The majority of the Earth’s seismic activity occurs within the top 30 km of the crust suggesting that plate tectonic processes are driven by relatively shallow dynamics. At the same time large, shallow and tsunamigenic earthquakes are among the most destructive natural disasters faced by Australia and the Pacific ‘ring-of-fire’. As an example, the 2004 Great Sumatra earthquake and associated tsunami reagically killed ~230,000 people in 14 countries. In this context, accurate estimation of the location, duration, depth and mechanism of shallow earthquakes is crucial for seismic hazard assessments and a better understanding of the near-surface dynamics of the Earth.
Since four decades ago, earthquake source parameters have been estimated at through waveform modelling of low-frequency data and the results collected in earthquake catalogues such as the global-centroid-moment-tensor (GCMT) and USGS. These catalogues rely on two assumptions: 1) one-dimensional spherically symmetric Earth models, and 2) the point source approximation of seismic source. However, it is well-understood that the Earth's structure varies significantly with lateral position (as is observed in tomographic maps of the Earth’s interior). It has also been shown that earthquake originate at elongated (sometimes very complex) ruptures and not point wources.
I have developed and implemented methods to incorporate the 3D heterogeneity of the earth into seismic source parameter estimation. Based on these developments, I have drawn a new earthquake catalogue for Papua New Guinea and the Solomon Islands using a 3D continental model of the Australasian region and I will discuss the new insights that this dataset provides. Using high-resolution 3D Earth models, I present a high-frequency simulation of the 2016 Petermann Ranges earthquake in central Australia, the 2013 nuclear explosion in North Korea and the 2007 Caldera collapse on Reunion Island.
I present my ongoing research to go beyond the point source approximation where I seek new intelligent methods to simulate the length, complexity and duration of large earthquakes.