Modelling seismic catalogues by cascade models: Do we need long-term magnitude correlations?

Abstract

We consider sequences of earthquakes from northern and southern California. We study these data sets for long-term persistence using the concept of natural time and fluctuation analyses. We also construct a simulation model for regional seismicity using random background seismicity and the BASS model for aftershock occurrence. We include in the simulation model corrections for long-term persistence in the background seismicity and for missing data early in aftershock sequences. We find excellent agreement between the California data and the simulations indicating significant long-term correlations in earthquake magnitudes.