Lee, R.L; Bradley, B.A.; Hill, M.P.; de la Torre, C.A.; Kaiser, A.E.; Wotherspoon, L.M. 2023 3D ground motion simulation-based site amplification considering multiple basin geometries: a Wellington, New Zealand, case study. Lower Hutt, NZ: GNS Science. GNS Science report 2022/40. 51 p.; doi: 10.21420/R21F-3C28
Abstract:
This report provides a region-specific study on site amplification in Wellington, New Zealand, predicted through 3D physics-based ground-motion simulations. The Wellington central business district (CBD) overlies a complex sedimentary basin that has been observed to strongly amplify surface ground motions through historic events, such as the 2016 Mw 7.8 Kaikōura earthquake. As Wellington is exposed to high seismic hazard, it is important to understand the role and influence of the basin on site amplification. Several recent models of the Wellington basin are tested in simulations to quantify the predicted site amplification and provide insight into the corresponding epistemic uncertainty. Response spectral amplification factors are shown at several locations across the Wellington CBD, along 2D transects and also across the area in spatial maps. This highlights the site amplification and uncertainty patterns across the Wellington CBD and ties those to features in the basin models. Specifically of interest is the role of the southern basin margin and how it causes a double-banded amplification structure due to a combination of concave and convex features. The analyses also highlight the limitations resulting from simulation spatial resolution, where impedance effects could not be adequately modelled at locations with shallow basin depth. When compared to observed site amplifications, the simulations were able to capture the level of site amplification at some deeper basin sites, as well as broad spectral amplification peaks. At shallow basin depth sites, the simulations under-predicted the level of site amplification due to the spatial-resolution limitations. At the strong-motion station TEPS, the level of amplification was well-matched, but the period at which the simulated spectral site amplification peak occurs (T = 0.7 s) was shorter than in observed site amplification (T = 1.1 s). Through investigating 2D transects, it appeared that the large short-period amplification was through a combination of basin-depth, basin-edge and wave-guide effects along the Te Aro sub-basin. Simulated spectral amplification factors were as large as 8.0 in some locations and consistently large at the CentrePort area, where significant liquefaction and structural damage has previously been observed. Improvements can be made to the simulations with more constrained basin-depth geometry and characterisation of the shear-wave velocity of soils within the basin. Numerical modelling improvements can also be made with higher spatial-resolution simulations, reduction of the minimum S-wave velocity in the simulations and proper near-surface nonlinear site response modelling using physics-based approaches. While each of the noted improvements are likely to be necessary to achieve accurate modelling of basin site response, the immediate next phase of investigation will look to utilise improved velocity characteristics of basin sediments in Wellington.
