CouldM_maxBe the Same for All Stable Continental Regions?

Abstract

In probabilistic seismic-hazard assessment for stable continental regions (SCRs), the maximum magnitude M max truncating the earthquake magnitude-frequency distribution is commonly based on Bayesian updating of a global prior distribution derived from the distribution of observed M max in superdomains (groups of tectonically similar domains). We use randomly simulated earthquake catalogs to test if this observed superdomain M max distribution could also be explained by a global uniform M max value in SCRs, given our limited periods of observation. Using published average recurrence parameters per continent, catalog completeness thresholds for different regions within each continent, and assuming a Poisson temporal occurrence model, we simulate 10,000 random catalogs for each SCR domain, combine them into superdomain catalogs, and determine the largest sampled magnitude and the number of sampled earthquakes in each. Imposing an M max of 7.9, the largest magnitude observed in SCRs to date, and catalog lengths similar to those presently available, we obtain superdomain M max distributions similar to that observed. Hence, we cannot presently distinguish whether M max in SCRs is spatially variable or uniform. As a result, using a single value of M max in seismichazard analyses for all SCRs might make sense. Simulations with larger M max and longer catalogs confirm that catalog length is the limiting factor in our knowledge of M max. Online Material: Table of data available for the 255 stable continental region domains.