time period, there are a couple of cyclones pumping warm air up the East Coast of Greenland, quite visible at 850hPa, especially on the 11th-13th. By the 14th, temp at that elevation has returned to below freezing. At the surface, temps stay above freezing between Greenland and North of Europe.

  29. I suspect instrumentation differences and that the newer polar orbiters have better sensors. It’s difficult to distinguish between clouds and ice during polar winters and this was especially true for older polar orbiters. In 2005, they deployed NOAA-18 which was the first to use a significantly different and upgraded microwave sounding instrument.

      • You are wrong. Roy Spencer’s site quotes

        “We describe the major changes in processing strategy, including a new method for monthly gridpoint averaging; a new multi-channel (rather than multi-angle) method for computing the lower tropospheric (LT) temperature product; and a new empirical method for diurnal drift correction. We also show results for the mid-troposphere (“MT”, from MSU2/AMSU5), tropopause (“TP”, from MSU3/AMSU7), and lower stratosphere (“LS”, from MSU4/AMSU9). “

        • UAH measures mid-troposhere temperatures, not the surface.

          tty is correct, UAH v6 takes data from 0-14km altitude with the peak sensitivity at 4km.

      • The microwave data is used to establish atmospheric temperature profiles and this instrument went through a significant upgrade starting with NOAA-18 in 2005. It’s also not just the result of one instrument, but the combined result from many instruments. There’s also a whole lot of processing involved, which is subject to change and when it does, it’s generally not applied to older data, especially since its often specific to new sensors that aren’t available on older satellites.

  30. Possibly related to the stability of the polar vortex. Stable vortex == stable arctic air (cold)
    Unstable vortex == unstable arctic air (cold air goes south replaced by warm air from somewhere else)
    The unstable vortex is associated with solar minima.

  31. Ooops! Kip……my apologies! Please scrub my last comment regarding left/right.

    There’s no typo……only me reading your post badly on a small screen. Sorry about that!

    However, you do have a typo just above your graphics showing the latitudes encompassed by the two different measurements.

    “NoPol” is defined as 60N-980N, is larger yet…..should read “60N-90N, is larger yet” Thanks and, again, sorry about the earlier screw-up.

  32. The Arctic area has a smaller land mass than the other areas measured, so fewer data points were used in its calculation.

    Fewer data points means you are going to have more deviations.

    If you want to do an apples-to-apples comparison, you would need to get the raw underlying data and use the same number of data points from each region to calculate the regional averages.

    Otherwise, of course the areas with more data points will have more smoothing and less deviation.

    • Bart Tali ==> But UAH is a satellite data set “measuring” temperature (radiance) of the atmosphere.

      • No Kip, UAH is not a data set “measuring” temperature. UAH is the output of a MODEL that translates microwave brightness into a “temperature.”

        You think the output of a model is “data?”

        • Dave ==> Thus the quotes riund measuribg….radiance is used as a proxie for temperature and checked against other data, such as radiosone balloons,

          • Sorry Kip, the “riund measuribg” is meaningless.

            Model output is a proxy?…… mean like tree rings?

        • Just like a thermometer is a MODEL that translates expansion/contraction of a medium (or more recently changes of the electical resistivity) into temperature.

          Temperature is actually a statistical distribution of molecular movement and not directly measurable. If anything radiation temperature is less indirect.

      • RSS takes the same inputs that UAH takes, but the output of the RSS MODEL is different than the output from UAH.

  33. How does the question fit in with various oscillation changes from the mid 90’s forward (PDO, NAO etc..)? We did also reach a significant ice minimum in 2007 that could provide some logical linking relationship.

  34. Waxing sarcastic: Despite being a non polar molecule (both charge and moment being balanced by its symmetry), CO2 can be seen in recent NASA visualizations migrating to both poles in the upper atmosphere. The effect is pronounced in the arctic because both human and natural (soil) CO2 is predominantly northern hemisphere.

