Spurious periods in the terrestrial impact crater record
L. Jetsu 1,2 and
J. Pelt 3,4
Received 8 September 1999 / Accepted 21 October 1999
We present a simple solution to the controversy over periodicity in the ages of terrestrial impact craters and the epochs of mass extinctions of species. The first evidence for a 28.4 million year cycle in catastrophic impacts on Earth was presented in 1984. Our re-examination of this earlier Fourier power spectrum analysis reveals that the rounding of the impact crater data distorted the Monte Carlo significance estimates obtained for this cycle. This conclusion is confirmed by theoretical significance estimates with the Fourier analysis, as well as by both theoretical and Monte Carlo significance estimates with the Rayleigh method. We also apply other time series analysis methods to six subsamples of the currently available more extensive impact crater record and one sample of mass extinction epochs. This analysis reveals the spurious "human-signal" induced by rounding. We demonstrate how the data rounding interferes with periodicity analysis and enhances artificial periodicities between 10 and 100 million years. Only integer periodicities connected to irregular multimodal phase distributions reach a significance of 0.001 or 0.01. We detect no real periodicity in the ages of terrestrial impact craters, nor in the epochs of mass extinctions of species.
Key words: Earth comets: general minor planets, asteroids methods: statistical Galaxy: solar neighbourhood
Send offprint requests to: L. Jetsu
This article contains no SIMBAD objects.
© European Southern Observatory (ESO) 2000
Online publication: December 8, 1999