Over at WUWT, Willis has been up to his usual trick of mangling data in a vain attempt to discredit scientists who find strong links between the Sun’s variation and Earth’s weather and climatic patterns. This time it’s Le Mouel et al who get the treatment in his ‘analysis’ of their 2010 paper “Solar forcing of the semi‐annual variation of length‐of‐day”
As usual, Willis gets things upside down. He asks: “So … is there a correlation between sunspots and zonal wind speeds?” The answer to which is no, and the paper’s authors never claimed there was. However, as Fig 1 of Le Mouel et al’s paper shows, there is a strong anti-correlation between solar variation and the semi-annual variation of Length of Day (LOD) which is itself well correlated with changes in zonal wind speeds. For obvious reasons, Willis doesn’t show his readers Fig 1, reproduced here for your academic study.
window is used. (a) Comparison of the semiannual amplitude of lod with the sunspot number WN (red); WN is both
reversed in sign and offset by one year (see text). (b) Comparison of the detrended semiannual amplitude of lod (blue) with
the sunspot number WN (red); WN is reversed in sign and offset by one year. (c) Comparison of the semiannual amplitude
of lod (blue) with galactic cosmic ray flux GCR (red); GCR is neither reversed in sign nor offset (see text).
Nor does he mention the strong direct correlation between galactic cosmic rays (GCR) and semi-annual LOD variation, probably because his fellow warmist Leif Svalgaard wouldn’t like that at all, since it supports Henrik Svensmark’s GCR-cloud hypothesis.
Willis makes other spurious complaints in his attempt to discredit the author’s methods. For example he can’t see any reason why they would use a four year sliding window on the LOD data.
In a word (or acronym in this case), the answer is ENSO. The strongest weather oscillation on Earth affects zonal wind speed and has a 3.7 year period on average: it’s one third of the solar cycle length. The authors even go to the trouble of testing the data with one and two year sliding window periods to ensure they are not fooling themselves in their findings:
Discrepancies between phase estimates should probably be
partly attributed to the filtering processes. Our estimates do
not change when windows of 1 or 2 years rather than 4 years
are used, although of course short period noise becomes
larger.
I won’t bother trying to work out when Willis says “I used a CEEMD analysis which breaks out the underlying frequencies of the two signals” whether he’s referring to the semi-annual variation the authors are interested in or the annual variation he graphs immediately above which is mainly due to Earth’s transition from aphelion to perihelion and back (note that the troughs fall at the end of June and start of January). It’s all just obfuscation with periodograms when he could simply have discussed the excellent and clear data displayed in Le Mouel et al’s Fig 1.
I highly recommend reading the original paper. It’s commendably short and clear.
via Tallbloke’s Talkshop
December 28, 2021, by tallbloke
You must be logged in to post a comment.