Ian Wilson: ENSO epoch’s and Earth Rotation LOD

Posted: December 3, 2011 by tallbloke in climate, Ocean dynamics, solar system dynamics

A. The above graph is part of Figure 2.1. from Klyashtorin, L.B., Climate Change and Long-Term Fluctuations of Commercial  Catches – The Possibility of Forecasting, FAO Fisheries Technical Paper No. 410, Rome FAO, 2001.

It shows the close correlation between the rotation rate of the Earth (measured by the Length-of-Day) and the zonal component of the Atmospheric Circulation Index (ACI). This graph shows that the zonal circulation patterns evident in the Earth’s atmosphere can be broken up into four 30 year epochs starting in the years 1880-85 [LOD curve only], 1905-1910, 1940-1945 and 1970-1975.

B. The above graph comes from figure 2.2 of Klyashtorin, L.B., Climate Change and Long-Term Fluctuations of Commercial  Catches – The Possibility of Forecasting, FAO Fisheries Technical Paper No. 410, Rome FAO, 2001.

The above graph shows that if you shift the LOD curve forward by ~ 6 years you get an excellent fit between LOD curve and the de-trended world mean temperature anomaly. Again the overall pattern can be broken up into four distinct 30 year epoch starting in the years 1880, 1910, 1940 and 1970.

C. The above graph comes from figure 2.23 of Klyashtorin, L.B., Climate Change and Long-Term Fluctuations of Commercial  Catches – The Possibility of Forecasting, FAO Fisheries Technical Paper No. 410, Rome FAO, 2001.

The above graph shows that if you shift the ACI curve forward by ~ 4 years you get an excellent fit between LOD curve and the de-trended world mean temperature anomaly (dT). Again the overall pattern can be broken up into three distinct 30 year epoch starting in the years 1910, 1940 and 1970. The ACI index does not extend far enough back to set a starting date for the first epoch but the dT and LOD curves suggest a date sometime around 1875 to 1880.

The (Extended) Multivariate ENSO Index

The Multivariate ENSO Index is defined at the NOAA web site located at:

The Extended Multivariate ENSO Index is defined at the NOAA web site located at:

The important point to note is that Multivariate ENSO Index is the most precise way to follow variations in the ENSO phenomenon:

Negative values of the MEI represent the cold ENSO phase, a.k.a. La Niña, while positive MEI values represent the warm ENSO phase (El Niño).

The Cumulative Sum of the MEI

If the cumulative sum of the MEI over a given epoch steadily increase throughout the epoch then the impact of El Ninos exceed the impact of the La Ninas over this epoch.
 If the cumulative sum of the MEI over a given epoch steadily decrease throughout the epoch then the impact of La Ninas exceed the impact of the El Ninos over this epoch.

The dotted red line in the above graph shows the cumulative sum of the extended Multivariate ENSO Index (MEI) between the years 1880 and 2000 A.D. The cumulative sum has been taken over each of the four 30 year epochs, starting in the years 1880, 1910, 1940, and 1970.

The solid blue line in the above graph shows the cumulative sum of the extended Multivariate ENSO Index (MEI) between the years 1886 and 2006 A.D. The cumulative sum has been taken over each of the four 30 year epochs, staring in the years 1886, 1916, 1946, and 1976.

It is clearly evident from this plot that whenever the cumulative MEI index is systematically decreasing over a 30 year epoch i.e. between 1886 and 1915, and between 1946 and 1975, the world’s mean temperature decreases. It is also evident that whenever the cumulative MEI index is systematically increasing over a 30 year epoch i.e. between 1916 and 1945, and between 1976 and 2005, the world’s mean temperature increases.


1. The ratio of the impact of El Ninos to the impact of La Ninas upon climate can be monitored over multi-decadal time scales using the cumulative MEI.

