de Vilder, S.J;. Brideau, M.-A.; Massey, C.I. 2022 Empirical and physics-based runout models. Lower Hutt, N.Z.: GNS Science. GNS Science report 2019/38. 42 p.; doi: 10.21420/HYMS-9W10
Abstract
The Stability of Land In Dynamic Environments (SLIDE) research project aims to improve the resilience of New Zealand’s buildings and infrastructure through better knowledge of the behaviour of slopes in response to strong earthquake shaking and significant rain events and the development of strategies for more robust remediation approaches. The SLIDE project’s main objectives are: 1.to establish inventories and data relating to slope processes, and 2.to use these data to underpin landslide hazard and risk assessments. This project is currently focused on Wellington, where a combination of various research funds have been used to establish a broader project to determine the effects of earthquakes and rain on slope stability. As Wellington is urbanised this phase of the project has focused primarily on anthropogenically modified slopes.This report presents both an empirical database of landslide runout and calibration of modelling parameters required for numerical landslide runout analysis. Understanding the distance a potential landslide can travel and the area it could impact are important components of hazard and risk analysis. Back analysis of previous landslides is traditionally used to predict the runout behaviour of potential future landslides. This can be done using empirical – statistical relationships (typically adopted for regional scale studies), or through numerical modelling of landslide run-out (typically used for more detailed, site-specific analysis). Both datasets are suitable for use and reflect the types and sizes of landslides that can occur within Wellington. The empirical database comprises national and international case studies, and specific relationships are derived for different landslide types, including: i) earthquake-triggered debris avalanches, ii) rainfall-triggered debris avalanches, iii) rainfall-triggered debris flows, and iv) fill slope failures. The numerical modelling calibration uses landslides observed during the Kaikoura earthquake, as well as landslides triggered along the Kaikoura coast during the ex-tropical cyclone Gita rain event as back analysis case studies. An iterative modelling procedure is employed, using RAMMS software, until the best-fit between model simulation results and observed landslide runout are obtained. The best-fit is quantified using: the difference in deposit distribution between the simulation result and observed landslide deposit. The friction parameters of the best fit model are then used to inform future numerical modelling. (auth)
