Ground motion processing and observations for the near-field accelerograms from the 2015 Gorkha, Nepal earthquake

Processed accelerograms define the ground motion input for a variety of engineering and scientific analyses. The Mw 7.8 Gorkha, Nepal earthquake of 25 April 2015 currently has data from six near-field seismograph stations, recorded from within Nepal, available in the public domain. However, only the USGS sensor from the Kantipath (KATNP) station in Kathmandu has a processed data set available. The remaining acceleration time series are only provided in raw form, which require additional signal processing to be properly utilized for analysis. This paper discusses three different processing schemes for the accelerograms that may be used based on the intended application. Based on these processed records, some observations of the ground motions are then discussed. Fourier and response spectra results are presented to illustrate the effects of the processing procedure and the frequency contents. The processed ground motions are compared to applicable ground motion prediction equations (GMPEs) for peak ground acceleration (PGA) and velocity (PGV). It is shown that existing GMPEs generally overestimate the PGA and PGV. Due to the presence of long period pulse-like behavior in the time series, two pulse extraction models are used to estimate the pulse parameters, including the pulse period (TP). By comparing the pulse period to an MW-TP relationship from the literature, it is shown that the recording from the rock site is an excellent match, while the other five soil sites have shorter periods than predicted. The MW-TP relationship shows longer pulse periods in rock sites than soil sites for magnitudes greater than MW 7.8, which is also observed from the results of this study. The observed long period peaks in the response spectra are attributed mainly to directivity effects.