The New Zealand National Seismic Hazard Model: Current Models and Future Developments
Dr. Matt Gerstenberger
(Risk and Engineering Team Leader, GNS Science)
Auckland Presentation: Thursday 28 January 2016
Wellington Presentation: Monday 1 February 2016
Christchurch Presentation: Tuesday 2 February 2016
Video recording link:
New Zealand seismic hazard modelling has relied on probabilistic seismic hazard analysis (PSHA) since the 1980s. Beginning with early models using large regional zones the models have become more detailed and precise (Stirling et. al.,1998, 2002) with the use of active faults, gridded smoothed seismicity, and ground motion prediction equations (GMPE; e.g., McVerry, et. al., 2006) developed for New Zealand. The most recent revision (Stirling, et. al., 2012) now contains more than 500 faults (with some use of time-dependence) which control the hazard over much of the country. The source models of the NSHM have recently been made available for download from the GNS Science webpages. Future iterations of the model will be developed using Open Source platforms such as OpenQuake and OpenSHA. We are also working toward making additional outputs more readily available, including information for displacement based design and deaggregations for ground-motion selection.
In other work we have developed a time-dependent hazard model for the Canterbury region in response to the earthquake sequence. We have used a hybrid seismicity model that captured time-dependence on three different time scales, from short-term to long-term. Statistical testing of this hybrid model (Rhoades et al, 2015) shows that the hybrid out performs any of the individual models, including the NSHM, over long-term forecasts such as used for building design standards. Additionally, we combined the McVerry et al (2006) GMPE with the Bradley (2010, 2013) GMPE to recognise the uncertainty in ground-motion prediction.
Our current work in seismic hazard is focused on rethinking some of the fundamentals inputs to PSHA. A primary driver is to understand the influence of uncertainties and how they may propogate through to policy. We are investigating what defines a source model, what information should be included, how to best capture epistemic uncertainty, updating the subduction zone model, and how we can verify if a new model is indeed an improved model. We are also exploring the impact of various currently unquantified uncertainties on the downstream risk-based applications of the model. Finally we are investigating how modelling of synthetic seismicity and ground motions based on physics can be used in future seismic hazard estimation in New Zealand.