Amplification of seismic waves due to surface topography and subsurface soils has often been observed to cause intensive damage in past earthquakes. Understanding the topographic amplification of ground motions can provide scientific guidelines for the design of buildings and evaluation of regional scale landslide hazards in mountainous regions. This study aims to numerically quantify the amplification of earthquake ground motions as well as the associated landslide hazards due to 3D realistic topography and subsurface soil conditions. A series of large-scale 3D numerical simulations are conducted to investigate the seismic topographic amplification effect of Hong Kong Island using Spectral Element Method (SEM). The major findings of this thesis are as follows:

(a) Parametric m...[ Read more ]

Amplification of seismic waves due to surface topography and subsurface soils has often been observed to cause intensive damage in past earthquakes. Understanding the topographic amplification of ground motions can provide scientific guidelines for the design of buildings and evaluation of regional scale landslide hazards in mountainous regions. This study aims to numerically quantify the amplification of earthquake ground motions as well as the associated landslide hazards due to 3D realistic topography and subsurface soil conditions. A series of large-scale 3D numerical simulations are conducted to investigate the seismic topographic amplification effect of Hong Kong Island using Spectral Element Method (SEM). The major findings of this thesis are as follows:

(a) Parametric models for ground motion amplification considering 3D topography and subsurface soils are developed. Results show that the topographic amplification factor is frequency dependent, and the topographic amplification pattern is significantly influenced by the thickness of the subsurface soil layer due to the coupled soil-topography effect. The parametric models characterize the topographic amplification considering subsurface soils, material damping and input wave frequencies using simplified parameters such as smoothed curvature, slope angle, elevation and soil depth. The predictions are proved to be accurate with standard deviation of residuals generally between 0.1-0.15.

(b) Geostatistical method is proposed to construct 3D realistic subsurface soil profiles with borehole data from field investigation. Variability of soil depths and the correlation between the shear wave velocity and the SPT-N values is considered in this subsurface soil velocity model. Ground motion simulations with the realistic subsurface soil model show that the amplification factors are significantly large compared with rock outcrop in the time and frequency domain. The amplification factor could reach as high as more than 10 at some frequencies. If different sources of variabilities on soil profiles are considered, the local amplification factor could become more variable. The phenomenon is in line with large amplification factors observed in some case studies.

(c) The influence of heterogeneity of soils on the amplification effect of horizontally layered soil grounds is quantified using 2D SEM analyses. Compared with a uniform soil profile, soil heterogeneity, represented by spatially correlated random field, shifts the resonance frequencies of the site and results in reduced amplification factors at high frequencies. In general, spatially correlated random field with shorter correlation distances behaves more homogeneously. Moreover, the spatial correlation distance of the amplification factor at high frequencies is highly related to the horizontal correlation distance of the spatially correlated random field.

(d) A physics-based method is proposed to evaluate the regional scale earthquake-induced landslide hazards. The numerical simulation directly considers the amplification effect of topography and soil layers, while the associated landslide is indicated by the Newmark displacement analysis. Results show that the topographic amplification effect and the vertical acceleration have a great impact on the results of Newmark displacement analysis. Under the 2475-year-return-period earthquake, about 4.9% of the terrain has a probability of slope failure of more than 5%. This physics-based method can be performed as long as the topographic data, subsurface soil profiles are provided and the seismic scenario is assumed.