In our article Kulkarni et al. (2013), we favored a clustered model for great Cascadia subduction zone (CSZ) earthquakes and our statistical analysis resulted in a probability of 0.65 that clustering was present in the turbidite record. The CSZ clustering analysis was originally motivated by the need to develop a CSZ logic tree for use in probabilistic seismic‐hazard analysis. In recognition of the two principal models that could explain the pattern of great earthquake occurrence observed in the turbidite record, we gave weights of 0.65 and 0.35 to clustered versus quasi‐periodic behavior in the logic tree, respectively (Wong et al., 2014).
In Lindh (2016), Al Lindh questioned the validity of the turbidite data and our analysis of temporal clustering and states that the CSZ actually behaves in a quasi‐periodic behavior. He summarizes his argument in four points, and we wish to respond to each point in the following. We confine our responses to the technical issues raised by Lindh. With regards to his discussion on seismic risk in the Pacific Northwest, we support his views. Certainly, understanding the behavior of the CSZ, particularly its recurrence, is a critical element in earthquake‐risk reduction in the Pacific Northwest.
Little Evidence for Clustering of Great Earthquakes
We appreciate Lindh for pointing out the more recent studies that indicate that for a few of the examples we cited, clustering of full‐rupture earthquakes may not be characteristic of the behavior that was previously suggested. However, the point we were trying to make is that earthquake clustering, regardless of scale or magnitude, is a behavior that is observed along plate boundaries and crustal faults worldwide. There are physical models that can explain temporal clustering, and the quasi‐periodic behavior favored by Lindh is unusual over long time spans. Many factors can come into play that can alter the long‐term behavior of a …