Abstract
A long-standing debate in active tectonics addresses how slip is accumulated through space and time along a given fault or fault section. This debate is in part still ongoing because of the lack of sufficiently long instrumental data that may constrain the recurrence characteristics of surface-rupturing earthquakes along individual faults. Geomorphic and stratigraphic records are used instead to constrain this behavior. Although geomorphic data frequently indicate slip accumulation via quasicharacteristic same-size offset increments, stratigraphic data indicate that earthquake timing observes a quasirandom distribution. Assuming that both observations are valid within their respective frameworks, I want to address here which recurrence model is able to reproduce this seemingly contradictory behavior. I further want to address how aleatory offset variability and epistemic measurement uncertainty affect our ability to resolve single-earthquake surface slip and along-fault slip-accumulation patterns. I use a statistical model that samples probability density functions (PDFs) for geomorphic marker formation (storm events), marker displacement (surface-rupturing earthquakes), and offset measurement, generating tectono-geomorphic catalogs to investigate which PDF combination consistently reproduces the above-mentioned field observations. Doing so, I find that neither a purely characteristic earthquake (CE) nor a Gutenberg-Richter (GR) earthquake recurrence model is able to consistently reproduce those field observations. A combination of both however, with moderate-size earthquakes following the GR model and large earthquakes following the CE model, is able to reproduce quasirandom earthquake recurrence times while simultaneously generating quasicharacteristic geomorphic offset increments. Along-fault slip accumulation is dominated by, but not exclusively linked to, the occurrence of similar-size large earthquakes. Further, the resolution potential of tectono-geomorphic records is dominantly affected by measurement uncertainty, that is, by the ability to correctly infer a geomorphic marker's pre-earthquake(s) morphology. Typically reported values for measurement uncertainty suggest that it is improbable to resolve more than five large earthquakes from tectono-geomorphic records.