Posted by: paulgarner | February 6, 2012

Sedimentary cyclicity: time for a rethink?

One of the burgeoning research areas in the field of sedimentology has been the detection in stratigraphic sequences of cyclicity attributed to orbital forcing. The main orbital cycles that might leave a signature in the rock record include the 100,000-year eccentricity cycle, the 41,000-year tilt cycle and the 22,000-year precession cycle, collectively termed Milankovitch (M-) cycles.

Most of the published studies involve the measurement of rock properties at regularly spaced intervals through a stratigraphic section. Statistical analyses are then applied to detect meaningful cyclic variations that can be matched to the orbital cycles. M-cycles are being sought because they would allow chronological calibration of stratigraphic sections with a much finer resolution than that permitted by radiometric dating.

The apparent detection of M-cyclicity in the sedimentary record poses an obvious challenge to the creationist model of earth history which says that these sediments accumulated over time scales much shorter than those over which M-cycles operate. Some critics of creationism have highlighted specific instances of this problem, such as Glenn Morton writing on cyclicity in the Eocene Green River Formation.

Over the years, scepticism has occasionally been expressed about the detection of M-cycles in stratigraphic data. Back in 1996, the claimed Milankovitch origin of six hundred cycles in Triassic platform carbonates of the Italian Alps was called into question by new biostratigraphic and radiometric data. In 2009, a cautionary review article by Robin Bailey was published in Terra Nova. Summarising some of his concerns in a ‘Soapbox’ column in Geoscientist magazine last year (Vol. 21, No. 5, June 2011, p.11), Bailey pointed out the circularity of reasoning that is often involved in cyclostratigraphic studies:

Where ‘cycle’ refers to repeated facies, the intervals between similar beds and the thicknesses of those beds usually vary unpredictably. Clearly, then, the only way that repetitions can achieve ‘cycle’ status, is by showing that they are all the outcome of M-forcing. Counting and ‘tuning’ aim to show that the number of cycles in a section of known time-span is consistent with this theory. For this to work the section must be free of significant hiatuses – something that can only be verified by a detailed time calibration. But in practice, the theory chases its own tail, concluding that there are no gaps, and M-forcing therefore proved, if the expected number of irregular cycles is counted!

Furthermore, Bailey pointed out that random data noise can give spectral peaks very similar to those generated by ‘real’ cycles and that, in the absence of accurate statistical models to handle the noise, false cyclicities are likely to be recorded. He concluded by emphasising that cycles reported in the literature were ‘mostly suspect’ and provided ‘unreliable time calibrations’.

Now, Vaughan, Bailey and Smith have published a detailed re-analysis of four datasets for which cyclicity had previously been claimed. They were able to reproduce the cyclicity detection in only one case. Their conclusion is that popular methods of cycle detection via spectral analysis are flawed and that the published literature on stratigraphic cycles is dominated by ‘false positives’. In a short summary published in Geoscientist under the heading ‘Cycle pathology’ (Vol. 22, No. 1, February 2012, p.7), Vaughan concludes

that the majority of published reports of regular Milankovitch-band cyclicities (eg in Paleogene, Mesozoic, and older strata), and the resultant astrochronological time calibrations, are based on statistically unsound detections.

These are interesting developments and it seems to me that the field is wide open for creationists and others to propose alternative mechanisms for stratigraphic cyclicity that do not depend on Milankovitch forcing.



%d bloggers like this: