The Pause Update: January 2016

UAH v6.0 data for January were released this week.  The dataset has moved to Beta 5, which has caused some interesting changes.  Remember, Version 6.0 is a work in progress, with slight tinkering, until the final version is released.

Here are updated graphs for various regions showing the furthest back one can go to show a zero or negative trend (less than +0.1C/ 100 years) in lower tropospheric temperatures.   In some regions the pause has lengthened, in others it has shortened.  Note: The satellite record commences in December 1978- now 37 years and 2 months long- 446 months.  12 month running means commence in November 1979.



trends jan 16 globe

There has been zero trend for 18 years and 11 months.

Northern Hemisphere:

trends jan 16 NH

The Pause is back!

Southern Hemisphere:

trends jan 16 SH

The Pause has lengthened again.  For well over half the record the Southern Hemisphere has zero trend.


trends jan 16 tropics

Tropical Oceans:

trends jan 16 tropic oceans

22 years 5 months.

Northern Extra-Tropics (20-60N- where most people live):

trends jan 16 nh extr

Southern Hemisphere Extra-Tropics (mostly water):

trends jan 16 sh extr

North Polar:

trends jan 16 NP

The Pause is 13 years 11 months long- 13 months short of 15 years.

South Polar:

trends jan 16 SP

For the whole of the satellite record, the South Polar region has been cooling.  So much for a fingerprint of warming due to the enhanced greenhouse effect being greater warming at the Poles!


trends jan 16 Aust

20 years 10 months.

USA 49 states:

trends jan 16 USA49

The Pause lives!


Tags: ,

12 Responses to “The Pause Update: January 2016”

  1. MikeR Says:


    You say the following above-
    ”The dataset has moved to Beta 5, which has caused some interesting changes. Remember, Version 6.0 is a work in progress, with slight tinkering, until the final version is released.””

    However yesterday at 4.01 pm you sung a different tune.

    With respect to beta 5 you say in your comments at

    “”Australia- winter warming much faster ( 3 times faster). This seems odd to me as well.
    Thanks for bringing this to my attention.”

    Other aspects of beta 5 as compared to beta 4 are as follows. He relationship between rain UAH and rainfall that you demonstrated clearly for the Australian data (see ) is worse with beta 5 . The correlation coefficient for beta 5 is Rsq = 0.68 versus Rsq = 0.76 for beta 4.

    The correlation between UAH beta 5 and RSS is slightly worse than for beta 4 in many regions.
    eg. for the South Pole the correlation between RSS and beta 4 is 0.642 and with beta5 it has dropped to 0.618.

    Rather than a slight tinkering at the edges there appears to be radical changes in trends for some regions as you move from version to version and now from beta to beta.

    For Australia back in the days of v5.6 the trend was 0.16 C deg/decade, it then increased to a value of 0.24 C deg/decade for V6 betas1, 2, 3 and 4 and now it is back to a more reasonable 0.15 deg/decade for V6 beta 5. These changes are huge and what is more disturbing are the differences between this beta and the last for Australia- see .

    There are some people obsessed by changes at a particular land based station that are much less significant in comparison to these changes over the entire continent for the UAH data.

    There are similar differences for USA49 see .

    I understand Roy Spencer has reworked the land data ( ) and the ocean data ( ) which show significant differences.

    I am glad that the global data do not show much difference ( ). Thank goodness. Maybe there some method in his madness by means of ad hoc adjustments to reflect what he believes is reality
    Before I update all my spreadsheets once again, does anyone believe he will ever achieve some finality or will it always be a work in progress? I hope I live long enough.

    Finally Ken do you have any comment as to why the difference data for all the data sets starts to go berserk around the middle of 1998, which just happens to coincide with the start of the so called “pause” ?

    No wonder this version managed to raise the global pause from the dead, at least for one more month.

  2. MikeR Says:


    The other aspect of your work that concerns me is this 12 month mean averaging.

    The effect on the pause is obvious. First see the unprocessed raw data for the relevant period from March 197 until the present – .
    The trend has a positive value

    Now compare it the 12 month mean average data you employ- ..

    Notice how the steeply climbing part of the graph for the most recent data has been made to disappear which of course lowers the trend to a negative value.

    I discussed this trick last month. Sorry to repeat myself – “To elaborate, this sleight of hand is accomplished by averaging the last few warming months with the preceding cooler months, that make up the remainder of the 12 month interval. As this lowers the right end boundary values of the graph the this naturally forces the trend line to a lower (in this case negative value).”

