Attenuation Characteristics and Seismicity Analysis of Himalayan Region with Linear Joint Time-frequency Analysis Techniques
About 59% of the land area of the Indian subcontinent has high chances of moderate to severe earthquakes, of which the Himalayas are the most seismically active. Regardless of the high seismic activity, the seismic network in the region is largely insufficient. Macro-level seismic zonation of the Himalayan region, along with the synthesis of Himalayan earthquakes using novel Linear Joint Time-Frequency Analysis (LJTFA) techniques will be the primary focus of this study. For seismic zonation, estimation of Peak Ground Acceleration (PGA) at bedrock level is important to understand the seismic hazard in the region. For this, region-specific Ground Motion Prediction Equations (GMPEs) are generated using multiple-regression analyses, separately for North and Central Himalayas (NCH) as well as North East Himalayas (NEH), considering the difference in their tectonic setting, purely based on recorded strong motion data. An updated larger dataset is used in the present study, making the new equations applicable to a larger magnitude and distance ranges, thus overcoming a major limitation of currently available region-specific GMPEs. After validation, the new GMPEs are used for SHA considering various source models. Based on the regional seismic source zoning using the Gutenberg-Richter (GR) parameters, the NCH is delineated into 5 zones and the NEH into 4. The whole study area is divided into grids of size 0.05° × 0.05° and the PGA at the center of each grid point is estimated using deterministic and probabilistic approaches, using the region-specific GMPEs. Seismic Hazard maps are then generated using a GIS platform which showed that the PGA estimated in the regions is comparatively higher than what is reported in the codal provisions for design horizontal acceleration. Seismically active regions with an insufficient strong motion network require synthetic ground motions and response spectra with reliable frequency contents for seismic design, geophysical studies, and seismic analyses. Temporal distribution of the frequency contents of a multi-component signal like seismic motions are not captured and well-represented in commonly used Fourier Transform techniques. To understand the non-stationary behavior of earthquake motions in terms of the temporal variations of amplitude and frequency contents and generate synthetic ground motions, a Linear Joint Time-Frequency Analysis (LJTFA) technique is advocated. For this, actual recorded time-histories from NCH and NEH are classified into 19 and 11 different Magnitude and Distance (MxDy) categories respectively. All signals in these MxDy categories are normalized and Gabor Transform (GT) is applied to transform and evaluate the Mean Gabor Amplitude Coefficients (MGAC) of all the normalized actual time-histories in each of the above-mentioned categories. Using an inverse transformation process; Gabor Expansion (GE), the MGAC are used to reconstruct and synthesize a time-history which is representative of each of the categorized MxDy combinations. The synthetic signals do not compromise on the quality of their spectral and frequency contents, thus yielding reliable seismic motions. A quantitative comparison between the actual and synthetic response spectra is made and a statistically good fit is observed between the actual and synthetic response spectrum developed. Further, Fourier Transforms of the actual and synthetic signals are also compared for a qualitative comparison. This further showed that the synthetic waves can represent reliably the wave propagation characteristics, tectonic setting, and fault characteristics of the actual waves transmitted in the region, preserving its temporal frequency and amplitude energy concentrations.
- Civil Engineering (CiE)