One of the reasons I decided to undertake doctoral studies with Bob Dott at Wisconsin was his interest in evaluating so-called cyclothems in the Rocky Mountain region. An early reading assignment was the paper by F. Zeller, titled "Cycles and Psychology". Bob stressed that order in stratigraphic successions ("cyclicity") needs to be demonstrated using objective methods/criteria and not just accepted a priori. This simple commitment, as well as other related ideas, has influenced my pursuit of recognizing stratigraphic order.

I have deployed various quantitative methods in the assessment of stratigraphic order. While at Wisconsin, the Permo-Pennsylvanian Amsden Formation and Tensleep Sandstone of Wyoming were studied using a combination of the statistical methods known as log-linear modeling and Markov chain analysis. Log-linear modeling was also used to assess the classic Illinois basin cyclothem against the original measured sections compiled by H. Wanless and J. Weller. Both examples are clearly not random stratigraphic successions. However, the former example contains geographic and stratigraphic variations (i.e. are not fully stationary throughout their occurrence), and the later indicates a poor fit for the classic cyclothem. While this method works well on discrete data such as facies transitions, it is not well suited for continuous data such as wireline logs. In more recent years, I have been using partitioned Fourier transform and wavelet transform analyses to assess order or patterns in wireline logs. Wind-ripple vs. non-wind-ripple Tensleep reservoir sandstones were discriminated based on spatial frequency spectra of microresistivity logs. Gamma ray logs of Pennsylvanian successions from the Paradox basin were also analyzed with the results yielding stratigraphic and geographic variations in the hierarchy of "cyclic" sedimentation.