Kulkarni et al. (2013; hereafter, Kulkarni2013) present a detailed statistical analysis of records of possible repeat great earthquakes on the Cascadia subduction zone, offshore of Vancouver Island, Washington, Oregon, and northern California. The data set they analyze is derived from turbidite cores collected off the Washington, Oregon, and northern California coasts (Goldfinger et al., 2012; hereafter, Goldfinger2012). They conclude that there is a 65% probability that the data are in fact clustered, as previous authors have speculated. This means that the cluster hypothesis does not meet the usual 95% threshold for acceptance of a hypothesis, although they then engage in a fairly complicated second‐order analysis which concludes the data are significantly clustered. They acknowledge that this may not be a unique interpretation, and essentially invite alternative models.
We recognize that the total reliance on the interpretations of Goldfinger et al. (2012) begs the question, would we reach the same conclusions if the turbidite data were interpreted differently by another set of eyes? Although the Goldfinger et al. (2012) study has undergone extensive review, that question cannot be answered until future interpretations become available. (Kulkarni2013, p. 16)
What follows is my attempt to present an alternative interpretation, one with rather different implications for the seismic hazard in the Pacific Northwest. (For reference, the top half of fig. 2 from Kulkarni2013 is plotted below as Fig. 1.)
In summary, my argument …