2. The cumulative MEI shows that since roughly 1880 there have been four main climate epochs, each 30 years long. There are have been two 30 year periods of cooling (i.e. from 1886 to 1915, and from 1946 to 1975) and two 30 year peiods of heating (i.e. from 1916 to 1945, and from 1976 to 2005).

3. Periods of warming occur whenever the impact of El Ninos exceeds the impact of La Ninas. Periods of cooling occur whenever the impact of La Ninas exceed the impact of El Ninos.

Ian’s excellent blog is here: http://astroclimateconnection.blogspot.com
This work provides a complimentary elucidation of the area of investigation which is the subject of the first post on this blog
.Ian very kindly wrote to me two years ago to recognise the independence of my own work in this area. In turn I recognise the originality and excellence of his work.
  1. tallbloke says:

    Ian, I tried to post the following on your blog but found I couldn’t get past the authentication system.

    “Excellent work!
    I’ve reposted this on my blog to try to give it as much exposure as possible.

    Well done.

    Rog TB”

  2. vukcevic says:

    Two out of three major climatic oscillations are the North Atlantic and the Equatorial Pacific. As far as I see it they are natural processes which can be only properly understood by studying ocean currents circulation, and why there is a natural variability. Once I understood what is happening in the North Atlantic (I am still sorting out many of the minor details) the Equatorial Pacific appeared to be much simpler affair.
    Here is result of my preliminary investigation, inspired by dr. Archibald’s post on WUWT some time ago:
    I had the driver worked out, but it was the dr. A’s cumulative SOI that cracked it. Once I am finished with the Atlantic, hopefully I’ll get to this one too.
    On the LOD I favour dr. Andy Jackson’s hypothesis.

  3. tallbloke says:

    Hi Vuk,
    I’ve exchanged a few emails with Andy, as well as his former PhD mentor who is now retired. I tried to get them interested in the excellent correlation I found between LOD and ‘z’axis relative motion between the Sun and the solar system’s centre of mass.

    It is a correlation which has a better R2 than most, but they refused to bite. I guess they don’t want anything to do with something which might be regarded by others in their field as ‘astrology’.

    It saddens me that science is retarded by these entrenched prejudices.

    The other planets affect the rate of Earth’s spin. What’s the problem? It’s a no-brainer once you consider it.

  4. adolfogiurfa says:

    @Tallbloke: Remarkable consonance!. Now we need to discover its nature.

  5. Ninderthana says:

    Thanks Tallbloke for hosting my short post on your site. I have added very little to what people like yourself, Vukcevic and others have already done on many previous posts on this topic. The two small additions to this topic that come from this post are that if you reset the cumulative SOI (or MEI) to zero at the beginning of 30 year epochs beginning in 1886, 1916, 1946 and 1976, you get continuous increases in the cumulative SOI (or MEI) when the World’s mean temperature warms and continuous decreases in the cumulative SOI (or MEI) when the World’s mean temperature cools.

    Vukcevic, what is Dr. Andy Jackson’s hypothesis on the LOD? I would be very interested to here what he has to say. Do you have a link to his work.

    Tallbloke’s ssb-z solar motion vs LOD is absolutely remarkable and it matches my parallel result which show that long term deviations in the LOD also closely match assymetries in ssb-(x,y) solar motion.

    http://www.wbabin.net/Science-Journals/Research%20Papers-Astrophysics/Download/3811 – page 19.

    Like you, Tallbloke, I am totally baffled as to why our joint discovery has received so little interest and attention from “main-stream” scientists. Vukcevic (and a few others) appears to be trying to get to the bottom of this problem by investigating possible mechanisms for this link.

  6. tchannon says:

    A definite extraterrestrial effect on earth which has escaped notice demands extreme evidence.

    Noticing as countless school children did that Africa fits with South America took lifetimes to become accepted via plate movement.

    From what I recall the data is suspect and the fit does not survive so well on close examination. A point here is that any external influence cannot be sometimes or vaguely and very particularly must not drift in time, which is I think the awkward problem.