    Seriously: Despite being anathma around here, adiabatic effects from higher winter pressure could easily account for the temperature increases. After 2005 (possibly in conjunction with a “great Pacific” shift), this pressure could exacerbate arctic oscillation ie more meridional exzcursions.

  35. I would put it down to a lack of high pressure over the arctic region in recent years in winter if there was high pressure dominating the weather then the Arctic region would be losing a lot more heat by radiation then it is now and would be colder despite cold air being pushed out of the Arctic. This year we have the threat of high pressure becoming dominant in the Arctic region from the weather models but like last tear it is taking a long time to happen. I remember when we had warm temperatures in December in the UK recently when it rose to I believe 15 degrees centigrade maximum temperature one day, that day however everyone had to drive with headlights on at noon which shows that it is the warm air that is pushed into the Arctic that we have to consider not cold air being pushed out of the Arctic.

  36. Kip,

    Very interesting post, thank you. As you know, I have been looking at the HadSST3 data recently and have not got very far as yet, but some initial comments follow. The reason for bringing this up here is not to distract from your post, but to add an additional and possibly relevant factor. First, bearing in mind that this is sea surface anomaly data and not air temperatures, look at this graph:

    There is a clear divergence in character from early 2003 onwards, with the northern hemisphere (NH) anomolies being heavily influenced by the annual (seasonal) cycle. Prior to 2003, NH and SH anomalies tracked pretty closely. Further, the NH peaks reflect the summer months (generally August) while the troughs, which are more consistent with the SH data, reflect the winter months (generally March). So, in a sense, the exact opposite of the air temperature anomalies that Kip has highlighted. More detailed analysis is required to see if/how the NH air and sea surface anomaly data may be linked.

    Preliminary analysis indicates that the appearance of the annual cycle in the HadSST3 anomaly data is definitely linked to the more northern latitudes. There are data issues above 75N with zero or a limited number of observations in each 5×5 degree cell, but these have only a minor effect on the overall NH values due to area weighting. The more interesting area seems to be from 75N down to about 45N between June and September. Very provisionally, there is some evidence that the highest average cell anomolies (based on 5 degree latitude zones) shift southwards as the summer progresses. Incidentally, you can view the distribution of individual cells and their values on a monthly basis over the last few years here:

    If anyone has done a detailed analysis of the HadSST3 data please share your conclusions. If not, perhaps someone with more analysis expertise than me would like to have a go.

    • Jim, I have seen the HADSST3 NH vs SH time series graph posted here before. To me the sudden change in the NH time series looks very artificial, as if something drastic changed in the way the data are compiled/analysed. I can’t imagine that strong seasonality would just suddenly turn on like that in the real world. This artifact raises questions about what was changed and why, as well as how legitimate is the change.

      • Bryan,

        I agree that it looks artificial, but I have starting digging into the data at a 5×5 degree cell level (values and number of observations) and, as yet, have not found any obvious “local” busts that are sufficient to explain what we see. There certainly are some very questionable values. There is also a marked increase in data observations around that time. According to the QC procedures, anomalies greater than 8C are deleted, but that still leaves some that are in 7C range, which is stretching credibility. For example, I have found a few very high values in the middle of Hudson Bay, which may possibly be explained (I have not checked this) a consequence of ice melt combined with relatively sheltered waters (i.e. not open ocean), but I see little point in speculating at this point. A thorough analysis is necessary before drawing conclusions and it is very time consuming due to the volume of data.

        I remain suspicious, but it clearly needs explaining.

          • Kip,

            Yes, will do. Assuming that the HadSST3 data are “correct” and not a conseqence of a change in procedures as suggested by Bryan-oz4caster (big assumption as he may be right), the annual cycle range of 0.2C up to 0.5C (see 2014) would indicate a significantly increased annual temperature variation compared to that recorded in the averages of the 1961-1990 baseline data. In other words, the summer anomalies are increasing at roughly twice the rate of the winter anomalies. The problem is that these data, clearly incorporating annual cycle effects, are then merged with the SH data, completely changing the character of the global HadSST3 data (e.g. the time series response), which are then incorporated into the HadCRUT4 data.

