A 60 year oscillation in Global Temperature data and possible explanations

Abstract: A temperature variation with a time period of a few decades has been previously reported [1] and observed in SST data. In this post,  a 60 year  oscillation with an amplitude of 0.1C is identified in the HadCrut global temperature data[2]  and will  likely cause a halt in global warming over the next decade or so.  An astronomical origin  is proposed as a possible cause of this effect. No single astronomical effect  can explain  the 60 year time period, so the possibility of an interference between the 11 year variation in solar irradiance with magnitude 0.05C and an 18.6 year 0.05C ocean tidal component due to precession of the moon’s orbit plane has been investigated. However, there is no convincing evidence that this can drive the 60 year 0.1C temperature oscillation either.

Introduction: The long term trend in global average temperature data [2] fits rather well to a simple logarithmic dependence on CO2 concentrations[3]. A simple model based on incremental surface warming to the radiative forcing of CO2 in order to restore overall energy balance, predicts a dependence deltaT = 1.6Ln(C/C0), where C is the CO2 concentration in any given year and C0 is a reference concentration in a given year to which deltaT is measured. However a better overall fit to 160 years of data [3] is given by deltaT = 2.5Ln(C/C0). This simple model has been criticised because it assumes that the Earth’s surface responds within a one year period to an increment in radiative forcing. For land surfaces this is certainly the case, whereas for oceans the response time is likely longer. This simple model for the CO2 greenhouse effect predicts about 2/3 of the observed warming (1.6). The remaining 1/3 may simply be due to other increases in greenhouse gases (CH4, CFCs)  as summarised in the 2007 IPCC report (1.6+ 0.9 = 2.5) [4].

Contribution of CH4 and CFCs radiative forcings compared to CO2

Contribution of CH4 and CFCs radiative forcings compared to CO2, IPCC 2007

Others may disagree with this interpretation. The objective of this post, however,  is not to argue one way or the other on this, but rather to concentrate on the  regular multi-decadal variation that is clearly  apparent in the data over and above any long term trends.  This regular oscillation is clearly evident below.

Evidence of a 60 year oscillation in global temperature data

Evidence of a 60 year oscillation in global temperature data

The dashed red curve is the fit deltaT=2.5Ln(C/C0). The measured CO2 concentration data is plotted on the right hand axis and is used for the calculation of the increment in temperature using 1850 as the reference value C0 (290 ppm). The residual data clearly shows the existence of an additional  regular variation of about 0.1C with a time period of about 60 years. A much better overall fit to the data is given by the solid red curve which adds  a 60 year harmonic with phase zeroed at 1865. The best fit solid curve  is

DeltaT(year) = -0.34 + 2.5ln(CO2(year)/290) + 0.1sin(0.105*(year-1860))  – equation 1.

By comparing differences in independent measurements of  deltaT  between ground and satellite –  HadCrut and UAH[5]  since 1979 one can get an estimate of the errors for  a single yearly value. This error  is  about 0.1C. Using this error estimate we can then compare just how well the calculated red curve fits  the 160 year HadCrut series by calculating Chi**2 of equation 1. to the data.  When this is done we get a total Chi-squared   value of 165 for 157 degrees of freedom  (or Chisquared/Ndf approx= 1.0) showing that the curve is rather a good fir with a high confidence level. This also strongly supports the view that the oscillation is a real effect.

Several others have previously proposed and discussed the existence of a multi-decadel variation in temperature [see discussion ref 1]. Many argue that the  oceans must be the vehicle of such a long term temperature effect since they alone have the required heat capacity. However the primary cause for  such a variation in ocean temperature is uncertain. It is natural to look for an astronomical origin of this effect as it seems to be global rather than say restricted to  just one region. Only the Sun and the Moon are likely to have sufficient direct effect on Earth to be possible  causes, but neither display any obvious 60 year cycle.

The 11 year cycle of sunspot activity causes a small but periodic variation in solar radiance on the Earth’s surface. The magnitude of this variation is about 0.1% of the  radiant energy on the Earth’s surface or a variable forcing of about 0.2 watts/m2. One can estimate that such an effect has capacity to  vary surface temperatures by around 0.2 degrees C with an 11 year cycle.

Solar Variability ( from Wikipedia)

Solar Variability ( from Wikipedia)

The moon produces diurnal tides in the world’s oceans and strong tidal currents near continents. The orbit plane of the moon precesses between +5 and -5 degrees to the ecliptic plane with a time period of 18.6 years. The angle of the orbit plane to the Earth’s equator varies from 28 to 18 degrees and this changes the latitude of the maximum tidal bulge on an 18.6 year cycle. It seems reasonable to assume that this varies the ocean driven transfer of heat across the Earth [6]. Once more the expected periodicity is18.6 years and not the observed 60 year cycle. The magnitude of this effect is difficult to estimate. However let’s assume that it is capable of producing a 0.05 degree C global oscillation . We can then investigate what the combined effect of both the solar variability and the tidal variability would be.

