AUTOMATED DETECTION AND CHARACTERIZATION OF SWARMS AND MAINSHOCK-AFTERSHOCK SEQUENCES IN SOUTHERN MEXICO (Invited Presentation)

Earthquakes in Mexico are frequent and dangerous. Over the last decade, there have been a dozen major earthquakes including two larger than M7.5. The Mexican subduction zone is considered a natural laboratory for studying slip processes due to the relatively short trench-to-coast distance which brings broad portions of the seismogenic megathrust inland. The recent discovery of earthquake swarms and strike-slip slow slip events associated with a >1000 km sliver fault provides an opportunity to pursue detailed characterizations of relationships between seismicity and aseismic slip on a complex plate boundary system. Using an automated detection algorithm that identifies clusters of events using the nearest neighbor distances in the space-time-energy domain (Zaliapin and Ben-Zion, 2013) we focus on sequences from 2012 to 2020. We developed an automated characterization algorithm to characterize the sequences on a spectrum of more swarm-like to more aftershock-like, which reduces time and bias due to the number of events and human perception. The automated algorithm using quantitative forms of these attributes: (1) the magnitude difference between the largest event and the next largest events, (2) the percentage of the sequence after the largest event, (3) the slope of seismicity rate over time, (4) the magnitude range divided by the number of events in the sequence and, (5) the rate of maximum magnitude decay over time. Overall, the automated characterization method yields similar results to manual characterization and is effective at identifying average properties when there are discrepancies among manual ratings for complex sequences. Surprisingly we found more swarms than aftershock sequences despite the prominence of large megathrust mainshock-aftershock sequences over the past decade. Temporally, some swarm sequences show an interesting pattern where the seismicity shuts on or off depending on nearby megathrust activity. Spatially, swarms help define a vertically dipping fault in the upper plate indicating the sliver fault may be closer to the coast than previously thought. We anticipate standardizing the characterization process will provide opportunities for more in depth studies of seismic sequence types and their causes.