Smith & Bailey (2017a) have raised concerns over the methodologies we applied for time series analysis in our study of the cyclic Devonian lacustrine deposits of northern Scotland (Andrews et al. 2016). Similar concerns have been raised on a number of occasions for similar studies (Bailey et al. 2009 on Kemp & Coe 2007; Smith & Bailey 2017c on Howe et al. 2016; Smith et al. 2016 on Fang et al. 2015 and Smith & Bailey 2017b on Perdiou et al. 2016). Indeed, a recent in-depth and technically robust rebuttal of a Smith and Bailey comment by Hinnov et al. (2016) largely negates the concerns raised about our own work. Repetition of this debate is not considered worthwhile here. However, the important point to be made concerns the nature of the use of time series analysis. In our study, time series analysis is utilized as a tool to examine data from a succession known to contain clearly developed, and widely recognized, cyclic deposits (Crampton & Carruthers 1914, Donovan 1980, Andrews & Trewin 2010). Detailed sedimentological analysis of well exposed outcrops was used to characterize these successions, and illustrate their cyclic nature. The best exposed section was then logged using a hand held gamma meter to investigate if the cyclic successions displayed a distinctive petrophysical character. This having been demonstrated, nearshore well data (UK 11/25-2 ST1) was examined to provide a more continuous stratigraphic record through the succession. Time series analysis was then applied and provided evidence for cycles of similar thickness to those found in the onshore successions at their respective stratigraphic levels. Therefore, the use of time series analysis was not to prove that cycles exist, which Smith and Bailey suggest, but rather to investigate their expression in offshore data.
This approach is in accord with the conclusions of Kemp (2016), and also referenced by Smith and Bailey, that time series analysis used in isolation to identify astronomical forcing should be treated with extreme caution. Instead we present strong independent evidence for the presence of cyclicity, and astronomical forcing, and then use time series analysis to look for these known cycles in adjacent datasets.
Scientific editing by Stuart Jones
- © 2018 The Author(s)
