Landslides DOI 10.1007/s10346-016-0684-8 Received: 30 July 2015 Accepted: 1 February 2016 © Springer-Verlag Berlin Heidelberg 2016 C. Massey I F. Della Pasqua I C. Holden I A. Kaiser I L. Richards I J. Wartman I M. J. McSaveney I G. Archibald I M. Yetton I L. Janku Rock slope response to strong earthquake shaking Abstract The 2010–2011 Canterbury earthquakes triggered many mass movements in the Port Hills including rockfalls, debris avalanches, slides and slumps, and associated cliff-top cracking. The most abundant mass movements with the highest risk to people and buildings were rockfalls and debris avalanches sourced from up to 100m high cliffs inclined at angles >65°. Cliffs lower than 10m in height generally remained stable during the strong shaking, with only isolated release of a few individual boulders. We used site-specific data to investigate the factors that controlled the response of the cliffs to the main earthquakes of the Canterbury sequence, adopting two-dimensional finite element seismic site response and stability modeling that was calibrated using the field observations and measurements. Observations from the assessed cliffs in response to the earthquakes show the taller cliffs experi- enced larger amounts of permanent cliff-top displacement and produced larger volumes of debris than the smaller cliffs. Results indicated a mean K MAX amplification ratio for all sites under study of 1.6 (range of 1.1–3.8). Field data and numerical modeling results, however, show that amplification of shaking does not necessarily increase linearly with increasing cliff height. Instead, our results show that accelerations are amplified mainly due to the impedance contrasts between the geological materials, corresponding to where strong differences in rock mass shear wave velocity exist. The resulting acceleration contrasts and rock mass strength control cliff stability. However, the amount of permanent slope displacement and volume of debris leav- ing the cliffs varied between the sites, due to site-specific geometry and rock mass strength. Keywords Canterbury earthquakes . Co-seismic landslides . Port Hills . Rock slope response . Slope stability . Site effects Introduction The 2010–2011 Canterbury earthquakes, New Zealand, triggered many mass movements in the Port Hills of Christchurch including rockfalls, debris avalanches and slides and associated cliff-top cracking, and soil slumps (Fig. 1). About 100 homes were damaged by rockfalls and debris avalanches, leading to the temporary evac- uation of many hundreds of residents. The 2010–2011 Canterbury earthquakes commenced on 4 September 2010 (New Zealand time UTC + 12 hours) with the M W 7.1 Darfield earthquake, situated 40 km west of the Port Hills (Fig. 1, inset). The damage and loss inflicted by the Darfield earthquake was eclipsed by the M W 6.2 Christchurch earthquake of 22 February 2011, which occurred directly under the Port Hills (Fig. 1). Widespread mass movements were triggered in the Port Hills including—using the scheme of Keefer (1984)—disrupted rockfalls, debris avalanches and associ- ated cliff-top cracking, and coherent soil slumps and slides (e.g., Dellow et al. 2011). Of the mass movements triggered in the Port Hills by the Canterbury earthquake sequence, rockfalls and debris avalanches were the most abundant type and caused the highest risk to people and buildings (Massey et al. 2014a). Rockfalls, debris avalanches and cliff-top cracking were also triggered by after- shocks on 16 April, 13 June, and 23 December 2011 (Massey et al. 2014a). This paper presents the results of our investigations into the response of several largely bedrock cliffs in the suburban areas of the Port Hills to the main 2010–2011 Canterbury earthquakes. The cliffs investigated are as follows: (1) Quarry Road, (2) Redcliffs, (3) Cliff Street, and (4) Richmond Hill. Seismic response of rock slopes Previous research has shown that the dynamic response of a slope during an earthquake comprises a complex interaction between seismic waves and the hill slope (e.g., Sepulveda et al. 2005). The response of a slope to an earthquake is thought to be controlled by the following: (1) the nature of the earthquake source; (2) wave propagation path effects; and (3) local site conditions and their effects on amplifying or de-amplifying shaking (Kramer 1996; Sepulveda et al. 2005 Kaiser et al. 2013). Path effects are taken into account in our modeling with application of the same regional attenuation functions for each event and site. In this study, we focus on the way ground amplification varies with different sources and site conditions. This is because the sites are very close to each other and to the earthquake sources, meaning that that the earthquake source to rock-slope site path lengths are short (Table 1), and therefore, a strong variability in path effect is unlikely. Research has shown that firstly, amplified ground motions in slopes can result from near-surface impedance contrasts associat- ed with material velocity contrasts caused by (i) local surficial deposits (fill, colluvium, alluvium, etc.) overlying rock (e.g., Bourdeau and Havenith 2008; Del Gaudio and Wasowski 2011); (ii) weathered materials overlying less weathered materials; (iii) highly fractured zones within more intact materials and discrete large-scale fracture zones (e.g., Moore et al. 2011; Gischig et al. 2015 ). Secondly, focusing of seismic waves by surface morphology—mainly slope inclination, height and shape, e.g., convex, concave, or planar—may result in topographic amplifica- tion (e.g., Geli et al. 1988; Benites and Haines 1994; Meunier et al. 2008; Hough et al. 2010), at larger “ridge-scales” and at smaller “site-scales” (Kaiser et al. 2014). For slopes, the characteristic site period can be influenced by both local slope materials and their contrasts, together with topography, and provides an indication of the frequency at which the most significant amplification can be expected (Kramer, 1996). Earthquake ground motions of similar amplitude, duration, and location are thought to affect the slope in different ways depending on the frequency content of the earth- quake, which is strongly influenced by earthquake magnitude and source-to-site distance. Therefore, the effects of amplification and resultant permanent slope displacement are likely to be larger when excited by earthquakes with predominant frequencies simi- lar to the fundamental frequency of the slope. A laboratory study Landslides Original Paper