Exposure of Australian dams to seismic hazard from proximal faults

Abstract The purpose of this study is to obtain estimates of ground surface rupture and ground motion hazards at standardised scales useful for general reference and regional comparisons of dams (n = 548) registered with the Australian National Committee on Large Dams (ANCOLD). Geospatial and statistical methods are used to investigate the exposure of dams to seismic hazard from 409 faults in the Geoscience Australia Neotectonic Features Database (NFD). We identify 216 faults at distances less than 100 km from 428 dams and measure 4055 fault-to-dam distances. At least 31 dams are located within 1 km of NFD fault traces and at least 16 of these dams could reside within NFD primary fault zones. Estimates of NFD maximum moment magnitudes (M w,max ) from fault area and length regressions range from 5.6 ≤ M w ≤ 7.9. Average (AD) and maximum (MD) NFD fault displacements for M w,max events range from 0.3 m ≤ AD ≤ 4.4 m and 0.9 m ≤ MD ≤ 8.4 m. Distance-probability regressions for distributed ground surface rupture suggest approximately 40 dams have a ≥ 10% probability of ground surface rupture occurring in the area encompassing the dam in a M w,max event. Peak ground accelerations (PGA) and pseudo-spectral acceleration at 1.0 s (PSA[1.0 s]) are estimated for M w,max at dam sites using the same set of ground motion models (GMMs) for the 2023 National Seismic Hazard Assessment (NSHA23), assuming engineering rock site conditions with a time-averaged shear wave velocity between the surface and 30 m sediment depth (VS30) of 760 m/s. Of the 4055 M w,max scenarios considered in total, 3579 M w,max scenarios produce 85th percentile PGA ≥ 0.1 g at dam sites and 2844 scenarios produce 85th percentile PSA[1.0 s] ≥ 0.1 g. Comparison with 1:5,000 annual exceedance probability (AEP) PGA and PSA[1.0 s] estimates from the NSHA23 indicate there are 404 out of 428 dams where NFD M w,max PGA are greater than NSHA23 values, and 422 instances where PSA[1.0] > NSHA23. Proximity to NFD faults imparts a first-order control on relative hazard. A large increase in the number of identified NFD faults over the last decade suggests further research will continue to increase NFD fault populations. NFD fault slip rates and M w,max are important parameters in seismic hazard analysis for many dams but exhibit large epistemic and aleatoric uncertainties. Improving the characterisation of NFD faults through acquisition of direct fault-specific seismic hazard information (e.g., single-event displacements, slip rates, inter-event times, frequency-M w distributions, segmentation scenarios) will assist with hazard profiling for many ANCOLD dams.