    From work I did for PV the lunar effect on LoD is very tightly coupled but fails before the far superior instrumentation era, where the better data starts 1962, still not wonderful. The US military only give the data from 1991.

    A useful addition would be a completely independent dataset.

    I’m not dismissing the idea.

  7. Ninderthana says:


    You have made some good points which any proposal must get around. Like any good sceptic, I do not blame you from holding back.

    The problem with the objections you have made is that [in general] they have been made by people who are not willing to look past the obvious connections before giving up their [tepid] search.

    I can testify to you that a few people have taken the Lunar tidal connection a little further and that they may have an explanation as to why the Lunar tidal mechanism only works during certain epochs and not in others. I agree that they will have to present an extremely good case in order to convince those who are holding back like yourself. That is reasonable.

    You might want to look up the research of Claire Periguad if you want to see how “main-stream”: scientists are spending hundreds of thousands of research dollars following up this link.

    Another lead you might want follow is to look at the third last figure in this post:

    [this item]

    And the first figure in the post:


    and the following two posts/publications:




  8. tchannon says:

    Second link you give. I assume the last point is 1990 and the one before that, just before 1800, do you have a date?

    This post might interest you, one reason why I mention 1990, Figure 6 especially
    (I am co-mod, post articles sometimes, such as above — Tim)

  9. tallbloke says:

    From what I recall the data is suspect and the fit does not survive so well on close examination. A point here is that any external influence cannot be sometimes or vaguely and very particularly must not drift in time, which is I think the awkward problem.

    Good points Tim.
    The similarity in the curves for LOD and SSBz motion I discovered have the important caveats that the LOD curve lags the SSBz curve by 30 years, and that I smoothed monthly data over two Jupiter orbits. On the face of it this is not very scientific, but there is some reasoning behind my choice of smoothing period, which is to flatten out the large shorter term swings caused by Jupiter; the dominant driver of solar motion WRT the barycentre.

    So what is my justification for that? Well, I’m considering the hysterisis of Earth’s internal flows. I believe they create the equivalent of a ‘well damped oscillation’ which will in effect smooth shorter term external forces in a way analagous to the averaging of the data over the timescale I have used.

    So why the thirty year lag? I’m not sure. Instantaneous and well coupled is great when you can isolate it. I suspect the lag is apparent rather than real, and that the mechanism behind the effect of planetary motion on Earth’s length of day is not simple gravitational mechanics, but rather an electro-magnetic dynamical force. This may well have effective power not when gravitational alignments reach inflexion points, but rather when suitable angular offsets set up resonances in the complex field interactions.

    If all that sounds a bit woolly to you, then fair enough, it’s still a fairly vague hypothesis in my mind too. Part of my motivation for blogging is in the hope of attracting minds more acute and with greater data handling capabilities than I possess to engage with the possibilities pointed up by the curve matches I’m discovering. Maybe it’s a long shot and a vain quest, but I’ll keep trying, because my intuition is telling me there is something important going on here.

  10. tallbloke says:

    It’s interesting to note the ‘wartime bump’ on Ian’s extended MEI cumulative plot. Which areas does the extended dataset cover which are not in the standard MEI Ian?

    Could skewing introduced by extra military naval activity in these areas account for some of the wartime bump in the global SST figures we were discussing in Tims datasets thread?

  11. vukcevic says:

    Andy Jackson lecture to AGU:

  12. Ninderthana says:


    You need to be careful about how you interpret the cumulative MEI diagram.

    First the extended MEI uses two variables (i.e. sea-level pressure (P; details in Allan and Ansell, 2006), and sea surface temperature (S; details in Rayner et al., 2006)) instead of the normal six.

    Second the MEI (based on six variables) used in my post covers the period 1950 to 2005, while the extended extended MEI (based on two variables) covers the period from 1871 to 2005.