      • Bryan,

        I have checked HadSST2 as well and it shows the same thing (it includes data up to end-2014). It is interesting to note that HadSST2 was introduced around the time of the “sudden change” (, but the same procedures would have been applied to all previous and, when available, subsequent data. I am keeping an open mind for now, in the absence of any specific information that provides a definitive explanation.

        WARNING: WFT has mis-labelled the HadSST2 datasets and shows NH as SH and vice versa. I have confirmed this against the Hadley published files.

    • V ==> ” if the N G is correct” — that is quite unlikely. NG gave up real science for activism many years ago.

  37. Why is the Lower Troposphere Temperature in that circle at the top of the world, 60N-90N, behaving so differently than the rest of the world ?

    November 10, 2018 at 8:55 am
    pointed at the “How much does it take to warm 245K air 5 degrees vs air at 275K? Compare that to trying to warm 300K air in the tropics just .5 degrees.”

    Another point is you lose much more energy at higher temperature (t**4), so at lower temperature you can have higher variations for smaller amounts of energy.
    Antarctica is more or less isolated through the circumpolar current (and included in the S Hemisphere graphic) whilst the Arctic is directly pumped from global oceans through the Gulfstream.
    There was also mention of the latent heat in the above thread – which is much more then the heat to warm air 1°.
    This for sure can create huge variations for the Arctic environment – a bit warmer or colder ocean means a lot in the balance of energy for the Arctic.

    The DMI modelled Arctic Temperature, for north of the 80th parallel, also shows anomalously warm winters and springs, seemingly confirming that there is something going on, but only since 2005. Why is that?

    If this is not an artifact of measurement (some instruments or location or whatever has been changed) could mean a change in global current circulation?
    Warmer Arctic air resulting in greater heat loss which could explain the pause.
    These just my two cents = ‘armchair specialist’ thoughts 😉

    • Dave ==> Thanks for the link to the equipment change in MSU — I’ll have to see what Spencer had to say at the time.

  38. If you look at this graph of sea ice volume
    you can see an increase in sea ice volume formation each year since about 2007. Open water sea ice formation releases latent heat at the ocean very surface hence we get an increase of surface heat and an increase in deep water formation (surface warming /deep water cooling). The very surface heat escapes to the atmosphere quickly over the winter and then is lost to space. Along the way it weakens the polar vortex which then allows the influx of more warm air from lower latitudes. This is a feature of low summer sea ice extent. The slow cooling of the arctic ocean at depth over years will enable the slow accumulation of pack ice with or without an increase of fresher water from the pacific.

      • Abstract
        The relationship between climatic parameters and the Earth’s magnetic field has been reported by many authors. However, the absence of a feasible mechanism accounting for this relationship has impeded progress in this research field. Based on the instrumental observations, we reveal the spatio-temporal relationship between the key structures in the geomagnetic field, surface air temperature and pressure fields, ozone, and the specific humidity near the tropopause. As one of the probable explanations of these correlations, we suggest the following chain of the causal relations: (1) modulation of the intensity and penetration depth of energetic particles (galactic cosmic rays (GCRs)) in the Earth’s atmosphere by the geomagnetic field; (2) the distortion of the ozone density near the tropopause under the action of GCRs; (3) the change in temperature near the tropopause due to the high absorbing capacity of ozone; (4) the adjustment of the extratropical upper tropospheric static stability and, consequently, specific humidity, to the modified tropopause temperature; and (5) the change in the surface air temperature due to the increase/decrease of the water vapor greenhouse effect.

        (PDF) Geomagnetic Field and Climate: Causal Relations with Some Atmospheric Variables. Available from: [accessed Nov 10 2018].

        • That is the line of research I was looking at. Here is an interesting paper.

          It looks at the variations of the lat/long of the north magnetic pole wrt global temps.

