タイトル：

Understanding and modeling the occurrence of earthquake rupture in New Zealand

要旨：

Recent work in Aotearoa New Zealand has identified several key challenges we face related to the occurrence of earthquake ruptures. At a high-level, these relate to complexity of rupture processes both in terms of the physical nature of the ruptures themselves and in terms of the temporal interdependence and relations between ruptures. The last 200 years of large earthquakes in New Zealand, and particularly in the last 50 years, have demonstrated complex surface rupture patterns and have shown the possibility for the Hikurangi subduction interface to jointly rupture with crustal faults during large earthquakes. Additionally, there is strong evidence for spatial clustering of large earthquakes, and, importantly, uncertainty around the long-term behaviour of earthquake occurrence and whether or not there is a true mean rate. Non-stationarity of the mean rate suggests that care must be taken when using the mean of past observations as an indication of the number of earthquakes that may occur in the future.  Also, as observational data sets increase in size and quality they can help us constrain these and other problems; yet important questions remain about the quality of completeness of the observational data sets. Ultimately, much of what we use as data are models themselves, involving interpretations of the base observations which can have implications for the downstream use of these observations. Such things have the potential to impact Gutenberg-Richter b-values, non-stationarity understanding, and, e.g., our understanding of the past behaviour of faults. To overcome observational data limitations, it is important that we bring together more and more the observations with the use of multi-cycle simulators. Quasi-dynamic and fully dynamic simulators show great potential to supplement our understanding from observational data, yet come with their own challenges such as understanding how well they usefully represent the real world or earthquake physics; also, to fully help understand the problems outlined above requires multi-cycle simulators to run for numerous cycles, which brings computational difficulties. However, bringing together earthquake observations with simulators is likely to be where our biggest gains in understanding of earthquake processes are likely to come from. In this seminar I will discuss my work and how it has contributed to understanding these problems. This includes, at a high-level, topics such as earthquake systems modelling, short-term earthquake forecast modelling, and testing of observational models.