    The WWII bump in the cumulative extended MEI is not a bump in a continuous record. The variable is artificially reset to a value of zero every thirty years, effectively making the long-term shape of the cumulative MEI curve meaningless. What is important is what the cumulative MEI does over a thirty year period once it has been set to zero. Does it systematically increase from zero or does it systematically decrease from zero? That’s what we are really measuring in my plot.

  13. Ninderthana says:


    Figure 6 shows that the times when Solar/Lunar tides had their greatest impact upon the Earth are closely synchronized with the times of greatest asymmetry in the Solar Inertial Motion (SIM). Over the last 800 years, the Earth has experience exceptionally strong tidal forces in the years 1247, 1433, 1610, 1787 and 1974 (Keeling and Whorf, 1997). A close inspection of Figure 6 shows that these exceptionally strong tidal forces closely correspond in time to the first peak in the asymmetry of the SIM that occurs just after a period low asymmetry. These first peaks in asymmetry in the SIM occur in the years 1251, 1432, 1611, 1791, and 1971, closely correspond the years of peak tidal force.

    The two dates for the peak asymmetry in the SIM (Solar Inertial Motion about the Barycentre) are
    1791 and 1971. They are roughly 179-180 years apart i.e. occurrences of the same planetary configurations for the Jovian planets which take place after one Jose cycle.

  14. tallbloke says:

    Ian, thanks. So I guess the Extended MEI data will already contain whatever anomaly the WWII bump represents anyway via sst records?

    Thanks also for the info on peak tides. Interesting.

  15. vukcevic says:

    Ninderthana you will find section devoted to the LOD in the AGU lecture linked above.

  16. Ninderthana says:


    Andy Jackson thinks that the 6 year oscillations in the core are just the period of the fundamental mode for the toroidal oscillations. Here is why he is wrong.

    1. He makes the argument that the LOD oscillations peaking at 6-7 years matches the oscillations
    expected from the fundamental mode for the toroidal oscillations. The [erroneous] assumption
    that he makes is that there are no external forces acting the solid Earth. However, it is well known
    that there are variations in the Earth’s LOD which match the 18.6 year lunar draconic cycle.
    [this shows that Lunar/Solar tidal forces does in fact affect the Earth’s rotation rate]

    2. Remarkably the ~ 6 year oscillations in the LOD seem to almost perfectly phase-match the
    6 year variations in the Lunar orbit which characterize the time required for the Lunar line-of
    -apse to realign with the Earth-Sun Line. In other words, it is just possible that the torsional
    oscillations in the core are being driven by the fact that Moon reaches perigee at New Moon
    once every 6 years.

  17. tallbloke says:

    Vuk, I’m halfway through watching Andy’s lecture. The LOD stuff is around 23 to 29 mins in. I’m pleased to see the finding of the ~6 year cylinder oscillation period corresponding approximately to the 6-7 year peak in the spectrum of LOD. This fits well with a consideration of the relative motions of Jupiter, Saturn , the Sun (and Moon!) and the nose of the heliosphere as seen from Earth I think. I’ll reflect on it some more after watching the rest.

  18. tallbloke says:

    Ian: Andy does mention ‘lunar’ briefly in his discussion of the ~6 year torsional oscillation period. i wonder if this has anything to do with the notion of a linkage between full/new moon at perigee/apogee, and earthquakes as the relative motion of the cylinders reverses.

  19. tallbloke says:

    Leif just gave me these links, along with an insult. 🙂


    Click to access ar2000_055.pdf

    From the first one:
    Excess of the duration of the day to 86400 s and angular velocity of the earth’s rotation, since 1623

    This table gives mean annual values of the duration of the day D, which are available for the last four centuries. For the interval 1623-1955, the data are those provided by L.V. Morrison, Royal Greenwich Observatory, interpolated for the middle of the year. The mean solar time has been referred to the dynamical time scale derived from the time argument of the lunar ephemeris. The duration of the day has been obtained :

    – from 1623 to 1860, by derivative of cubic splines fitted on individual values of the difference between mean solar time and dynamical time (13 knots),

    – from 1861 to 1955, by a 5-point quadratic convolute.