    I did recommend Ken to use Loess smoothing last month , equivalent to 12 month averaging (it is very easy to use) and the result is shown here- .

    Notice for this form of smoothing, while fitting the data better over all, the data at the ends are smoothed appropriately. Needless to say the trend is positive.

    I won’t go into the issues of serial correlation with 12 month smoothing (using either method) as this would be complete overkill.

  3. Mason I. Bilderberg (MIB) Says:

    Reblogged this on and commented:
    I follow the Mann-made global warming debate.

  4. Neville Says:

    Ken you say this is version 6 beta 5 and all zones and countries listed now show a pause. Their maths etc are beyond my understanding, but I’d like to ask whether you think the previous beta 4 would also show similar results? For example would beta 4 show a pause in the NH? And if not why not?

    • kenskingdom Says:

      No- beta 4 would show different results. Whether or not the NH still had a pause would depend on the beta 4 January value. There has been a major rework of land v ocean algorithms as Mike R has pointed out. I don’t know if this will be the final version but I suspect not. There seem to be a couple of oddities. So we shall see.

  5. MikeR Says:

    Neville and Ken,

    According to beta 4 the N.H. pause is dead and buried. It was 3 feet under last month, is it is now 6 feet under. As a result of beta 4 probably not getting updated beyond December 2015 , I have done the appropriate trend calculation and used the beta 5 figure of 0.7C for January 2016.

    It is unlikely that if a beta 4 figure does appear it will be below 0.66 C . In fact it would have to below 0.23 degrees C to give a trend value below 0.1 degrees C per century which is Ken’s new benchmark.

    This is for the raw, unadulterated UAH data.

    Of course if you smooth it over 12 months then you delay the demise of the pause by a few months but, if in the unlikelihood it recovers due to a significant long term LA-Nina, this will be delayed by a similar amount.

    I think it is all going to end in tears for the global and N.H . data, unless the smoothing is increased to say 450 months for which you will certainly get the required result of a pause since 1979.

  6. kenskingdom Says:

    Mike, I am pleased you recognise 12 month means are neutral. A start. I (and many others) have long known and made public that the pause is likely to disappear in early 2016 because of rising temperatures as a result of the El Nino. While this will be cheered by GWEs it proves very little. Trends are capricious creatures, and especially when the data are non-linear- they don’t go where expected. The whole point of watching the pause is that there has undeniably been a period of more than 15 years when global temperatures have shown very little if any increase despite a number of ENSO events. This followed a very rapid rise over the previous 10 or so years. That is interesting in itself. Now whatever happens to the pause over the next 12 months or so, this period will remain. There may be another step change up; there may be a levelling off; there may be a drop down with a cooling La Nina. But there will still be this inconvenient period of rather flattish temperatures.
    I will be posting shortly a follow up to my earlier post on trends, but from a different angle.

  7. MikeR Says:

    I am not sure what you mean by the statement 12 month means are neutral”? Neutral in what sense, They certainly smooth out the latest El-Nino rise and consequently decrease the trend.

    Apologies, I was blissfully unaware that you have any other references to the imminent demise of the pause other than in a comment on February 2nd. I may have missed all you other numerous comments.

    You also say above- “Now whatever happens to the pause over the next 12 months or so, this period will remain”.

    Yes it will obviously.

    However this pause can only be located by careful selection of the period, by selecting both the start and end dates of the correct data set. It has to be UAH v6 as the pause disappeared a year or two ago with UAH V5.6, the minimum slope for this version is 0.1 deg/decade (10 times your brand new criterion) for any period longer than 7 years.

    So the pause depends on selection of the right version of UAH (clearly not any of the land based data sets or the radiosonde data) and carefully selecting the start and end dates.

    This a contortion equivalent to a backward somersault with triple pike difficulty 6.5.

    Now I have to bring up a delicate matter. That is statistical significance .

    For the data of figure 1 and using the Linest function of Excel to provide extra statistics, the slope for the range is -0.00028 ( slightly different to your figure Ken, maybe a slight mismatch with starting date or the offset for the 12 month smooth, my fit was until June 2015 as the last 6 months are not appropriately smoothed) with a standard error in the slope of 0.01485 i.e the standard error is 5 times larger than the value itself. Which indicates meaninglessness.

    The confidence levels are -0.0032 to +0.00265. So again this result is totally meaningless in a statistical sense.

