Seismic Response of Alluvial Deposits due to Vertical Component of Near-Fault Earthquakes

The characteristics of vertical component of earthquakes are essentially different from those of their horizontal components in terms of amplitude and frequency content. This difference is mainly due to the fact that the vertical component is rather influenced by P-waves while the horizontal component is associated with S-waves. It is widely accepted that the energy of the vertical component of earthquakes is usually concentrated over a narrower range of high frequencies compared to that of the horizontal component. This concentration is destructive to engineering structures having vertical fundamental frequency within the high frequency range. During past earthquakes, the collapse of several buildings and bridges has been found to be directly caused by vertical excitations. Previous researches have illustrated that vertical ground motions are mainly influenced by earthquake magnitude, source-to-site distance as well as local site conditions such as geometry and material properties of subsurface soil layers.
Current paper aims to investigate the effects of various local site conditions on vertical component of near-fault earthquakes. To this end, a series of seismic site response analyses based on time-domain equivalent-linear approach were carried out using Quake-W to determine the ground response under simultaneous horizontal and vertical excitations. The site condition was categorized into different soil profiles comprising of different soil types with different V_s30 values rather than the generic site classifications used in seismic design codes. The input seismic motions used in the analyses comprised of 23 acceleration time histories from near-fault earthquakes recorded worldwide in tectonically active regions within distances less than 10 Km from the active faults (R_JB ≤ 10 km) and with magnitudes ranging from M = 5.6 to 7.6.
The results of numerical analyses were finally presented in terms of 5% damped V/H spectral ratio at the ground surface. The results were also compared with recent empirical V/H attenuation equations developed by other researchers. It was found that the obtained V/H spectral ratios show better agreement with empirical equations in case of higher-magnitude earthquakes at stiff soils than lower-magnitude ones at soft soils. In all of these cases, Poisson ratio (υ) of the soil that controls the ratio of S-waves to P-waves velocities was considered equal to 0.35 as a general value. It should be added that although υ is influenced by parameters like soil type, confining pressure and void ratio, it critically depends on the degree of saturation of the ground as well as the drainage conditions during loading. The saturation of a soil can result in rapid increase in υ, which in turn increases the P-wave velocity. To better understand this issue, some models were analyzed with a constant V_s30 value of 230 m/s but with different υ values or alternatively different P-wave velocity profiles. It was observed that as υ increased, the agreement between V/H spectral ratio predicted by numerical analysis and those obtained based on empirical equations enhanced. It can be concluded that the degree of saturation of the ground can profoundly affect V/H spectral ratio. This issue is not reflected in current V/H empirical relationships.
In conclusion, the results of current study demonstrate that V/H spectral ratio depend on parameters such as the natural period of the ground for P-wave, Poisson ratio of the soil and the frequency content of the input shaking. Moreover, it was revealed that the inverse of quarter wavelength impedance contract of P-waves shows a good correlation with V/H spectral ratios and can be used for the implementation of site effects into V/H empirical equations.