Detection of Significant Seismic Quiescence Patterns in the Mexican Subduction Zone Using Extended Schreider Algorithms

This study investigates the implementation of Schreider's quiescence algorithm and two variants that utilize spatiotemporal data to identify patterns of seismic quiescence. These patterns are of particular interest as they may serve as precursors to major seismic events, specifically large earthquakes (M>7), within the Mexican subduction zone associated with the Cocos Plate. We identify two characteristic stages: the α-stage, where a notable deviation is observed in the Schreider convolutions, and the β-stage, where the convolutions return to their background levels. In addition, we identify that the Schreider algorithm cannot discern quiescence patterns when earthquakes M>7 are too close in space and time. Consequently, we explore the behavior of the convolutions in three cases where the algorithm is restarted after the mainshocks, and we find apparent advantages for the spatiotemporal variants of the convolutions. The findings contribute to a more profound understanding of the stages preceding large subduction earthquakes and aid in the identification of precursor patterns in this region.