Analysis and synthesis of East Himalayan earthquakes using a hybrid wavelet synchrosqueezed transform

The work presented in paper is focussed on the development of an accurate time-frequency representation of accelerograms, representing the seismic activity of the Himalayalan regions. The Himalayas have a long history of high levels of seismicity due to the movement of tectonic plates in the region. The effectiveness of the advanced time-frequency analysis (TFA) method known as Wavelet Synchro-Squeezed Transforms (WSST) has been exploited for processing accelerograms of the regions. Obtaining simultaneous time and frequency localizations with high levels of resolution is always challenging for non-stationary signals such as accelerograms. To have better time and frequency localization, the first part of the paper proposes an advanced TFA technique using WSST. The second part of work is then focused on devising a technique for accurate synthesis of the accelerograms to represent the actual and limited seismic records. A WSST based denoising is further proposed to remove the unwanted noise from the signals. Using the available actual recorded data from the region, respective synthetic accelerograms and response spectra have been developed for different magnitude and distance ranges using the WSST technique. The synthetic signals are then authenticated with the actual signal data using different Joint-TFA (JTFA) approaches. The effectiveness of the proposed WSST based synthesis method is further confirmed by higher structural similarity metric (SSIM) scores when compared with the time-frequency representations of the reference ground truth accelerograms.