Variation in Diurnal tides (from: Richard Ray, Goddard, Nasa)

Variation in Diurnal tides (from: Richard Ray, Goddard, Nasa)

To investigate this possibility I have taken  a linear superposition of two periodic oscillations coinciding with the actual minima and maxima observed years from the two graphs above.  Both effects interfering with each other  with the independently observed  phase differences  produce an interference curve  but not a simple periodic variation of 60 days as observed.

combined solar variability and 18.6 year tidal term

The presence of this oscillation in the data can explain the apparent rapid increase in  temperatures observed in the 1980s and 1990s. Since 2000 no significant increase in temperatures has been observed and a further 20 years of flat or decreasing temperatures can be expected. The enhanced greenhouse effect due to increasing CO2 levels is still present and a further increase can be expected from 2030 for a further 30 years. Taking the emission scenario A1B, (which is consistent with current rises), from the IPCC/CIESIN reports[4], and applying the  fitted data as described above to CO2 levels, predicts a temperature increase below 2C by 2100 but with periods of enhanced increase followed by flat periods, each of about 30 years.

Extrapolation to 2100 of fit described in text

Extrapolation to 2100 of fit described in text

90 years is a very long time both in technological and economic terms, so CO2 emissions are likely to be reduced anyway below the A1B scenario envisaged. As stated before I am not an expert on climate science and perhaps more complex feedback effects are also present.  However the data itself show no evidence to date of any additional feedbacks once this  natural 6 decade oscillation  is folded in to a logarithmic temperature rise from an  enhanced greenhouse effect. Only modest temperature rises are predicted in this case during this century.


1. A periodic oscillation of 0.1C in global surface temperature data has been isolated. It can be parameterised by deltaT = 0.1sin(2pi/60*(year-1860)

2. The rapid rise in temperatures observed during the 1980s and 1990s is probably caused by  this oscillation superimposed on an underlying greenhouse warming. From 2000 to 2010 temperatures have remained static and no further increase is expected for another 10 years  despite an expected increase in CO2 levels.

3. There is no single astronomical effect which can explain the 60 year time period. I have looked into the possibility that a  superposition of both the 11 year solar variability and the 18.6 year lunar tide could produce the observed 60 year oscillation. There is no convincing evidence that this is the case.


[1] Empirical evidence for a celestial origin of the climate oscillations and its implications . Scafetta,N. Journal of Atmospheric and Solar-Terrestrial Physics

[2] Hadcrut Temperature anomalies, http://www.cru.uea.ac.uk/cru/data/temperature/

[3] see http://clivebest.com/blog/?p=2241

[4] IPCC 2007 Working Group 1 Report: http://www.ipcc.ch/publications_and_data/ar4/wg1/en/contents.html

[5] http://vortex.nsstc.uah.edu/data/msu/t2lt/uahncdc.lt

[6] Richard Ray, GSFC, NASA, Decadal Climate Variability Is There a Tidal Connection? http://www.decvar.org/documents/hawaii_workshop/agenda/ray.pdf

This entry was posted in Climate Change, Physics and tagged , . Bookmark the permalink.

7 Responses to A 60 year oscillation in Global Temperature data and possible explanations

  1. jmrSudbury says:

    One thought was that the period was somewhere between 60 and 66 years depending on whether or not the AMO and PDO are in phase. For the peak near 1940, they were in phase. They are slowly going out of phase as the AMO is something like 5 years shorter.

    One thing that I wondered about as soon as I saw that 66 year period was whether or not the 22 year polar flip solar cycle had anything to do with it as it was one third the cycle length, but it just seems to be a coincidence.

    Last week, I happened across a graph I had not seen in a few years, but I did not bookmark it. It shows the temperature trend with the ocean variation removed. It was similar but better than the modicum of work I did on this subject. If I find it again, I will let you know. This graph is similar to your first graph above, except it projects to 2100, and I think it goes back farther, perhaps using a proxy. Its also different because it uses a more linear trend than your red dashed curve. It seems to me that the ends of your red dashed curve are both too high — the slope at the beginning is too shallow, and the slope at the end is too steep. Of course, another 60 or 120 years of data would correct that, but it seems no one wants to wait another 60 years.

    John M Reynolds

  2. jmrSudbury says:

    I found the graph. It is at the top of the page at the following link:


    John M Reynolds

  3. clive says:

    Yes – this graph is prophetic. When I started looking into all this stuff I remember reading his paper, thinking that being based in Alaska he would know whether hype about the retreating arctic ice cover was fact or spin.


  4. uglybb says:

    Ran across your site because I am trying to understand where this 60 year cycle thing originates.

    The earliest reference I can find to it is some old book “the Jupiter effect” and lots of pseudo-science quacks who want to connect a gravity force diectly from Jupiter to Earth with two body Newtonian calculations I assume to link in.

    I actually can’t find a scienctific or astronomical reason for a 60 years cycle to exist do you know the source of this belief?

  5. clive says:

    Reply uglybb: I don’t really undertsand the 60 year oscillation although it seems to be linked to the “Pacific Decadel Oscillation (PDO) and the Atlantic Multidecadal Oscillation (AMO). The origins of these is not really explained either. There is a paper here: Nicola Scafetta, Empirical Evidence for a celestial origin of the cimate oscillations and its implications. http://arxiv.org/pdf/1005.4639 which does seemingly link the oscillation to the orbit of jupiter. In the next post A Fit to Global Temperature Data I find that a fourier analysis shows there are 11 year, 9.3 year and 60 year terms present in the data.

  6. Pingback: Getting real with Climate Inertia and Feedbacks | Clive Best

  7. Pingback: Tidal component to CSALT | context/Earth

Leave a Reply