Site response analysis for estimation of seismic site amplification in the city of Durban (South Africa)

Abstract

The city of Durban has previously experienced higher than expected ground motions from large distant earthquakes. It is potentially exposed to significant seismic hazard due to seismic site amplification, which needs to be estimated for effective mitigation efforts. Detailed stochastic one dimensional (1D) seismic site response analyses were performed at 90 sites in the city. Analytical models have demonstrated that they can simulate reasonably well the seismic ground motions amplification. The most widely used model is the equivalent linear approach. The approach computes the ground response of horizontally layered soil deposits subjected to transient and vertically propagating shear waves through a 1D soil column. Seven earthquake time histories together with developed sub-surface models were selected as input parameters to estimate the seismic site amplification at the 90 sites in the city. The used time histories were taken from the 2014 M5.5 Orkney earthquake with distance range (4.8–46.9 ​km). The uncertainties in ground motion input, variation in the shear wave velocity and variations in the shear modulus reduction and damping curves (i.e. variation of non-linear properties) were carefully modelled. Results obtained from this study were used to prepare maps of peak ground acceleration (PGA) at the surface and amplification factors. The minimum and maximum PGA at surface are estimated as 0.01 ​g and 0.30 ​g respectively. Based on the results of the analysis, the city may sustain amplification in the range of 0.7–4.7 ​at PGA with high values along the coast. The results indicate that the low shear wave velocity values, weak and soft material at shallow depths are responsible for the higher amplifications observed especially along the coast. Therefore, a site-specific design approach should be adopted for the seismic design of critical structures.