    More information on the computation of the duration of the day is available in Stephenson and Morrison (1984), with an estimation of the accuracy of these evaluations. From 1956 up to present, the duration of the day has been obtained from the BIH/IERS values of UT1-TAI ; the table gives annual averages. At the level of precision of these values of the duration of the day, the unit of the dynamical time and the unit of TAI can be considered as having the same duration. Thus D is expressed in present SI units. The table gives also the values of the angular velocity of the earth’s rotation w derived from the listed values of D.

    So, how well does this match the LOD series of the Russian scientist Ian collaborated with in his SSBx,y study? Fig 3 of your paper here:

    Does the method used by IERS for earlier results take any account of ‘other factors’?

    Here is Leif’s plot from the IERS data

  20. Roger Andrews says:


    Bob Tisdale posted a graph that’s very similar to your cumulative MEI sum graph in June 2009; .
    He claims that El Ninos and La Ninas have cumulative warming effects – they don’t cancel out. Thought you might care to check his results if you haven’t already done so.

  21. tchannon says:

    Ninderthana, getting a bit messy without threading here. I’m going to have a think about the dates etc., a slow process of absorbing info. The figure 6 is hinting about the connection with the solar magnetic cycle, which is more important than the sunspots but we have little long data. How this ties in with anything else seems little known.

    LoD, hmm. Well I didn’t want to overload this thread, so this is shorthand.
    Chandler wobble, http://www.gpsl.net/climate/data/chandler-wobble-c.pdf
    and something done quickly for elsewhere hinting about how the older data might drift off

    I failed to find a driver for the ~6.4y so this is an ongoing work which I look at from time to time as things turn up. I don’t buy the idea of some mysterious 1.18y ocean floor flapping as a causal.

  22. Stephen Wilde says:

    This is looking helpful to get a link from planetary movements to variations in the rate of energy release from the oceans and thence to the troposphere on a 60 year cycle.

    I’m thinking that ENSO has a cause related to the solar energy input imbalance either side of the ITCZ given that the ITCZ has a mean position north of the equator. That gives us the observed interannual variability.

    Seperately I think the 1000 year cycling MWP to LIA to date is solar induced. That would give the upward stepping observed from one 60 year cycle to the next which has probably been ongoing since the bottom of the LIA as the system recovered. Likely there was downward stepping from MWP to LIA.

    That leaves a gap to be filled as regards the cause of the 60 year oceanic changes whereby El Nino dominates over La Nina for 30 years (warming troposphere) and then vice versa (cooling troposphere) for 30 years.

    It seems entirely plausible to me that planetary movements could induce such shifts of balance in the main body of the oceans to produce a 30/60 year oscillation.

    Given the huge thermal capacity of the oceans as compared to air it wouldn’t need much change to generate the observed effects on tropospheric temperatures.

    Mind you, whatever the cause, the effect is ultimately going to be a bottom up ocean induced shift in the permanent climate zones as the energy in the oceans moves to and fro.

    So whatever forcings operate on the system they will be countered by latitudinal shifts in those zones as a negative reaction that maintains system stability.

    I don’t see that we need climate cycling on more than those three timescales to produce every climate phenomenon that has been observed.

  23. tallbloke says:

    Tim, threading comments seems to work at climate audit. We can give consideration to enabling it here. I wonder how well it’ll work with mobile apps such as the android wordpress app. I worry that it will lead to a less democratic feel to the blog too.

    Sub-conversations which merit their own threading could be put up as a new post perhaps?

    [Climate Audit has a much larger main text width. Could create a junk WP blog cloned from here and experiment to see what happens. For the time being the view of keep it as it is seems sensible. We don’t often get into Byzantine dialogs — Tim]

  24. tallbloke says:

    “I don’t see that we need climate cycling on more than those three timescales to produce every climate phenomenon that has been observed.”

