Massey, C.I.; McColl, S.T.; Lukovic, B.; Sadashiva, V.K. 2024 Development of vulnerability functions for residential buildings in New Zealand affected by earthquake-induced landslide displacement. Lower Hutt, NZ: GNS Science. GNS Science report 2024/23. 160 p.; doi: 10.21420/7D21-0W59
Abstract
The main objective of the research presented in this report is to explore how permanent ground displacement (i.e. slippage) caused by earthquake-induced landslides (EIL) triggered by the 2010/11 Canterbury Earthquake Sequence (CES) influenced the damage to predominantly 1- and 2-storey single-dwelling residential buildings located on hillslopes in the Port Hills, Christchurch, New Zealand. The main outputs from this research are: (1) the identification of factors that might control the vulnerability of 1- and 2-storey single-dwelling residential buildings impacted by landslide displacement; and (2) to investigate whether new landslide displacement-based vulnerability functions, and their uncertainties, can be derived for use in loss modelling. The main outcome that we hope to achieve from this research is that end users have greater confidence in the vulnerability functions that they use to estimate loss, i.e. that the outputs of risk models better represent expected impacts. To investigate the relationships between building damage ratios – the ratio of repair cost to replacement cost – and EIL displacement, we have used building-damage reports, claim information and repair-cost estimates collated by EQC Toka Tū Ake from the 2010/11 CES, mainly relating to the 22 February 2011 earthquake. The variables (i.e. predictor variables) used to predict the damage ratios in this research were chosen based on the variables previously found to influence both earthquake shaking on slopes and the damage to buildings exposed to earthquake shaking. These variables can be grouped into: (1) the hazard intensity, in this case, the level of earthquake shaking and/or the amount of permanent ground displacement; (2) the geology and morphology (shape of the topography) of the site on which the building is founded and which may influence the response of the site to earthquake shaking; and (3) the characteristics of the building. The findings suggest that the damage ratios for the buildings located within the landslide (mass movement) areas in the Port Hills are overall higher than those for those outside the landslide areas with equivalent levels of ground shaking (i.e. the Deep Dive shake-only dataset). This indicates that earthquake-induced permanent ground displacement caused more building damage than shaking alone. However, the results suggest that interactions between the earthquake shaking and the mechanisms and amounts of permanent ground displacement, as well as the location, size and other building-specific factors, are complex, with significant scatter shown by the relationships. The logic underpinning the damage-ratio model fitting was that earthquake-induced ground shaking causes shaking damage to buildings and permanent ground displacement, which also leads to building damage. The results show that the best-fitting models are those that adopt the variables of amplified peak ground acceleration, net permanent ground displacement, building strain and floor area. However, these all have relatively low statistical relationships. Given these results, there are multiple options that might be used, with caution, to estimate the number of buildings likely to be damaged from permanent ground displacement caused by landslides and their subsequent loss, at different levels of shaking, at the national scale. (auths)
