Site and topography effects are integral part of strong ground motion recorded during an earthquake. Site effects due to shallow subsurface velocity and topographic changes have been clearly seen in the Gorkha earthquake, 25th April, 2015 ($M_w$ 7.8) at Kapkot and Berinag stations, which lies at an epicentral distance of 507 and 485 km, respectively. The high peak ground acceleration was recorded at Kapkot station that is at valley, while comparatively low peak ground acceleration was recorded at Berinag station that is at hill. This paper investigates the effect of site topography and shallow velocity structure on ground acceleration generated due to propagation of SH wave generated by a finite far-field rupture. The propagation of SH wave in a shallow subsurface earth model with the vertical variation of velocity can be modelled by finite difference (FD) method based on staggered algorithm that can effectively model the propagation of the seismic wave in isotropic as well as heterogeneous elastic medium. This paper discusses the role of staggered algorithm in the generation of particle motion at the surface of modelled earth characterized by surface topography and vertical distribution of elastic constants. The developed software for FD modelling of the medium has been tested for SH wave propagation in a purely elastic medium in terms of numerical stability, dispersion and boundary conditions. Numerical experiments show that the method effectively models the topography and thin surface velocity layers in the model for varying cases. The obtained surface acceleration records from the propagation of SH wave at Kapkot and Berinag clearly show that both the site amplification and topographic effects have played a vital role in shaping the accelerograms at these stations.

This paper investigates the three-dimensional frequency-dependent attenuation structure of the Garhwal Himalaya in the Indian subcontinent. Based on the distribution of earthquakes and recording stations in the Garhwal Himalaya, the entire region of 152 ${\times}$ 94 km$^2$ is divided into 108 three-dimensional uniform rectangular blocks. These blocks are assumed to be of thickness 5 km that extends to 15 km depth. Each block represents the rock of different attenuation coefficients. The S-phase of strong motion records has been used to estimate the shear wave quality factor in each block by the inversion of spectral acceleration data. The inversion of spectral acceleration data is based on the modified technique of Joshi (2007) and Joshi et al. (2010) which was initially given by Hashida and Shimazaki (1984). The earthquake data of 19 events digitally recorded by 33 stations of the strong motion network between 2005 and 2017 have been used in this paper. The outcome of the inversion process is the shear wave quality factor at all frequencies present in the records. The three-dimensional attenuation structure at various frequencies is presented in this paper and is correlated with the regional tectonics of the Garhwal Himalaya. The correlation of attenuation structure at 10, 12 and 15 Hz with the tectonics of the region indicates that the shear wave quality factor has a strong relationship with the tectonics of the region. The values of the shear wave quality factor at different frequencies obtained from inversion have been used to obtain a relation of shear wave quality factor Q$_{\beta}$(f)=107f$^{0:82}$ for the region of Garhwal Himalaya for frequencies 10–16 Hz. The comparison of obtained shear wave quality factor with other studied relations clearly indicates that the obtained relation is close to what has been obtained in earlier studies and thereby indicates the reliability of obtained three-dimensional shear wave attenuation structure from inversion of spectral data.