Some Thoughts on Estimating Maximum Magnitude and Corner Magnitude

Abstract:

Mmax, the size of the largest possible earthquake in a region, is one of the least certain parameters used to assess seismic hazard. In a strict sense, Mmax can only be interpreted if you assume that the distribution of earthquake magnitudes has some upper limit, i.e., the truncated Gutenberg-Richter distribution. Unfortunately, you cannot estimate Mmax and its uncertainty using only the magnitudes recorded so far; you also need to incorporate independent data such as those from geodesy or paleoseismology, and these data have additional uncertainties. Rather than struggling with Mmax, some seismologists have suggested using a different parameterization of the magnitude distribution: one that is tapered rather than truncated, shifting the focus to the point at which the taper begins, the so-called “corner magnitude.”

In this presentation, we describe frequentist and Bayesian methods for estimating the corner magnitude; we demonstrate that you can obtain a confidence interval for the corner magnitude using one or more earthquake catalogs of sufficient length; we suggest that the corner magnitude varies with tectonic regime; and we explain why our results are different from those of previous studies.