Holden, C.; Kaiser, A.E.; Massey, C.I. 2014 Broadband ground motion modelling of the largest M5.9+ aftershocks of the Canterbury 2010-2011 earthquake sequence for seismic slope response studies. Lower Hutt, N.Z.: GNS Science. GNS Science report 2014/13 48 p.
Abstract: We propose to produce broadband free-field rock outcrop synthetic seismograms for the largest M5.9+ aftershocks of the 2010-2011 Canterbury earthquake sequence at sites of interest for slope failure studies. These synthetic seismograms will be used as rock input motions for site-specific modelling of slope responses – for slopes in the Port Hills of Christchurch – currently being investigated by GNS Science, for the seismic slope stability project being funded by the New Zealand Natural Hazard Platform. There were no instrument recordings at these particular sites and it is not ideal to use the nearest strong motion station due to large distance between the target site and strong motion site or strong local site effects (including non-linear behaviour). We employ a stochastic approach to compute the seismograms (Motazedian and Atkinson, 2005) that is not only controlled by detailed source models but also by regional parameters derived using spectral inversion of the extensive Canterbury strong motion dataset (Oth & Kaiser in press; Kaiser et al., 2013). We validate the method by comparing and obtaining good agreement between recorded seismograms and synthetic seismograms computed at strong motion sites within 20 km of the earthquake sources. Our results show that using appropriate stress drop value and site-specific amplification functions helps greatly to reproduce key engineering parameters such as Peak Ground Acceleration (PGA) and response spectra. The technique is particularly efficient for “moderate shaking” events such as the local Mw 5.3 April 16th event and the distant Mw 5.9 December 23rd event. For larger events such as the February 22nd and the June 13th events, key parameters are also well reproduced for sites exhibiting linear response behaviour such as LPCC, D14C, MQZ and CACS. These sites are classified as site class B rock site (NZS1170), except for CACS which is a class D deep/soft soil site (Note, that this site behaves essentially linearly as it is both further away from the earthquake source and located on stiffer soils to the west of the city). However for sites behaving non-linearly such as HVSC, CMHS, REHS and SHLC, adding a site-specific amplification function results as expected in over-predicting ground shaking (by factors of 2 to 4 times PGA) and respective response spectra. CMHS and REHS are classified as site class D (NZS1170). These sites require detailed non-linear soil-response analyses. These results support the method we have chosen to produce rock ground-motion time histories and characteristics at sites of interest in slope-response analysis. (auth)