    As a final exercise why don’t you find the number of 20 year intervals in the whole data set (there are 220 intervals, first interval from May 1979- April 1998, last interval June 1995-May 2015) where the trend is negative. To save you the calculations. The answer is none.

    You could also try 19 years and you will get the same answer. In fact you can try any period longer than 18 and 6 months years and you will never get a negative trend.

    In fact the closest you get, for an interval 18 years 7 months , are 11 out of the 220 that are less than your new criterion of +0.01 degree/decade. Again there are none for the 19 or 20 year intervals that come even close to being below your new criterion.

    If you find a negative trend for any interval greater than 18.5 years for the 12 month smoothed global UAH beta 5, let me know and I will check my calculations.

    Last but not least, we come to the presence of the much discussed decrease in slope from around 2002 until about 2012 , before the current rise from 2013 onwards. For an appropriate visual representation ,see Roy Spencer’s current page for January and his 13 month running average, that has both the raw data and the smoothed data on the same plot, unlike Figure 1 above.

    This reduction in trend for this period has been discussed to death along with the other numerous pauses (see ) that have been attributed to the usual range of suspects i.e ENSO, PDO cycles, etc..

    There is so much available with a simple Google search. I think at this stage, any further contribution from myself would be superfluous. But if you ask, I can provide so many links they will be coming out of your ears.

    I wouldn’t recommend that to my worst enemy.

    • kenskingdom Says:

      1. “Neutral” : will apply equally to future higher or lower temperatures; not selected for bias one way or the other. Referring to your comment that a La Nina led recovery “will be delayed by a similar amount.”
      2. I will pretend I didn’t see your referencing Skepticalscience. For goodness’ sake!
      3. You expend a large amount of time and energy trying to persuade me that I am wrong. This is wasted. As is my time and energy in replying to your comments.

  8. MikeR Says:

    By the way, all the above did not include the extra uncertainties due to serial correlation. If you want to see trend calculations with these included then these two sites are what you need.
    and .
    They are wonderful resources where you can see trends for almost every global temperature data set and their associated uncertainties (that include the required serial correlations).

  9. Neville Says:

    For me I’m starting to doubt a lot of the temp data. At WUWT Bob Tisdale’s latest post shows the surface temp average from 1979 to Jan 2016 to be 0.16C per decade and satellite average to be about 0.12 C per decade.
    So for the entire 37 years the difference is about 0.04 C per decade, or four hundredths of a degree C more warming at the surface than shown by satellites. I don’t think we can measure the surface or lower trop that accurately. Of course AGW theory dictates that there should be more warming in the trop than the surface, so the theory is busted anyway. IOW no hot spot no AGW and certainly no CAGW.

    Also of interest is the difference in warming between 1998 to 2016. Average for the surface data is about 0.13 C per decade and satellite only about 0.001 C per decade. Amazing that this shows about 130 times more warming per decade at the surface when compared to the satellites. And the warming from 2001 to 2016 shows about 8.5 times more warming per decade at the surface compared to the satellites. Does any of this make any sense?

  10. MikeR Says:


    Like you I am not just starting to doubt, but have been a long term doubter of the veracity of temperature data particularly the satellite data.

    The originator of the RSS data set , Dr Mears believes the surface based measurements are more reliable than his own RSS satellite data. To quote him ”A similar, but stronger case can be made using surface temperature datasets, which I consider to be more reliable than satellite datasets (they certainly agree with each other better than the various satellite datasets do!)”.

    Roy Spencer of UAH fame , unlike Mears has the utmost faith in his data, so much that he blissfully continues to make changes that significantly change trend values by ridiculous amounts. He went from v5.6 trend of 0.143 deg/decade to his latest and greatest V6 trend of 0.113 to more closely match the RSS data , the same data that Mears thinks is inferior to the land based data.

    Going from UAH v6 beta1 to beta 5 the difference between RSS is getting greater see for global data –

    Then there are the significant regional changes of greater than 50% going from one beta to the next. Australia is the prime example see

    UAH has now become such a dog’s breakfast that my kelpie would promptly regurgitate it, if it had the misfortune to accidentally swallow it. It has become a running joke.

    The only thing going for it now is it confirms that the past 6 months are the warmest on record ( from August to January) in line with RSS, Giss, NOAA, HadCrut and the radiosonde data (RatPac A) .

Comments are closed.

%d bloggers like this: