Do Large Earthquakes Occur at Regular Intervals Through Time? A Perspective From the Geologic Record

Abstract

We analyzed a catalog of 31 published earthquake chronologies to assess the commonality of quasiperiodic earthquake recurrence across a range of fault types and tectonic settings. The statistical approach we employ differs from previous methods in that it explicitly incorporates numeric uncertainties in the earthquake chronologies while recognizing that random sequences of events (against which the chronologies are tested) may appear to be less disordered over the short time scales typical of most published records. Our results show that 58% of the chronologies support an interpretation of quasiperiodic recurrence (probability of random recurrence < 10%). These include strike‐slip, normal, and reverse faults in both plate‐boundary and intraplate environments, which exhibit evidence for quasiperiodic recurrence with comparable frequency. We conclude that quasiperiodic failure is likely the norm for faults in the seismogenic crust and that stress renewal is a first‐order control on fault rupture across a wide range of tectonic settings.

Periodicity and Clustering in the Long‐Term Earthquake Record

“…Our results show that large earthquakes recur more regularly than expected from exponentially distributed interevent times (i.e., a Poisson process), instead showing weakly periodic behavior similar to that expected from a renewal process (Figure 4). These findings are consistent with those of Sykes and Menke (2006), Moernaut et al (2018), and Williams, Davis, and Goodwin (2019). The majority of records also have memory coefficients indistinguishable from zero; for most faults, the paleoearthquake record does not show correlation between successive interevent times or clustering.…”

“…Interevent time variability does not scale with long‐term activity rate for faults with moderate‐to‐high activity rates (Figures 2a, 2e, and 2f), consistent with the findings of Williams, Davis, and Goodwin (2019); however, burstiness increases for annual rates < ~2 × 10 −4 , with the greatest variability in interevent times observed on low activity‐rate reverse faults (Figures 2e and 2f). We can reject a renewal process model for low activity‐rate faults, but not random recurrence (i.e., a Poisson process; Figures 3c and 3i).…”

“…Previous studies have suggested a correlation between slip rate and periodicity (Nicol et al, 2016; Yuan et al, 2018); however, in their global analysis, Williams, Davis, and Goodwin (2019) concluded that most faults rupture regularly regardless of slip rate, tectonic region, or faulting style. Less regular earthquake recurrence on some faults may be a consequence of factors other than the long‐term tectonic loading rate.…”

“…Yet observational confirmation of elastic rebound theory as a general model has remained challenging. Despite geological observations supporting quasiperiodic recurrence on some faults (e.g., Berryman et al, 2012; Scharer et al, 2010), many paleoearthquake records are too short or uncertain to confidently exclude random earthquake occurrence (Williams, Davis, & Goodwin, 2019). Furthermore, instrumental observations of earthquake sequences often show complex spatiotemporal patterns characterized by short periods of heightened activity (clusters or “bursts”).…”

U‐Th Dating of Syntectonic Calcite Veins Reveals the Dynamic Nature of Fracture Cementation and Healing in Faults

“…Thus, the fracture sealing rates documented here indicate that fractures in the Loma Blanca damage zone could seal fully with cements between successive ruptures, at least in the shallow crust. These rates, however, would be too slow to heal fractures of similar aperture in faster‐moving plate‐boundary faults, which typically experience large ruptures with a periodicity of several hundred years (Williams, Davis, & Goodwin, ). It is important to note that the rates we document for calcite growth are specific to the examined lithology, tectonic setting, crustal depth, and hydrogeological regime.…”