Bulletin of the Seismological Society of America, cilt.114, sa.6, ss.3127-3142, 2024 (SCI-Expanded)
Complex fault geometries with multiple sets of inclined active faults pose a challenge to the accurate representation of fault-based seismic sources in probabilistic seismic hazard assessment (PSHA). In the absence of slip rates associated with fault segments and the presence of sparse seismic and geodetic data, the estimation of segment-specific activity rate includes significant uncertainty. This study proposes a comprehensive procedure for defining the segment-specific activity rates and associated uncertainties in fault-based PSHA for extensional tectonic regimes and applies the procedure in the northern margin of Western Anatolian Extensional Province. The seismic sources are modeled as rupture systems with individual fault segments using the connections between available active fault traces, first-order geological complexities, earthquake catalog, and focal mechanism solutions. Alternatives to estimate and partition segment-specific activity rates based on annual slip rate, seismicity rate, and moment rate are explored. Each alternative is implemented in PSHA, and the results are compared in terms of peak ground acceleration (PGA) maps for a 475-year return period. A comparison of the 475 yr PGA maps showed that the activity rates based on annual slip rates translate into higher hazard estimates and a more uniform distribution of PGA; whereas, the activity rates based on seismicity and moment rate result in a PGA distribution that is more sensitive to the occurrence and location of previous large-magnitude events. The approach utilized to partition activity rates among parallel segments has a noticeable effect in the areas where highly asymmetric fault activity is inferred from morphology. Hence, alternative approaches for estimation and partition of activity rates are combined to model the epistemic uncertainty in segment-specific activity rates, and a repeatable procedure is developed to build fault-based seismic source models in the presence of complex fault geometries.