The present study reports the 1-D ground response analysis (GRA) results for various locations near Yamuna River, Noida, India. Yamuna river forms a major sub-basin of the Ganga River system that covers 42.5% of the total catchment area of the Ganga basin. Yamuna is the main river system of Delhi region and primary perennial source of water. Yamuna river experiences frequent flooding in the monsoon due to heavy rain and increased water level. It has been noted that the water table in all the locations is significantly high, which enhances the chance of the liquefaction of the sandy strata or cyclic mobility of the soft clayey strata. The sands obtained from the Yamuna bank fall into the boundary of the most liquefiable zone, hence earthquake studies are essential in the riverbank of Yamuna. In the present study, an attempt has been made to carry out one-dimensional ground response analysis by equivalent linear method using open-source software DEEPSOIL, for six borehole locations near the Yamuna riverbank considering 1991 Uttarkashi earthquake (M_w=7.01) as an input ground motion. Shear wave velocity is an essential input for carrying ground response analysis. Shear wave velocity is correlated from the SPT ‘N’ values using empirical relations. The peak ground acceleration value, Amplification Ratio, and response spectrum were determined for all the six locations.

This paper focuses on the stratigraphic effects of medium-stiff soil deposits and describes the results of 1D non-linear seismic response analyses carried out using the computer code Deepsoil. A set of 30 soil profiles were considered varying the bedrock depth H in the range of 20-60 m and assuming a compliant bedrock. A set of horizontal acceleration time histories, recorded on almost flat rock outcropping sites, with amplitude, energy and frequency content varying in wide intervals were used as input motions. The parameters defining the backbone curve have been calibrated using shear modulus reduction curves available in the literature for soils of low to medium plasticity. The analysis results are presented in terms of stratigraphic and spectral amplification factors obtained from the motion computed at ground surface. The analysis results show that the stratigraphic amplification factor is significantly influenced by the peak outcrop acceleration and bedrock depth while it is almost unaffected by the data sorting with reference to earthquake magnitude; accordingly, original empirical predictive models have been proposed to estimate the stratigraphic amplification factor as a function of the most influencing parameters.

Most parts of North-East (NE) India have witnessed a number of devastating earthquakes time to time. Guwahati being one of the important cities of Assam, NE India, with seismic zone V, has the highest zone factor (as per Indian Standard code). With the rapid urbanization and increasing population, the need of site-specific hazard assessment is in great demand in this city. This study attempts to determine the surface level Response Spectra (RS), Amplification Factors (AF) for PGA, Fourier Amplitude Ratios (FAR) for Guwahati. The soil properties at various locations are collected from the micro-zonation atlas for the region and these data are correlated with the V_S30 values. The material properties have been estimated using the standard reference curves. One dimensional (1D) equivalent linear approach has been adopted in this study. Spectrum compatible earthquake ground motions have been constructed as input motions for the ground response analysis, based on the Probabilistic Seismic Hazard Assessment (PSHA) results for the region. It has been observed that the local site-effects can significantly amplify the ground motions. Hence, the results are expected to give a good insight for the site-specific studies of Guwahati and be useful for future planning and design of structures.

The 1923 Kanto earthquake has not been sufficiently verified with respect to the strong ground motion generation areas with a period of less than 2 seconds, which is directly related to building damage. With the goal of determining the source rupture process by targeting the building collapse ratio, we first estimated the site amplification factors in the heavily damaged areas using the EMR method proposed by Kawase et al. (2018) and the VACF method by Ito et al. (2020). At sites with an observed collapse ratio of 100%, the amplification factor is large around a period of 1 second, and the site amplification factors may explain the occurrence of major damage, although the contribution of the source characteristic must also be taken into account. On the other hand, there are some sites where the site amplification is small but the damage is relatively large. In such municipalities, it is considered necessary to explain the damage exclusively by the source.

A ground motion model (GMM) predicts intensity measures (IM) at a site given an earthquake scenario, comprising source, path, and site effects. Among these, the site effect is typically constrained by site parameters such as V_S30 and site natural period. However, estimating site effects using a single parameter like V_S30 can lead to high uncertainties, as it might not adequately represent the complex geological structures at a site. To reduce uncertainties of non-ergodic site effects, accurate geotechnical information or seismic records of the target section are required. Generally, temporary seismometers are installed for earthquake preparation, response, and research purposes following a main earthquake. After the 2017 M_L5.4 Pohang event in the southeastern region of the Korean peninsula, several temporary stations were installed near the epicenter, recording multiple aftershock events. Using measurements from temporary seismic stations, we developed non-ergodic site effects for the temporary station locations and could interpolate them to create a more accurate IM map for the mainshock near the epicenter. We validated the developed IM map by comparing records that were not used for generating the IM map. The estimated IM at a test station had less error when compared to the IM predicted directly from the GMM.

During a strong earthquake, the local geological and geotechnical conditions can significantly alter the earthquake shaking in terms of amplitude, frequency content, and duration, a phenomenon often referred to as the site effect. In the past decades, the shear-wave velocity of the top 30 m (V_s30) and the fundamental period of vibration (T₀) were established as indicators for the seismic site effects. To improve understanding of subsurface ground conditions and their spatial variation in the Saguenay region, Canada, a Monte-Carlo based approach is proposed for V_s30 and T₀ modelling. First, a detailed probabilistic 3D geological model was developed considering three soil types: glacial till and postglacial fine and coarse sediments. The study area was modelled with 3D 75x75x2 m grid cells using sequential indicator simulation assigning probability of occurrence of each of the soil types to each cell. In parallel, a comprehensive Vs database was created based on invasive geotechnical measurements. Interval Vs probability distributions were determined for each soil type at each 2m depth. Monte-Carlo (MC) simulations were conducted considering Vs as the random variable. The resulting probability values were used to create V_s30 and T₀ maps and associated uncertainties. These results demonstrate the capacity of the proposed MC approach to incorporate the variability of the subsurface conditions in the seismic hazard assessment process.

Estimation of site amplification is important for a seismic hazard assessment. To develop site amplification prediction models, the models have utilized both measured and randomized profiles to compensate for the limited number of measurement database. One-dimensional site response analyses were performed to calculate linear and nonlinear responses, which were then separated into training and test data. While most amplification models are based on regression analysis using specific types of functions, this study utilized machine learning (ML) models with algorithms independent of prior functional forms. Random forest (RF) and deep neural network (DNN) models were used to train the model, and the test data were used to predict the amplifications. The DNN-based model shows more accurate results than the RF-based model, and the ML-based models were shown to outperform the existing regression-based model