Polar temperature change during satellite era

Posted: August 18, 2011 by tchannon in climate



Figure 1

A minor fuss has errupted over ERA40 reanalysis data at Arctic latitudes so I thought it would be useful to post how I see some of the data.

[EDIT: SERIOUS ERROR CORRECTED, untested software, mistake by me, had an incorrect weighted mean showing a grossly small temperature range]

WUWT article

This pair of plots were produced here from gridded data for circles +70 latitude to the poles. UAH data does provide what seems reasonable data. Data to July 2011.

As with the south pole station data there is zero temperature change for Antactica.

Various results are possible for the Arctic and elsewhere I showed a reasonable inverse match between sea ice and UAH TLT temperature.

In private I have posited the idea that earth temperature wobbles around a pivot at the south pole but this is not shown. We need much longer data to see how true that is. There is agreement between satellite and HADCRUT3 but the latter shows signs of severe internal problems if the clock is wound back far.

I think there is plenty of evidence supporting a view of cyclic Arctic change, including in formal papers.

How the plots were done

Some time ago I knocked up some scripting software which processes published gridded datasets into a common format. Various processed output is provided.

In this case all latitude bands pre-weighted are output, a large grid of time series. It is then a simple matter to compute a combination.

(for UAH the outputs are exact clones of published, ie. the math is the same and for UAH there is no missing data)

One thing I have never done is figure out a list of latitude/longitude pairs for extraction which match known sea ice.

There is some question over the validity of polar data from these satellites given the poor view. Odd is that it works even during winter total darkness but the rough sensing altitude is 15k feet where maybe there is still enough light.

  1. tchannon says:

    I hope the image appears, WordPress seems to asleep on serving images.
    [update: strange, only fixed on clearing WP browser cookies]

    If the subject of this post is not of interest please let me know. There are many topics which could be covered.

  2. Doug Proctor says:

    The quantum of temperature change for the northern latitudes does not seem to be enough to justify Hansen’s claim that his “superior” Arctic data is the reason GISTemp has a 0.12C higher global temperature than HadCruT or UAH. The amount of specific Arctic temperature rise needed to add to HadCruT would appear to be several times what UAH sees.

    Do you see this disconnect?

  3. Tenuc says:

    Thanks again, Tim. So not polar amplification then – rather polar oscillation?

    Makes one wonder how ‘the team’ are going to manage without the now discredited ERA40 reanalysis data to scare us with. The ‘science’ of climatology really is in a bad state!

  4. tchannon says:

    I suppose it could be called polar oscillation. Done plenty on this without completing enough make a major point by showing comprehensively. (takes time and effort to put things together)

    The south pole being high up a mountain all surrounded by heat curtains does complicate matters.

    An amusement was discovering one of the core alarmists have available almost the first paper where their name appeared, with a famous lead, right at the start of the enviro money spinning where they cite Arctic change starting with aerial photos 1940s. Complete that to today and up and down it goes. Read old marine and exploration, up and down it goes.

    Why? Haven’t got close enough yet. My suspicion on something causing solar asymmetry unless that is more directly the cause.

  5. Tenuc says:

    Perhaps something to do with the magnetic poles and the charge field. Changes to field strength/charge density seem to effect both Earth and sun (and perhaps all the other planets).

    Where’s Vuk when you need him???

  6. Roger Andrews says:


    “So not polar amplification then – rather polar oscillation?”

    The work I have done on this admittedly isn’t diagnostic, but it suggests that what we are actually looking at is “extra-polar decay”; that is, the oscillations begin in the Arctic, decrease in amplitude as we go south and take a couple of years to make it down to the Equator.

    Also note that the positive correlation between temperature and the Scafetta 60-year cycle goes negative in the Antarctic.

  7. tchannon says:

    I have corrected a serious plotting error, software had never been tested when used in that way, Taken most of the day to discover two mistakes. Data value is a weighted mean, I had forgotten about that. I knew there was one missing data point and ignored it but finding the bug was hard. I had incremented the date _before_ output, so it lost the first datapoint. Idiot.

    Computation from gridded is
    sum (cosine(latitude) * grid cell temperature for latitude band)
    sum(weights used)

    This corrects for the variable area of each grid cell. (cells are tiny at the poles)

  8. P.G. Sharrow says:

    One of the problems with Arctic temperatures is that they are inferred from a distance and not measured locally. Polar ice cover seems to be a factor of wind paturns that move the ice into warmer waters.
    Antarctic temperatures are measured locally, by trained people. Harder to fudge the data. pg

  9. RACookPE1978 says:

    Reading quite a few original research papers about the Arctic radiation, albedo, sea-ice trends and clouds and temperature trends dating from the mid-60’s through the mid-2000’s …

    Several of the earlier papers talked about the the apparent oscillation between Antarctic and Arctic findings – as if the two slowly changed at the same periodicity, but in opposite “polarities”: The Antarctic total ice increased/temp’s decreased as the Arctic ice decreased and Arctic temp’s increased, etc.

