The differences in b values between foreshock and aftershock sequences can be shown to be a statistically significant property of real earthquake sequences if a sufficiently large number of cases is considered, i.e., if the catalogs are long enough. These differences depend on the particulars of the data processing procedures used to define the sequences, such as space-time windowing and definition of cutoff magnitude thresholds for completeness of catalogs, as well as other magnitude cutoffs. They are also a statistically significant property of long artificial, or synthetic catalogs in which all subsequent earthquakes are modeled as resulting from a stochastic time delay applied to predecessor earthquakes. As a consequence, we make the conjecture that the difference in b values is due to subtle asymmetries in data processing procedures rather than differences in the physical environment before and after a large earthquake.

First Page Preview

First page PDF preview
You do not currently have access to this article.