SIMULATION OF STRONG GROUND MOTION USING SEMI EMPIRICAL MODELLING TECHNIQUE

Abstract

Strong ground motion prediction is one of the indispensable prerequisites for earthquake resistant design of structures. Safe design of any structure in a seismically active environment require various engineering parameters which are obtained either directly from strong motion time histories or from simulated strong motion records. Strong motion data are not easily available at all construction sites therefore simulation techniques are required for generating useful strong motion data for such sites. Simulation techniques require several parameters of earthquake and other seismic information prior to the modeling of any earthquake ground motion in any tectonic setup. Estimation of such parameters is practically difficult task, especially in a region where we have limited information in hand. This has motivated a need for a technique of simulation of strong ground motion which should depend on the easily available parameters at any new site. The semi-empirical method was used for strong motion simulation of major to large earthquake in a broad frequency range by Kumar et al. (1999), Joshi et al. (2001), Joshi and Midorikawa (2004) and Joshi and Mohan (2008). This technique has never been tested for simulating ground motions due to great earthquakes. Modifications in the semi-empirical method have been made in the present thesis to incorporate the effect of radiation pattern and seismic moment of the target earthquake. These modifications have removed the dependency of semiempirical method on attenuation relation. The semi-empirical method has been also modified to resolve the obtained record into two horizontal components. The modified technique has been studied in detail to check the presence of various strong motion properties like directivity effect and variation of peak ground acceleration with respect to surface projection of rupture plane. Strong motion records have been simulated for the well recorded Niigata earthquake (Mw = 6.6) of October 23, 2004 to validate the modified technique. Iterative modeling suggests that rupture propagate bilaterally in northwestern direction at a depth of 13 km with rupture velocity 3.1 km/sec which is similar with the findings of Honda et al. (2005). Based on satisfactory results obtained from present simulation technique of this well recorded and well-studied earthquake, two earthquakes in Indian subcontinent viz., the Sikkim earthquake (Mw = 6.9) of September 18, 2011 and the Sumatra earthquake (Mw = 9.0) of December 26, 2004 have been studied to test the applicability of the modified technique. The technique is further applied to present a ground motion scenario due to a great hypothetical earthquake in the Andaman Island, India. The rupture ii plane is placed in the Andaman region for this hypothetical great earthquake and records have been simulated using both the modified semi-empirical and empirical Green’s function technique. The rupture responsible for the Sikkim earthquake has been placed at a depth of 44 km between Tista and Gangtok lineaments. The length and width of the rupture plane for the Sikkim earthquake are assumed to be 51 and 13 km, respectively. The strike of the rupture plane is assumed to be parallel to the Tista lineament and is 328°N which is close to that obtained from fault plane solution of this earthquake given by Global CMT. Acceleration records have been simulated for near-field as well as far-field stations. Several simulations from different source models and its comparison with observed records in the frequency range of 0.01–20.0 Hz indicate that the Sikkim earthquake was generated by a rupture originating at a depth of 47 km and propagating in southward direction with rupture velocity of 2.9 km/sec. The source of the Sumatra earthquake has been modeled using both modified semiempirical and empirical Green’s function techniques. The final rupture model of the Sumatra earthquake obtained after iterative modeling using modified semi-empirical and empirical Green’s function techniques has further been used to simulate both horizontal component of ground motion at PSI and MDRS stations which lies at an epicentral distances of 355 and 2060 km, respectively. It has been further observed that due to dependency of empirical Green’s function technique on aftershock record, simulations at a far-field station has been compared in the frequency range of 0.3 to 2.0 Hz; whereas modified semi-empirical technique simulations have been compared in the frequency range of 0.3 to 4.0 Hz for the great Sumatra earthquake. Iterative modeling of the source of the Sumatra earthquake with several possibilities of rupture parameters indicate that this earthquake was originated at 38 km depth and started propagating in northwestern direction with rupture velocity of 3.0 km/sec. The simulation technique developed in this work is further used to model a hypothetical earthquake of magnitude 8.5 (Mw) in the Andaman region of Indian subcontinent. This region lies close to the epicenter of the Sumatra earthquake. The causative rupture for this hypothetical earthquake is placed in the Source Zone 81 marked by Bhatia et al. (1999). This source zone has potential of generating 8.5 magnitude earthquake (Bhatia et al. 1999). The Andaman region itself has experienced earthquakes of both reverse and oblique strike-slip fault mechanism (Ortiz and Bilham 2003; Chatherine et al. 2009). Both NS and EW component of horizontal record has been simulated for reverse and oblique strike-slip type of earthquake source mechanism in this region. It has been observed that peak ground acceleration obtained from several iii records using both methods with different possibility of nucleation point lies within a range of 0.3 to 2.0 g for reverse source mechanism at POR station. This range of peak ground acceleration from simulated records for oblique strike-slip source mechanism is between 0.1 to 2.0 g at POR station. Several strong motion parameters are required for the estimation of seismic hazard in any region which is sometimes not easily available at all sites. This thesis presents modified semiempirical method which has advantage of using easily available parameters. The technique can be applied to a region having scarcity of observed strong motion data and can be effectively used for estimating earthquake resistant design parameters.