Engineering characteristics of ground motions recorded by Bangladesh and north-east Indian strong motion instrumentation network

Abstract

Ground motion attenuation relationship is the most significant and useful topic in engineering seismology. It provides the background for seismic resistant design, seismic zonation map and seismic hazard analysis. Indian Meteorological Department (IMD) is operating a widespread network of instruments for recording strong ground motion in North-East Indian region and Bangladesh. This network has so far recorded around 1608 three-component (North-South, East-West and Vertical) time history records from 160 earthquake events with 536 records having magnitude from 2 to 8 from 2005 to 2017. The goal of this research is to acquire strong ground-motion (SGM) data and develop attenuation law for the study region. The widely used ground motion attenuation models proposed by Joyner and Boore and Bommer and Akkar are adopted to construct the empirical attenuation equations by multiple regression method. The predicted peak ground parameters are expressed as a function of magnitude, distance (epicentral distance or hypocentral distance) and site category. The model uses a magnitude-independent shape according to geometrical spreading and anelastic attenuation for the attenuation relationships. Ground-motion prediction equations (GMPEs) based on horizontal ground motions (peak ground acceleration, peak ground velocity, peak ground displacement) and Spectral Values (0.3 sec, 1 sec & 2 sec.) have been developed in this study for rock and firm soil sites. Also GMPEs based on vertical component have been developed. Ground motion attenuation relationships based on North-East Indian strong motion data have been developed for firm and rock soils. The acceleration predictive values of horizontal and vertical component for firm soil site are larger than those of rock site under the same conditions for a given earthquake event. Moreover, attenuation curves are magnitude-independent as well as have no distance and magnitude dependency on residuals. The predictive equations for rock sites (273 records) shows good correlation as standard deviations are low and it is 0.37σ for peak acceleration, 1.25σ for peak velocity and 1.30σ for peak displacement. Moreover, standard errors for acceleration response spectrum (ARS) are 0.52σ and 1.35σ for velocity response spectrum (VRS). For soft rock sites (63 records) recorded data, standard deviations are 0.35σ for peak acceleration, 1.25σ for peak velocity, 1.31σ for peak displacement, 0.52σ for ARS and 1.31σ for VRS. Firm soil sites (200 records) have the least standard deviation of 0.30σ for peak acceleration. Additionally, 1.30σ for peak velocity, 1.40σ for peak displacement, 0.59σ for ARS and 1.36σ for VRS. Thus these models are more suitable for the firm soil sites. The depth of the earthquakes have no effect on the attenuation equation developed according to the stated models. However, initial amplification value at closer distance is slightly higher for Joyner and Boore prediction than Bommer and Akkar. Though the equations decays linearly with no dependency on magnitude of the earthquakes. Near field earthquakes (epicentral distance less than 60 km) amplifies more than far field (epicentral distance more than 60 km). Additionally, far field attenuation line amplifies more for Bommer and Akkar model than the model from Joyner and Boore. Epicentral distance parameter effects largely for Bommer and Akkar model. At larger hypocentral distance, near field model decreases rapidly. Predicted models of this study are compared with other researches and shows high correlation with results from other authors. However, at small distance, this research shows slightly higher amplification of predicted peak acceleration but at distance from 5 km to 50 km, this study decays faster than other models predicted by famous researches. Thus, this model is more reliable than the others at higher distance sites