          Two things struck me: the meandering of the magnetic pole is often in jerks, moving small distances some years, much larger distances in others. Perhaps a large movement around 2005, combined with other changes, resulted in the sudden polar temp extremes. (Actually, the paper suggests a 25-year time lag, so maybe a big shift in 1980?)

          The second thing is that one could hypothesize a linkage between the magnetic pole shift and temperatures. Seems to me a shift in the location would shift how and where solar and cosmic radiation most impact the atmosphere. That, in turn, could impact cloud cover, and temps.

          But I am way out of my depth. Maybe someone else will pursue it.

  39. Kip
    This change since 2005 has not just been confined to Arctic temps. lt also shows up in the spring and early summer NH snow cover extents.
    Since 2005 there has been a notable decline in NH snow extents during the months of May and June ( and to a lesser extent also during April and July). The Rutger snow lab monthly graphs show this up very clearly.

    Here is a update to my first snow record. Which recorded Oct 27th 10.04am for this season 2018/19.

      • Sorry l don’t know how to provide links.
        But if you are interested take a look at Rutgers snow lab home page. Where under the “graphs” list you can click onto monthly anomalies and check each month. Where you will notice a step change in the NW snow extent from 2005 and onwards during the months May, June, and July. The change for the month of June is very noticeable.

  40. Thank you. Interesting analysis. Slicing and dicing the data can lead us to the root causes of the changes. Now I have more questions.
    – Is this just UAH and CFSR? What do other data sets say?
    – Are there differences between sea locations and in-land locations?
    – Do we have correlations with ocean currents and wind patterns?
    – and so on

  41. Kip
    The answer to your question 1 – atmosphere intrusion from troposphere mid latitudes. The inflow has increased over the period starting mid 1980’s. These intrusions can cause vortex deformation.
    Look at the chart link below, note that starting December Barrow CO2 values go flat during the winter months.
    The period of greatest CO2 output in the NH and the values stop rising and go flat. Some years have a specific shapes depicting wind speed flow. Increased air flow = dilution of CO2. That same wind is transporting heat also, and other trace gases.


    Air ingress

    • Ozonebust ==> I can’t connect a change in 1980 to the observed change not taking place until at least 2005….?

      • Kip
        Using the data in the example below, NH average temperature anomalies started moving higher about 1989 – 1990. Prior to that the NH anomalies were colder.

        Also note the immediate and identical mirror effect between the NH and SH temperatures particularly the 2005 year. This is evident most years. From that an average is derived. No-one previously has noticed the mirror effect.

        I have researched the reason for the mirror effect, one day I will do a post on it. It is quite fascinating, but not that simple. The temperature anomalies should not be seen as isolated from earths atmospheric transport systems and limitations. To understand the temperature anomalies you first must delve wider and understand the mechanics of atmospheric distribution and what the influencing and controlling factors are. This is the controlling mechanism of the anomalies and why there is an Arctic amplification. Anyone stating that earth is warming without a comprehensive explanation of the influencing factors is talking nonsense.

  42. Thoughts.
    Despite the differences, land, elevation , temperature inversions etc etc we still need to compare both ends of the world. First request is a graph of Antarctic changes available,
    If not why not and if available what does it show?

    Second one may note that at the height of this cold the DMI had shown, last 5 years, below normal or normal summer temps so no problem there.
    The temperature increases occur as soon as the temp drops a couple of degrees and recently has been staying up and spiking throughout the winter independent of the ice refreezing rate.
    As soon as it reaches the same day on the other side, warming up, it ducks under the average for the 100 days of summer.

    DMI graph is unnaturally smooth and symmetric. I am sure it does not reflect the actual data summary/average which would have to be bumpier and asymmetric. Probably we do not have the right average put in to derive our anomalies from. Natural variability is very underestimated.

    • angech ==> You have as many questions as I do ;-0

      DMI data is modelled and not measured — but has been the go-to data set for the high north for many years.

      • FFS Kip, DMI is not the “go to data”
        Maybe at WUWT

        Over the years DMI have merely picked a weather model and cut out the portion above 80N

        And they dont even area weight it.

        Its not their data, its other people data and they dont even area weight it properly.


        • “Its not their data, its other people data and they dont even area weight it properly.”
          A bit touchy, young man.
          You of all people should restrain your comments on area weighting considering the way you have twisted the data (other people’s often, by the way) in the past.
          But since you have made the comment DMI is allowed to use the data. They are scientists and they weigh it according to thought out principals.
          Now I don’t agree with them all and you don’t but that does not make their attempts any less scientifically valid.
          As to go to data, if people go to it, and they do, it is go to data, full stop.
          Try being nice to people.

  43. “The DMI modelled Arctic Temperature, for north of the 80th parallel, also shows anomalously warm winters and springs, seemingly confirming that there is something going on, but only since 2005. Why is that?”

    I guess it has something to do with wind. I have used the CERES data 2000-2016. The monthly wind speed anomaly 2000-2016 is shown here,

    the anomaly of the skin temperature is shown here.

    The green lines are obtained by a Gauss low pass filter. Obviously, there is some correlation.

      • I seen that. What I’m thinking is the satellites can see volcanoes easier than some station away from there. Right now, there are 2 minor heating events at volcanoes in Alaska. The molten exchange from below could cause that.

  44. Some possible causes
    – Delayed heat transfer effects of the 1998 Super El Nino
    – Changes in the solar Ap-index, which possible affects the vortex.

    I recall Anthony had discusions on the Ap index drop in Oct 2005.

    The fact that the fluctuation are only in winter tends to make sense if the vortex is being impacted by the low Ap index. Of course we need to wait to see what happens when the Ap returns to higher levels.

  45. “Arctic amplification” of surface temperatures is often referred to, but seldom with any serious physical insight. For starters, it should be recognized that advection of heat from lower latitudes, rather than local isolation, is the dominant driver of temperature variations in that region–especially in winter. Furthermore, given the nonlinear S-B relationship between energy and temperature, any given change in the former necessarily produces a larger change in the latter as the local regional temperatures diminishes with increasing latitude.

    What the recent arctic upsurge seems to show is that advective heating has increased, thereby decreasing the energy content concentrated at lower latitudes–all in accordance with the natural principle of maximizing entropy in a closed system. That sort of disparate behavior should be expected primarily in the latter stages of a warming episode, when the nonlinear dynamics of global advection are operating at peak level.

    • I think the phrase also gives an insight into the mind-set of the climatist. “Arctic amplification” conjures up a getting-even-worse-than-we-thought scenario in the mind of the low information reader, so the journalist class also love the phrase. From there it is a short step to the infamous “Arctic death spiral”.

      The same thinking probably also applied in the case of the water vapor feedback: Not only did it turn out to be physically necessary to invoke a large, unproven, water vapor feedback term into the models to turn a mild global warming into a catastrophic global warming, but it also strongly appealed to the modelers in the linguistic sense I have just described. Almost everybody grasps the concept of feedback quite well, certainly well enough to help envisage a runaway train going down a steep hill, which is what they wanted. Confirmation bias is not just an affliction the experimentalist, but also of the theoretician.

      • Almost everybody grasps the concept of feedback quite well

        Well-grounded skepticism notwithstanding, I fear that you still give too much credit to “climatists.” The “feedback” that they often claim is not at all feedback in any rigorous system-analysis sense of the word. Instead of referring to a physical loop between system output and its excitation, it is used vaguely to describe evolving changes in the response characteristics of the system. The physical meaning and analytic implications of true feedback is seriously grasped by only a minute minority.

  46. Well
    There are several problems here Kip.
    First and foremost is you dont understand that UAH ( like DMI) is a ALSO a model of temperatures.
    In order to estimate the temperature UAH uses radiative physics ( for the microwave channel)
    One of the assumptions Spencer has to make regards the emissivity of the surface.

    They assume one value for all of the ocean
    They assume one value for all of the land.

    Looking at their code it seems clear they treat the area covered by ice in the artcic as if it were
    Ocean, and give it the emissivity of the ocean.

    They make no previsions for the emmissivity of snow or ice, and there is no accounting
    consequently, for the changes in emmissivity due to changes in snow cover or ice cover.

    From their document

    “From a practical perspective, however, the atmospheric temperature
    measurement (at least in the troposphere, where weather and climate variations are
    concentrated) cannot be made without also measuring at least some amount of thermal
    emission from the Earth’s surface shining up through the atmosphere for layers below TLS.
    Therefore, more surface-sensitive channels were included in the AMSU sensor design in
    order to better correct for this contaminating influence on the atmospheric measurements.
    This is important for the wide range of surface backgrounds in different regions,
    since the intended use of AMSU for monitoring regional temperature variations for input
    into numerical weather prediction models. We do not, however, perform any such
    corrections to our products.”

    Hint, they are telling you its no good for regional analysis
    Hint, Reanalsysis products will be better for the arctic (numerical weather prediction)

    • Mr Mosher, you are correct, as I remarked earlier, UAH doesn’t ‘do’ the white stuff.
      Adjusting the October NH accordingly gives a global anomoly of approx 0.14 , same as September.
      Of course to any sane person, 0.14 and ZERO are the same given the BS qualities of these models ( including yours).

  47. “Why is the Lower Troposphere Temperature in that circle at the top of the world, 60N-90N, behaving so differently than the rest of the world ?”

    Remember the earth in the tropics has net radiative energy flow in. Ie. The sun’s inbound radiation is consistently greater than the top of atmosphere outbound radiation.

    And at the poles it is of course the opposite. Especially in winter their is little or no inbound solar radiation.

    We have 2 poles, so let’s me stage that fact.

    In the winter the Antarctic interior is aabsolutely fridged! Why? The only source of heat is the air and air is a very low density carrier of energy. The air simply can’t keep up with the black body radiation and the land just gets ultra cold before equilibrium is established.

    The arctic isn’t as cold in its winter. Why? Most of the arctic is readily accessible to the “warm” ocean. In this case all liquid water is “warm”, at least relative to Antarctica’s winter.

    The arctic in winter is probably the only place on earth that is getting a majority of its heat from the ocean. That’s what makes it unique. It simply isn’t like the rest of the globe; its not even similar to Antarctica.

    So, back to your question. The surprise isn’t that it behaves differently. The surprise would be if it didn’t since the energy flows are totally different.

    • “Which do you believe, GISTEMP, HadCRUT4, UAH, or RSS?”

      I trust more those who do not change the past, very simple.
      I have not much trust in somebody who changes the past and exactly none in somebody who does that repeatedly.
      And you? Who do you trust?

      • henryp,

        you are correct. The LOESS curve calculation is not sensitive to changes near the end-points, so it misses the recent cooling.

        So don’t take too much notice of the end-points, but the rest of the LOESS curve should be ok.

        I am working on a new way of smoothing, using iterative smoothing with a binary filter (1, 2, 1). I am hoping that this will show the recent cooling, and also allow me to accurately calculate the “almost instantaneous” warming rate (rather than a warming rate averaged over a period of 10 years or more).

  48. The wider variation is due to errors in the map projection used to get unit areas of sampling. The errors/distortions are larger at the poles than at the equator. The minute you see latitudinal variability, you change map projection to see which one has the least temperature anomaly.

  49. “The Summers, however, have not been anomalously warmer. Summers show about 100 days of temperatures above freezing — and that by only a degree or so (never breaking above 275K — 0 °C = 273.15K)

    DO you know WHY the 2m temperature over ice would be limted during the melting season?


    the excess heat goes into melting ice

    • can’t have been much “excess heat” (who deems it to be excess rather than just heat ?) since 2000 given summer ice extent levels.