    Stephen, I agree except for the above quote. I don’t think we should discount the possibility that oceanic storage of solar energy and later release in a big surge might help account for the big swings between glacial/interglacial conditions which amplify the milankovitch triggers. Best discuss this on the glacial climate thread to help maintain focus here.

  25. adolfogiurfa says:

    @Tallbloke:Leif just gave me these links, along with an insult. 🙂

    “Once upon a time, in the middle of the jungle, there was a group of lions surprisingly hearing with deep attention a violinist playing….Suddenly appeared a new lion, jumped over the player and ate him…..one of the lions that was hearing the concerto said: “Here it came that damned deaf to spoil us the concert!” 🙂

  26. adolfogiurfa says:

    @Vukcevic: See this [snip]

    [Reply]Try Michele’s thread for this Adolfo too far OT here.

  27. tallbloke says:

    nice. I see interesting periodicities in the CET data.
    Please could you explain this item in the legend:

    CET(1880-2010) LOD


    Also, some interesting ponderings on LOD, solar activity, and other connections 2/3 down the page here;

  28. jim hogg says:

    It seems fairly obvious to me that the sun, the planets and the moon must be major players in our climate, and I’m glad to see that younger intellects are at work in this field. . .

    I’m assuming that everyone involved in this debate, and lurking, is familiar with the pioneering work of the very impressive Rhodes Fairbridge . . . . . .If not, a useful summary by Mackay can be found here:

    Click to access fairbridge_rhodes.pdf

    There are also many valuable references. . .

  29. tallbloke says:

    Hi Jim, and welcome. Yes, Rhodes Fairbridge was a fine polymath and cross disciplinary thinker. His work on collating data and evidence for cycles in climate has helped us along enormously.

  30. P.G. Sharrow says:

    Great hunks of metal traveling in space, dynamos spinning, within the fields of dynamos, all connected by the fabric of space. Mass/inertia, Electro-Magnetic forces, gravity, all parts of the whole.
    Weather/Climate is just a slight effect of enterplay of these giant forces.

    Tallbloke and his little ragged band are moving the ball down field while the “experts” are fixated on the commercials. Soon enough you will be able to post a big one on the scoreboard! Cheers 😎 pg

  31. vukcevic says:

    may be OT but it is great fun
    vukcevic drove Tamino ‘nuts’ on RealClimate:
    1 , 2 , even ‘Sceptical Science’s Daniel Bailey is getting on the bandwagon 3

  32. tallbloke says:

    Vuk, you drive us all nuts with your mysterious ‘drivers’ and elliptical legend items. 🙂

    Answer my question five comments up dammit!

  33. vukcevic says:

    Elementary my dear Watson, it was placed in a wrong place, now corrected:
    I’ll email the answer to Tamino and Bailey just in case the RC censors it.
    Dr. Svalgaard of Stanford and Dr. Steig of Washington Universities have all information required. None have protested any irregularity.

  34. tallbloke says:

    Aha! You’re welcome.

    I wish you’d make the LOD curve a nice prominent colour to separate it from the CET 1880-2010 curve though. I’m having trouble seeing the difference between the two colours on my lappy.

    It’s fascinating how different the longer period CET spectrum differs so much from the temp series which run from 1880.

    Any chance you could add the longer term LOD spectrum from the EIRS data too? (Not that I have much faith in it, but I may get a surprise).

  35. vukcevic says:

    I don’t think the LOD prior to 1860 is any good.
    Re: CET , longer the data set more selective spectrum analysis, what you see there is the 3 distinct components merge into one, you need at least 2-3 cycles for good resolution, the lobes are misleading, they just show transition between two neighbouring dominant components:

  36. tallbloke says:

    the lobes are misleading, they just show transition between two neighbouring dominant components:

    Does that mean the time periods at the peaks of the lobes don’t tell us anything about cycle lengths, because their position is ‘pushed around’ by the relative strengths of the lobes to either side?

