Y.W.M. Liao, B.Fry, A.Howell, A.Nicol, C.Williams, C. Rollins (2022) The role of frictional heterogeneities in the earthquake cycle. Geoscience Society of New Zealand Annual Conference 2022: Programme & Abstracts Volume. Geoscience Society of New Zealand Miscellaneous Publication 161A. Geoscience Society of New Zealand, Wellington, pp. 88. ISBN (online): 978-0-473-66216-5 ISSN (online): 2230-4495. https://gsnz.org.nz/assets/Uploads/Shop/Products/GSNZ_annual_conference/MP161_2022_Palmerston_North/MP161A_2022_GSNZ_conference_Palmerston_North_Abstract_Volume.pdf
Determination of earthquake source models for seismic and tsunami hazard assessments is difficult and highly uncertain. Information such as recurrence interval and origin of earthquakes and probability of multiple-segment rupture etc. is often unknown, especially for large earthquakes (M>7.0). Earthquake simulators such as RSQSim, based on rate-and-state friction models, generate long-term synthetic earthquake catalogues on a system of known faults. They also provide a clear pathway to future generations of seismic and tsunami hazard models in which ground motions are calculated by modelling the seismic wavefield, having the ability to include variability in the earthquake source process. Importantly, modelling the earthquake cycle also provides a previously elusive pathway toward the ambitious goal of probabilistic tsunami risk modelling.
Here we test the effects of varying a priori input parameters during RSQSim modelling. We explore the effects of different initial stress models and rate-and-state constants (a and b) for earthquake cycle simulations of the Hikurangi-Kermadec subduction zone, one of the most observationally undersampled earthquake source regions on the planet. We compare our results with the magnitude frequency distribution of the observed earthquake catalogue and empirical scaling laws as a first-order test.
Using variable values instead of a uniform value for initial stress on the fault plane generates more realistic synthetic earthquake catalogues. Co-seismic ruptures become less characteristic as well when the heterogeneity of initial stress is considered. The comparison of rupture area scaling, coseismic slip and stress drop obtained simultaneously with the synthetic catalogues indicates that smaller b (with fixed a) could result in smaller co-seismic slips and stress drops and scaling consistent with empirical scaling laws. Our results provide a more objective way to assess earthquake-related hazard and hope for hazard assessment and emergency response planning.