    • Mosher ==> Someday you might actually read the whole essay beofre commenting — or read nore closely with an honest intention to understand what the author is saying, rther than looking for things to snipe at.

      UAH 60N-90N (really 85N) also hugs the line — and that’s the troposphere — from 60N up.

  50. This has been around for quite some time:,74.18,1604/loc=5.942,77.915
    Notice the hot spot just west of Spitsbergen, nicely in front of the entrance of the bay on along which Longyearbyen is situated.
    SE of Spitsbergen is another, less pronounced hot spot.

    Could very well have something to do with this:

    Depending on the wind direction the warm water will directly influence the readings of the thermometers at Lonyearbyen airport.

  51. Arctic winter warming amplified by the thermal inversion and consequent low infrared cooling to space
    R. Bintanja, R. G. Graversen & W. Hazeleger
    Nature Geoscience volume 4, pages 758–761 (2011)

    Pronounced warming in the Arctic region, coined Arctic amplification, is an important feature of observed and modelled climate change1,2. Arctic amplification is generally attributed to the retreat of sea-ice3 and snow, and the associated surface-albedo feedback4, in conjunction with other processes5,6,7,8. In addition, the predominant thermal surface inversion in winter has been suggested to pose a negative feedback to Arctic warming by enhancing infrared radiative cooling9. Here we use the coupled climate model EC-Earth10 in idealized climate change experiments to quantify the individual contributions of the surface and the atmosphere to infrared radiative cooling. We find that the surface inversion in fact intensifies Arctic amplification, because the ability of the Arctic wintertime clear-sky atmosphere to cool to space decreases with inversion strength. Specifically, we find that the cold layers close to the surface in Arctic winter, where most of the warming takes place, hardly contribute to the infrared radiation that goes out to space. Instead, the additional radiation that is generated by the warming of these layers is directed downwards, and thus amplifies the warming. We conclude that the predominant Arctic wintertime temperature inversion damps infrared cooling of the system, and thus constitutes a positive warming feedback.

    • Relation between geomagnetic field and climate variability. Part 2: Probable mechanism
      N. Kilifarska, V. Bakhmutov, G. Melnik


      In this study we show that correspondence of the main structures of geomagnetic field, near surface air temperature and surface pressure in the mid-latitudes, reported previously in the 1st part of the paper, has its physical foundation. The similar pattern, found in latitude-longitude distribution of the lower stratospheric ozone and specific humidity, allows us to close the chain of causal links, and to offer a mechanism through which geomagnetic field could influence on the Earth’s climate. It starts with a geomagnetic modulation of galactic cosmic rays (GCR) and ozone production in the lower stratosphere through ion-molecular reactions initiated by GCR. The alteration of the near tropopause temperature (by O3 variations at these levels) changes the amount of water vapour in the driest part of the upper troposphere/lower stratosphere (UTLS), influencing in such a way on the radiation balance of the planet. This forcing on the climatic parameters is non-uniformly distributed over the globe, due to the heterogeneous geomagnetic field controlling energetic particles entering the Earth’s atmosphere.

      • It is well known that the unique source of charged par-ticles i n the lower stratosphere are the GCR, which consist from protons, alfa-particles and heavier nuclei (Z>3) with energy >1—2 GeV/nucI. The GCR depose their energy mainly at a height of 10-20 km, where the maximum of the air ionization (maximum of Pfotzer) is. This ionization takes place In the lower stratosphere, where the photo-dissociation rate decr eases ess entially and becomes comparable, even less than the CR ionization rate. The GCR are a source of ozone formation in the lower stratosphere, because of radiol ysis in the oxygen molecules . In t h i s way two mechanisms of ozone formation from GCR are observed: first, the radiolysis of the oxygen molecules; second, a participation of ions in the ion -molecular processes.

