Backbone ground-motion models for crustal, interface and slab earthquakes in New Zealand

SKU:
SR_2022-11.pdf
$0.00
(Inc. GST)
$0.00
(Ex. GST)
Write a Review

Atkinson GM. 2022. Backbone ground-motion models for crustal, interface and slab earthquakes in New Zealand. Lower Hutt (NZ): GNS Science. 61 p. (GNS Science report; 2022/11). doi:10.21420/QMJ6-P189.

Abstract

A ground-motion model (GMM) that strikes a balance between empirical- and simulation-based approaches is developed in support of the 2022 update of the New Zealand National Seismic Hazard Model. The development follows the backbone approach, comprising a central model to express the median ground motions for earthquakes in New Zealand, along with upper and lower alternatives to describe its epistemic uncertainty. Epistemic uncertainty expresses uncertainty in the median model due to our inability to completely characterise the appropriate functional form and its parameters; this uncertainty varies according to limitations in the completeness of the New Zealand ground-motion database and thus varies with magnitude and distance. Aleatory variability (or random scatter) of ground-motion amplitudes about the median is also characterised. Separate GMMs are developed for crustal, interface and in-slab earthquakes. The approach taken is to perform a regression analysis of the New Zealand response spectra database (Hutchinson et al. 2022) based on the generic GMM model of Yenier and Atkinson (2015b) and Hassani and Atkinson (2018), which employs the concept of equivalent point-source simulations. The model parameters that control the scaling of the GMM with magnitude and distance describe source effects (seismic moment and stress parameter), path effects (geometric and anelastic attenuation) and site effects (site shear-wave velocity). The overall approach and functional form have been validated and calibrated extensively using the California NGA-W2 database for crustal earthquakes (Yenier and Atkinson 2015a, 2015b) and the Japan NGA-Sub database for interface and slab earthquakes (Hassani and Atkinson 2018). The New Zealand database provides constraint on the model for M~4–7 for frequencies from 0.2 to 100 Hz at distances to ~400 km; extension of the GMM to larger magnitudes (M7–9) is constrained by the Hassani and Atkinson (2018) seismological model, which was developed for application to events of M3–9. The model background and its formulation for earthquakes in New Zealand is described in this report. All model coefficients, along with the aleatory variability components, are given in an electronic attachment (NZCoefficientsV8.xlsx). The model comprises a central branch (best-estimate) with upper and lower branches representing epistemic uncertainty. The central, upper and lower branches are evaluated for a table of magnitude, distance and frequencies in the electronic attachment spreadsheets for each event type (NZCentralBBV8.xlsx, NZLowerBBV8.xlsx, NZUpperBBV8.xlsx). Site terms for all stations are also provided in an electronic attachment (NZStationTermsV8.xlsx). The performance of the central model relative to the ground-motion database is illustrated in Appendix 1. (The author)