“Barycentric” influence of the planets on the sun is just statistically insignificant, and a previous paper that claims to find a signal in isotopic records is proven to be nothing more than a statistical artifact.
In 2012, Astronomy & Astrophysics published a statistical study of the isotopic records of solar activity, in which Abreu et al. claimed that there is evidence of planetary influence on solar activity. A&A is publishing a new analysis of these isotopic data by Cameron and Schüssler. It corrects technical errors in the statistical tests performed by Abreu et al.
They find no evidence of any planetary effect on solar activity.
In a new paper published in A&A, R. Cameron and M. Schüssler, however, identify subtle technical errors in the statistical tests performed by Abreu et al. Correcting these errors reduces the statistical significance by many orders of magnitude to values consistent with a pure chance coincidence. The quasi-periods in the isotope data therefore provide no evidence that there is any planetary effect on solar activity.
The paper (h/t to Dr. Leif Svalgaard)
No evidence for planetary influence on solar activity
R. H. Cameron and M. Schüssler
Max-Planck-Institut für Sonnensystemforschung, Max-Planck-Str. 2, 37191 Katlenburg-Lindau, Germany e-mail: [cameron;schuessler]@mps.mpg.de
Received 16 April 2013 / Accepted 24 July 2013
Context. Recently, Abreu et al. (2012, A&A. 548, A88) proposed a long-term modulation of solar activity through tidal eﬀects exerted by the planets. This claim is based upon a comparison of (pseudo-)periodicities derived from records of cosmogenic isotopes with those arising from planetary torques on an ellipsoidally deformed Sun.
Aims. We examined the statistical significance of the reported similarity of the periods.
Methods. The tests carried out by Abreu et al. were repeated with artificial records of solar activity in the form of white or red noise. The tests were corrected for errors in the noise definition as well as in the apodisation and filtering of the random series.
Results. The corrected tests provide probabilities for chance coincidence that are higher than those claimed by Abreu et al. by about 3 and 8 orders of magnitude for white and red noise, respectively. For an unbiased choice of the width of the frequency bins used for the test (a constant multiple of the frequency resolution) the probabilities increase by another two orders of magnitude to 7.5% for red noise and 22% for white noise.
Conclusions. The apparent agreement between the periodicities in records of cosmogenic isotopes as proxies for solar activity and planetary torques is statistically insignificant. There is no evidence for a planetary influence on solar activity.
The statistical test proposed by Abreu et al. (2012), a comparison of the coincidences of spectral peaks from time series of planetary torques and cosmogenic isotopes (taken as a proxy for solar activity in the past) with red and white noise, is logically unable to substantiate a causal relation between solar activity and planetary orbits. Furthermore, the execution of the test contains severe technical errors in the generation and in the treatment of the random series. Correction of these errors and removal of the bias introduced by the tayloring of the spectral windows a posteriori leads to probabilities for period coincidences by chance of 22% for red noise and 7.5% for white noise. The coincidences reported in Abreu et al. (2012) are therefore consistent with both white and red noise.
Owing to our lack of understanding of the solar dynamo mechanism, red or white noise are only one of many possible representations of its variability in the period range between 40 and 600 years in the absence of external eﬀects. This is why the test of A2012 is logically incapable of providing statistical evidence in favour of a planetary influence. Alternatively one could consider the probability that a planetary system selected randomly from the set of all possible solar systems would have periods matching those in the cosmogenic records. In the absence of a quantitative understanding of the statistical properties of the set of possible solar systems to draw from, the comparison could again, at best, rule out a particular model of the probability distribution of planetary systems. Here we have shown that the test in A2012 does not exclude that the peaks in the range from 40 to 600 years in the planetary forcing are drawn from a distribution of red or white noise.
We conclude that the data considered by A2012 do not pro- vide statistically significant evidence for an eﬀect of the planets on solar activity.