    • Winter temperature inversion: “We experience this temperature inversion almost every day while we are in the high Arctic at Eureka (Nunavut, Canada) for the Canadian Arctic ACE validation campaign. Due to the inversion, we experience much warmer temperatures (sometimes up to 20°C higher) at our research laboratory PEARL (the Polar Environment Atmospheric Research Laboratory), which is approximately 600 m (above sea level) higher than Eureka, which is at sea level. While our colleagues in Eureka are experiencing -50°C, those of us working at PEARL are often enjoying the relatively “warm” temperature of -30°C.”
      So the temperature can be 20 degC colder at surface than it is in the “lower troposphere”.

      • nobodysknowledge,
        An interesting direct observation of arctic inversion inducing + 20C delta T at 600M.
        Thanks for that!

    • And the UAH “arctic” consists of many weather systems. Greenland with much winter temperature inversion, Siberia, North Canada, Barents Sea and others.

      • Kip
        it IS the movement of earth’s inner core
        just google: map of movement of the magnetic north pole/

        it shows the movement over the last 1000 years but it accelerated quite fast over the last 10 years {first map].

        I call this the magnetic stirrer effect. It is just that the inside of earth has to re-align with whatever the sun is dictating.

    • …we find that the cold layers close to the surface in Arctic winter, where most of the warming takes place, hardly contribute to the infrared radiation that goes out to space. Instead, the additional radiation that is generated by the warming of these layers is directed downwards, and thus amplifies the warming.

      Might help to explain the contradiction between DMI surface and UAH TLT Arctic trends in recent years.

  52. First post over here. To me it does not seem to be related to ice extent as, according to Masie, the winter extent has even increased slightly from 2005, so this cannot be the explanation, see gif below:
    What has happened from 2005 is that solar cycle 23 has entered its minumum for a few years. As commonly understood, solar minums coincide with increased meridional patterns and less zonal winds. As the Atlantic is the main gateway to the Arctic the answers could/should be found in the NOA index as this leads to an increased exchange of heat between the subtropical areas and the arctic. One of the studies that has shown this link is listed below. This study seems to confirm, at least partially, the behaviour from 2005 onwards and would be interesting to extend this analysis with arctic winter temperature anomalies.

  53. The Arctic Ocean is a difficult place to live if you happen to be a human. Food, and water are scarce, and it is difficult to provide warm structures that allow humans to survive the cold, dark winters. We also have less permanent living organic systems to investigate the past warmth or lack of warmth. So both current measurements and past measurements are suspect, and averages of winter temps are even more suspect.
    Understanding of the variations in temperatures during the winter time is limited to a relatively short time frame from a climate perspective.
    The local areas of interest that we currently have some short term knowledge of are: the water temperature, the sea ice coverage, the land ice coverage, and the air temperatures.
    The air temperature is influenced by the local factors, and by non-local circulation patterns. The other areas of interest are also potentially influenced by a combination of local and non-local factors. Some second tier elements to keep in mind would be solar variations, changes in volcanic activity, changes in air chemistry.
    So you have a situation where you are investigating the end result of a matrix of many inputs, where some are local, some are not. We have many potential driving factors for changes in observed air temperature during recent winter years.
    It is likely that every driving factor will have to be analyzed to arrive at a consistent hypothesis. A few likely guesses would be:
    1) Arctic Ocean warm water intrusions / oscillation
    2) lack of cold air intrusions (from Siberia) -> warm air intrusions from warm water sources.
    3) less isolation between the warm water and the air
    4) greater cloud cover that slows the radiation of warmth to space
    5) a combination of many factors: warmer water leads to less sea ice, leads to more cloud cover, leads to warmer low temps…etc.

    More info is needed, but there is plenty to speculate about. And you have at least one thing working for you. Limited UHI 😉

    • Russ ==> Thanks for your input — I have mentioned that the UAH North Pole graph looks similar to a chaotic output overlaid by a seasonal signal. Many chaotic systems can shift to a chaotic randomness from a stable pattern under very small changes in conditions.

Comments are closed.