  10. RACookPE1978 says:

    So … What about NASA-GISS Hansen’s claim that the Arctic is now 4 degrees higher from a baseline of 1976? Doesn’t this contradict that claim completely, since the data is for slightly further north, but is less 1 degree for the entire period?

    DMI 80 north latitude shows a steady decline in summertime temps since 1958.

  11. P.G. Sharrow says:

    @RACookPE1978 says:
    August 22, 2011 at 5:20 am

    I am delighted to see someone is doing real examination of contemporary writings and research papers. Real science has known of this north south oscillation for a very long time. The choice of 1976 as a baseline was political not scientific. The 1970s was known by Hanson to be the point of swing from cooling to warming in the north hemisphere. pg

  12. Nic L says:

    Isn’t the UAH MSU TLT temperature data rather unreliable over most of the area polewards of 70S because of its high altitude? Per their Readme file at http://vortex.nsstc.uah.edu/data/msu/docs/readme.msu:

    The areas of poor anomaly values are : Tibetian Plateau, Antarctica, Greenland and the narrow spine of the Andes.”

    AVHRR satellite data (which measures surface skin temperature) is an alternative, but needs cloudmasking and appears to suffer from positive drift and various inhomogeneities. [Comiso’s cloudmasked AVHRR data for Antarctica for 1982-2006, used for the Steig et al 2009 temperature reconstruction and also for the O’Donnell et al 2011 improved reconstruction, is available at http://faculty.washington.edu/steig/nature09data/data/.%5D

  13. tchannon says:

    Nic L,

    “Isn’t the UAH MSU TLT temperature data rather unreliable over most of the area polewards of 70S because of its high altitude? ”

    A good question.

    My position is that I used to worry but reached the conclusion it’s good enough.

    RSS do blank out the high plateaus, UAH does not. (ask if you want to see this graphically)
    Cristy (UAH) does warn about it, shows it anyway (as you point out).

    The difference between the two is mostly about high frequency response. RSS pushs hard, UAH smothers more. No right or wrong on this.

    I spent many hours working on the mapping. During that time I did not spot any problem with the high areas for UAH. Put another way the detail there looks contiguous and sane.
    The exception is Antarctica where the area is large with no good compare. However, it seems to match the ground record well enough.

    I am aware of the AVHRR matter. All rather dodgy and as a reminder I highlighted the infamous AWS problem with Mcintyre. Little has been said about the detail but at the time I had some of the raw AWS data (some because there are GB of it) at the same time as BAS/NERC were playing games via their web publish version and I could forward/reverse engineer what they were kidding on about, particularly over missing data. All highly amusing and as far as I am concerned not actually important.

    Upshot is I do not trust anything about Antarctica temperature. For example, the detail of station location is obfuscated/missing yet I have unearthed station moves involving altitude. Keep in mind here it was altimeter (barometer) information with undid BAS.

    Feel free to disagree and discuss, I make mistakes same as anyone else.

  14. RACookPE1978 says:

    Nic L says:
    August 22, 2011 at 9:12 pm

    Isn’t the UAH MSU TLT temperature data rather unreliable over most of the area polewards of 70S because of its high altitude? Per their Readme file at http://vortex.nsstc.uah.edu/data/msu/docs/readme.msu:

    The areas of poor anomaly values are : Tibetian Plateau, Antarctica, Greenland and the narrow spine of the Andes.”

    No. Not a factor. Note the use of “altitude” and not “latitude” .. 8<)

    The very far north can be divided into three general regions:
    North of 80 North latitude: 90% of the area north of 80 north latitude are simply Arctic ocean. Sea level. No land at all of any altitude.
    The very few sq kilometers of Greenland included in this region begin at the ocean side (altitude = 0.0, then grow up towards the mountains. So, less than 2% are over 1500 meters altitude.
    South of 80 north: Land-side tundra, ocean-level waters and ice. Only the Greenland central plateua, Norwegian and Swedish mountains, and a little bit of the Urals are high. Canada's western mountains extend into Alaska, but look at the 'real world" geography: Barrow AK is sea level on the north coast of Alaska at latitude 72 north. So, even what little (flat, near-sea-level) tundra is between 72 north and 70 north, that land is not above 1500 meters elevation.

    The very far south varies: Antarctica has a high central plateau, but you'd need to establish that the satellite readings differ from the ground-based weather stations.

  15. tallbloke says:

    Thinking geographically, it seems natural that Antarctica has a more stable temperature regime than the arctic. It doesn’t mostly disappear in summer. It has the circumpolar ocean flowing freely and it is less affected by changes in wind patterns than the North with it’s resticted channels.

    So … What about NASA-GISS Hansen’s claim that the Arctic is now 4 degrees higher from a baseline of 1976?

    4 degrees… Fahrenheit or Centigrade?

  16. RACookPE1978 says:

    He claims +4 to +5 C across just about all of the Arctic tundra, then flows his red and orange clouds north 1200 km across the Arctic Ocean. I don’t recall his latest Russian-Siberian extrapolations accurately enough to mention them.

  17. tallbloke says:

    Heh, the guy is an embarrassment to NASA. When will they gently put him out to grass?