Abstract
Rapid characterisation of large-magnitude earthquakes is essential for Tsunami Early Warning (TEW). In the Southwest Pacific, several subduction zones (Hikurangi-Kermadec-Tonga, Puysegur, and Vanuatu) have the potential to host large-magnitude (Mw7+) subduction earthquakes. Such events can create a substantial vertical displacement, which can trigger tsunamis. In Aotearoa New Zealand (NZ), W-phase moment tensor inversion is used to obtain earthquake source information which is then used as input for tsunami simulations. The W-phase is a long-period low-amplitude phase visible between the P- and S-wave arrivals on a seismogram. It does not saturate with magnitude, making it useful for quantifying moment magnitudes (Mw, called Mww when obtained from a W-phase inversion) for the largest earthquakes. It also provides the centroid, rather than the hypocentre of an earthquake, allowing for a better estimation of the spatial distribution of shaking impacts and tsunamigenesis.
The W-phase is well-tested in its characterisation of global earthquakes, typically providing stable solutions within 30 minutes of the occurrence time. When used on a regional scale, challenges and limitations arise as fewer seismic records are available, and the W-phase may be less visible and usable. The network configuration, seismic source representation, and tectonic structure of the region have a direct impact on the W-phase solutions.
This thesis focuses on the specifics and limitations of regional W-phase inversions for a rapid characterisation of tsunamigenic earthquakes in the Southwest Pacific. The first chapter introduces the development and evolution of tsunami early warning systems over time, as well as the seismological principles of a W-phase inversion. The second chapter presents a first evaluation of regional W-phase solutions tested on a set of 12 recorded events. This study highlights the robustness of the retrieved magnitude compared to other parameters, suggesting that useful Mww can be obtained within 10 minutes of the earthquake occurrence. It also explores the use of regional Green’s functions to improve W-phase inversions for smaller epicentral distances. The third chapter addresses the impact of the source complexity on W-phase inversions. It examines four regional events with varying levels of source complexity. The fourth chapter focuses on estimating uncertainties associated with regional W-phase solutions for Mw7.5+ earthquakes. It presents a method using a physics-based earthquake catalogue to compensate for the lack of records of large-magnitude events. Looking at W-phase solutions for synthetic events tailored to the Hikurangi-Kermadec subduction context, it highlights the bias generated by the network configuration. Finally, a discussion summarises the specifics and limits of regional W-phase inversions in the Southwest Pacific and their implications for tsunami early warning in NZ.