Effects of subduction interface locking distributions on tsunami hazard: a case study on the Hikurangi/Tonga-Kermadec subduction zones

Abstract

Subduction zones have the greatest potential to generate large earthquakes and tsunamis. However, when undertaking Probabilistic Tsunami Hazard Assessments (PTHAs), subduction zones are a significant source of epistemic uncertainty. Therefore, understanding how the spatial distribution of elastic strain accumulation on the subduction interface influences the tsunami hazard is important for providing comprehensive hazard assessments, as well as quantifying uncertainty. This is especially important if the spatial locking distribution is undefined, and if it changes through time. Physics-based earthquake simulators allow different interpretations of the subduction interface locking distribution to be modelled, and how this influences the long-term seismicity, and tsunami hazard, can be explored. Using three physics-based synthetic earthquake catalogues, generated by the earthquake simulator RSQSim, we analysed the tsunami hazard in Aotearoa/New Zealand. Three alternative representations of the subduction interface locking distribution along the Hikurangi Subduction Margin and the Tonga-Kermadec Subduction Zone were specified in the simulator to generate the catalogues. We modelled the tsunamis generated by MW>8.0 earthquakes from each of the catalogues and undertook PTHAs. These assessments show that patches of high slip-deficit, both along strike and dip of the subduction interface, increase the tsunami hazard at the coast. Locking along the shallowest segments of the subduction interface also significantly increases the tsunami hazard. Our results show that careful consideration of the locking distribution in physical models is necessary before using them for PTHAs. They also show that by analysing multiple physical models of subduction zones, uncertainty in hazard assessments caused by the unresolved interface properties can also begin to be quantified.