2026-10-06T21:01:22Z https://api.figshare.com/v2/oai

oai:figshare.com:article/34041333 2026-10-01T07:32:15Z category_32621 portal_771 item_type_8 month_year_10_2026

Spatiotemporal Analysis of Two Decades of Seismicity Across Aotearoa New Zealand Codee Leigh Williams (14201846) Seismology and seismic exploration Machine Learning Seismology Earthquake Catalogue School: School of Geography, Environment and Earth Sciences 370609 Seismology and seismic exploration Degree Discipline: Geophysics Degree Name: Doctor of Philosophy Degree Level: Doctoral Existing catalogues of seismicity in Aotearoa, New Zealand, cover a range of scales and time periods. GeoNet’s nationwide catalogue is primarily intended to facilitate rapid hazard communication and response. This emphasis, combined with the heterogeneous nature of the New Zealand seismograph network, can mean that some earthquakes are mislocated or not detected at all. In contrast, regional catalogues developed for research purposes offer a much more in-depth view of seismicity but require additional data and/or manual or otherwise time-consuming analysis that are not efficient or available for a near real-time nationwide catalogue. We have developed and applied a modern automated workflow to generate a research-grade earthquake catalogue for Aotearoa New Zealand that balances the best of national and regional earthquake catalogues. This workflow expands and improves the current catalogue, enabling further research into New Zealand's seismo-tectonic setting and a better understanding of present-day seismic hazard. Overall, our methodology was able to pick and locate 99.7% of all earthquakes in the GeoNet catalogue between 1st January 2001 and 1st January 2021, totalling 407,553 earthquakes, and provide systematic and robust location uncertainties. In general, we find inconsistent changes in hypocentral locations relative to those located by GeoNet. We explore the diverse range of tectonic settings illustrated across the temporal and spatial extent of the catalogue and consider the impact of our locations on constraining subduction zone structure, as well as more accurate characterisation of spatial relationships between earthquake sequences. At a national scale, we calculate the seismogenic cut-off depth (D95), analysing seismic hazard at the Alpine Fault, Central North Island and the subduction margins. We have extended our nationwide machine learning earthquake catalogue to detect and locate earthquakes between 2011 and 2018 in the Southern Hikurangi Margin. Our catalogue contains 36,532 events detected by GeoNet, as well as 32,620 previously undetected events, to almost double the number of earthquakes located between 2011 and 2018 in the Southern Hikurangi region. We do this by further improving our workflow to identify events missing from our national catalogue and associating these events using a novel approach combining kurtosis and PyOcto association techniques.Using this catalogue, we investigate spatio-temporal variability in earthquake occurrence in the region. First, we discuss the key structural features of the region, including the presence of faults in the overriding plate, the depth to the subduction interface and the structure of the Wadatti-Benioff zone. We find no seismically active structures which would indicate fault linkages between the North Island Dextral Fault Belt and the Marlborough Fault Zone, and instead find disperse continuous seismicity across the Cook Strait. In the lower North Island, we analyse the intersection of crustal faults with the underlying subduction interface and speculate on the role this may play in past (e.g. 1855 Wairarapa Earthquake and 2016 Kaikoura Earthquake) and future large earthquakes.We build on our studies of seismicity in the Southern Hikurangi Margin, investigating triggering from teleseismic, national, and local events. Of the 683 potentially triggering events identified, we find 10 events to correlate with statistically significant increases in seismicity rate, most notably the Kaikoura earthquake and its aftershocks. We identify three separate mechanisms of triggering taking place in the 7 days following the initiation of the Kaikoura sequence.Finally, we bring together these results from the three chapters of this thesis and discuss how this improves our understanding of seismicity and seismotectonic processes in Aotearoa New Zealand. We offer insights into how this work may lay the basis for further research with and development of seismic catalogues both in New Zealand and globally. We discuss the opportunities and limitations of machine learning integrated catalogues and large datasets in seismology and suggest areas for future work in catalogue development. 2026-10-01T00:00:00Z Text Thesis 10.26686/8v5m-t3k1 https://figshare.com/articles/thesis/Spatiotemporal_Analysis_of_Two_Decades_of_Seismicity_Across_Aotearoa_New_Zealand/34041333 Author Retains Copyright