Are Asperities Persistent Over Time? Observations From The Mexican Subduction Zone

Abstract:

Fault slip during an earthquake is observed to be highly heterogeneous, with areas of large slip interspersed with areas of smaller or even no slip. The cause of the heterogeneity is not clear. One hypothesis is that the frictional properties on the fault are heterogeneous. The parts of the rupture surface that have large slip during earthquakes are coupled more strongly, whereas the areas in between and around creep continuously or episodically. The continuously or episodically creeping areas can partly release strain energy through aseismic slip during the interseismic period, resulting in relatively lower prestress than on the coupled areas. This would lead to subsequent earthquakes having large slip in the same place, or persistent asperities. A second hypothesis is that in the absence of creeping sections, the prestress is governed mainly by the accumulative stress change associated with previous earthquakes. Assuming homogeneous frictional properties on the fault, a larger prestress results in larger slip, i.e. the next earthquake may have large slip where there was little or no slip in the previous earthquake, which translates to non-persistent asperities. The study of earthquake cycles are hampered by short time period for which high quality, broadband seismological and accelerographic records, needed for detailed studies of slip distributions, are available. The earthquake cycle in the Mexican subduction zone is relatively short, with about 30 years between large events in many places. We are therefore entering a period for which we have good records for two or more subsequent events occurring in the same segment of the subduction zone, in addition to historical records. We have analysed records for most large earthquakes within the last 110 years, and compared the waveforms, to estimate the similarity of waveforms of events breaking the same segment. In general we find that the waveforms are different, indicating that the same fault patch was not broken. However, near the Guerrero-Oaxaca border we find one region that seems to have broken three times in the last 100 year. Our observations therefore indicate that both processes are important, perhaps with varying relative importance depending on the region. In other words, some asperities are persistent over time, whereas others are not.