  37. Ninderthana says:

    Most of the LOD data in the 17th to 19th centuries is based on meridian transit observations of “fixed” stars across the (local) meridian. These observations are done with instruments that are specifically designed to produce as accurate results as possible. The precision of the final number depends on the number of meridian transits observed on a given date. In many cases you are
    dealing with thousands of observations taken from various places around the globe.

  38. tallbloke says:

    Thanks Ian,
    I did an eyeball comparison of the IERS data from 1640 and the LOD data you used in your paper and there are some fairly noticeable differences. Is this due to the historic independence of the Russian data?

  39. tchannon says:

    First, a useful web site to do with LoD is the other world, Observatory of Paris

    And for LoD this page adds some background, particularly stating in plain language the mangling used on all non-recent data, which has no skill and will hide any detail present in whatever actual data is used.
    “from 1623 to 1860, by derivative of cubic splines”
    “from 1861 to 1955, by a 5-point quadratic convolute”

    Plot from obspm.

  40. tchannon says:

    Second, I cannot reproduce Vuk’s result.

    Hamming window (to make things simple)
    Not bothered normalising, scales unimportant.

    Perhaps I have misunderstood.

  41. tallbloke says:

    Hi Tim, yes, your LOD from ~1625 is simply the inverse of the plot from the IERS data Leif has on his site which I reproduced upthread. What I’d like to know from Ian is why his plot differs in fig 3 of his paper here:


    In some ways this is of historical interest only, because I agree with you and Vuk that using LOD data from prior to the start of Gross’ reconstruction is likely going to pollute the signal. I explained this to Dr Leif Svalgaard (somewhat forcefully) yesterday on WUWT. I wonder what x-y plane result Ian would get if he re-ran his calcs with the later data only.

    I hope both Vuk and Ian will tell us more so we can iron this out and agree on datasets and basic processing.

  42. vukcevic says:

    Sorry Tim, CET here

    was mislabelled (it appears I do have problem with labelling), that should be CET summer June,July,August, while this one
    is the annual CET

  43. tchannon says:

    Seasonal explain it.

    This is a bit misleading for England. We have predominantly near maritime from the Atlantic, prevailing south west winds except that critically blocking weather occurs, any season. Last year or so we have had odd patterns related to this.

    High pressure over Europe or Scandinavia reroutes the Atlantic weather or stops the flow. This lasts up to about 6 weeks and then unusual heat, cold and or dry occurs. During spring or autumn is is merely unuaul for the season without breaking min or max.

    Normal meteorology maths practice tends to mess up the data hiding some of this. (simple example, very cold around 1st December will not figure as an exceptionally cold December, shows as a lesser extent for November and December. Also note that say 3 weeks of exceptional tends to disappear.

  44. vukcevic says:

    At the moment I am very busy, writing an article dealing with seasonal differences, so graph was plucked out from that selection, not meant to mislead, since there is the other link with the annual spec. I shall replace or even better delete, so there is no confusion.

  45. tallbloke says:

    Vuk, don’t delete anything please. We’re just trying to clarify datasets, not making out that anything deceptive has been done. Anyway, I already saved copies. 🙂

    I think the seasonal graph is very interesting in terms of the periodicities revealed in the longer term CET data. We have nearly half a ~970 year cycle’s worth of data there, and it is valuable.

  46. Roger Andrews says:

    According to the latest thread at RealClimate (http://www.realclimate.org/index.php/archives/2011/12/global-temperature-news/#more-10128) we’ve got this all wrong.

    ENSO doesn’t control temperatures, all it does is generate short-term noise. When we remove this noise all of the commonly-used temperature time series – including TLT – show the same amount of warming (http://www.realclimate.org/images//FR_fig5.png).

    Our thanks to Grant Foster and Stephan Rahmstorf for this incisive analysis. 😉