RECONCILING SEISMOLOGIC AND PALEOSEISMOLOGIC VIEWS OF EARTHQUAKE RECURRENCE VIA LONG-TERM FAULT MEMORY

Meyers and Hamilton's (1964) observation that the 1959 Hebgen Lake, Montana, earthquake reflected repeated deformation in what we now recognize as part of the Pacific-North America plate boundary zone illustrated the paleoseismic approach to earthquake recurrence. Since the 1906 San Francisco earthquake, seismology's dominant paradigm has been the earthquake cycle, in which strain accumulates between large earthquakes due to motion across a locked fault and is released by fault slip in earthquakes. Over time, this process should cause quasiperiodic earthquakes and steady accumulation of displacement. In these models, the probability of a large earthquake increases with time until one occurs, at which point the probability drops to zero and the cycle begins again. The fault “remembers” only the last event, when the probability was renewed - reset to zero. Because the probability depends only on the time since the past one, the fault has only “short-term memory." However, long paleoseismic records often show large earthquakes occurring in supercycles, sequences of temporal clusters of seismicity, cumulative displacement, and cumulative strain release. Because supercycles and associated earthquake clusters are not described by commonly used models of earthquake recurrence, we are exploring an alternative model, Long Term Fault Memory (LTFM). In LTFM, the probability of an earthquake grows with time at a steady rate, simulating steady strain accumulation. The probability drops after an earthquake, but not necessarily to zero as in the traditional earthquake cycle model, simulating partial strain release. Thus the fault’s history over multiple cycles influences the future probability of an earthquake. We use LTFM to simulate paleoseismic records from subduction zones, transform faults, and intraplate regions. In some portions of the simulated earthquake history, events can appear quasi-periodic, while at other times, the events can appear more Poissonian. Hence a given paleoseismic or instrumental record may not reflect the long-term seismicity of a fault, which has important implications for hazard assessment.