- 1GNS Science,1 Fairway Drive, Avalon, Lower Hutt 5040, New Zealand
- 2Faculty of Engineering, Department of Civil and Environmental Engineering, University of Auckland, 20 Symonds Street, Auckland 1010, New Zealand
- 3National Institute for Earth Physics, 12 Calugareni, Măgurele 077125, Ilfov, Romania
The performance of the built environment during earthquakes is strongly influenced by local and regional variations in ground conditions that influence the amplitude and frequency content of ground motions. Developing models to predict these local site amplification effects is a key ingredient for the modelling of seismic hazard and risk. This study investigates the capability of various measured site parameters (e.g., fundamental frequency (f0)/period (T0), HVSR) and/or inferred site proxies (e.g., slope, rock classification, curvature) to predict local site amplification in New Zealand (NZ). To achieve this, we compiled an extensive database of relevant site parameters at 582 GeoNet seismic stations, derived from seismic data (ambient noise and earthquake recordings), geological and topographical maps, as well as site parameters included in the NZ-strong-motion database (Wotherspoon et al., 2024). Additionally, the NZ backbone model proposed by Atkinson (2024) was used to compute PSA site-to-site variability within the period range of 0.05 to 10 seconds, utilizing a comprehensive dataset of ground motion parameters from Manea et al. (2024). We then evaluated the robustness of correlations between site parameters and earthquake site-to-site variability to assess their performance both individually and in combination.
The results indicate that of any single metric, the strongest correlation with site-to-site variability is achieved by geological era, closely followed by site classes based on the 2004 NZ seismic design standard (SNZ 2004). Among measured parameters, VS30 shows the best performance at short periods, while T0 is more effective at longer periods. Conversely, Z1.0 and Z2.5 exhibit the lowest coefficients of determination, perhaps either reflecting the poor characterisation of these parameters, or implying that bedrock characteristics in NZ differ from those in regions where these parameters were originally developed. Inferred parameters such as slope, curvature, and relief perform similarly, although they may capture different aspects of site-to-site variability. In conclusion, while different geological and topographical proxies are effective for estimating site amplification at a regional scale, measured site parameters such as the fundamental frequency/period, VS30 and HVSR are also needed to capture the variability of site response at the local level.
How to cite: Manea, E., Kaiser, A., Wotherspoon, L., Stolte, A., Hill, M., and Gerstenberger, M.: Estimating Site Amplification in New Zealand Through Measured and Modeled Proxies, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13322, https://doi.org/10.5194/egusphere-egu25-13322, 2025.