1. Modeling of ground motions and stress transfer caused by the December 26, 2004 Sumatra earthquake M. B. Sørensen 1, K. Atakan 1, J. Havskov 1, N. Pulido 2, A. Ojeda 3 1 Department of Earth Science, University of Bergen, Norway, 2 Earthquake Disaster Mitigation Research Center EDM, NIED, Kobe, Japan, 3 INGEOMINAS, Bogota, Colombia Abstract On December 26 th 2004, a devastating earthquake of M=9 occurred offshore Northern Sumatra. Due to the size of this earthquake and the accompanying tsunami wave, disastrous consequences have been observed at several countries around the Indian Ocean with a total death toll of more than 200 000. The tectonics in the region are characterized by the oblique, NNE oriented subduction of the Australian and Indian plates under the Sunda microplate with a rate of 6-6.5 cm/yr. This oblique convergence results in strain partitioning, where the trench perpendicular thrust faulting along the subducting slab accommodates the E-W component of the motion, whereas the N-S component of the motion is probably accommodated by the right-lateral strike slip faulting along the Great Sumatran Fault which passes along the western part of the main Sumatra island parallel to the volcanic chain. Source parameters of the December 26 th 2004 event have been used for modeling the resulting ground motions in the nearby affected regions. This will give information about the importance of ground shaking in the total destruction of places like Banda Aceh, Northern Sumatra, Indonesia. The modeling is performed for a multi-asperity finite fault using a hybrid procedure combining deterministic modeling at low frequencies and semi-stochastic modeling at high frequencies. In addition, stress transfer is modeled to estimate the resulting stress distribution and give an insight to the issues of future earthquakes along the neighboring segments of the subducting slab or along strike-slip faults on mainland Sumatra. Hybrid ground motion simulation methodology The December 26, 2004 earthquake has been modeled using a hybrid method for simulating ground motions due to a finite-extent earthquake source: Low frequency: Deterministic wave propagation from an asperity model in a flat layered velocity structure (Discrete Wave Number Method, Bouchon, 1981). High Frequency: Semi-Stochastic Simulation based on an asperity model. The model combines the stochastic methodology of Boore (1983) with the empirical Green’s function method of Irikura (1986), and a high frequency radiation pattern model (Pulido et al., 2004). Seismicity depth distribution Seismicity of the region during the modern instrumental period since 1964. The earthquakes within 0-33 km are shown in red. The green circles represent 33-66 km depth. The blue circles are epicenters of earthquakes at 66-99 km. The deeper earthquakes are shown in purple (99-122), turquise (122-155) and black (>155). The two focal mechanisms are the two solutions for the Dec 26, 2004 event from USGS and Harvard University. 6.5 cm/year Great Sumatra Fault 3.8 cm/y 5.2 cm/y Mentawi Fault 2.8cm/y Tectonic setting The earthquake of Dec 26, 2004 (M=9.0) occurred along the NW part of the Sumatra-Java subduction zone. The accumulated stress as a result of the relative plate motions is released through a rupture along the plate interface between the Australia-India plates and the Sunda micro-plate. Recent studies in the region have revealed that this segment of the plate boundary has been locked (Simoes et al., 2004) for quite some time since the three large earthquakes (M>8.0) that occurred in the same general area towards both north and south of the 2004 earthquake in 1833, 1861 and 1881 (Ortiz and Bilham, 2003). The aseismic slip at greater depths is postulated to be the result of pressure and/or temperature induced steady-state brittle sliding along the plate interface, possibly favoured by the fluids released from the subducting slab. The mega-thrust earthquake on Dec 26, 2004, was the latest manifestation of the repeated earthquake cycles along the convergent plate boundary. Strain partitioning The general plate convergence in the region is oriented towards NNE with an average velocity of 5.5 to 6.0 cm/year. This oblique convergence results in strain partitioning, where the trench perpendicular thrust faulting along the subducting slab accommodates the E-W component of the motion, whereas the N-S component of the motion is accommodated by right-lateral strike slip faults such as the Great Sumatran Fault, which passes along the western part of the main Sumatra island parallel to the volcanic chain (McCaffrey et al., 2000). A large amount of trench-normal stress has been released during the recent thrust events. However, significant trench-parallel stresses are still present which may be released in future large strike-slip events. References Boore (1983): Stochastic simulation of high frequency ground motions based on seismological models of the radiation spectra, Bull. Seism. Soc. Am., 73, 1865-1894. Bouchon (1981): A simple method to calculate Green’s functions for elastic layered media, Bull. Seism. Soc. Am., 71, 959-971. Irikura (1986): Prediction of strong acceleration motion using empirical Green’s function, Proceedings of the 7 th Japan. Earthq. Eng. Symp., 151-156. McCaffrey et al. (2000): Strain partitioning during oblique plate convergence in northern Sumatra: Geodetic and seismologic constraints and numerical modeling, Journal of Geophysical Research, 105, 28363-28376. McCloskey et al. (2005): Earthquake risk from co-seismic stress, Nature, 434, 291. Murphy and O’Brien (1977): The correlation of peak ground acceleration amplitude with seismic intensity and other physical parameres, Bull. Seism. Soc. Am., 67, 877-915. Ortiz and Bilham (2003): Source area and rupture parameters of the 31 December 1881 Mw = 7.9 Car Nicobar earthquake estimated from tsunamis recorded in the Bay of Bengal, Journal of Geophysical Research, vol. 108. Pulido et al. (2004b): Strong ground motion estimation in the Sea of Marmara region (Turkey) based on a scenario earthquake, Tectonophysics, 391, 357-374. Simoes et al. (2004): The Sumatra subduction zone: A case for a locked fault zone extending into the mantle, Journal of Geophysical research, vol 109. Yagi, Y. (2004): Source model for the December 26 th 2004 Sumatra earthquake, available at http://iisee.kenken.go.jp/staff/yagi/eq/Sumatra2004/Sumatra2004.html Figure from US Geological Survey, 2005. Figure from M. Raeesi, 2005. Earthquake source model Ground motion simulations are based on the fault model given by Y. Yagi (2004). The following source parameters are used: Epicenter:3.3 ºN, 95.78 ºE Source depth:13 km (75 km along dip) Strike:329 º Dip:10 º Slip:110 º M 0 :4 · 10 22 Nm Rise time:6 ± 2 s Rupture velocity:2 ± 0.5 km/s Attenuation:Q = 100 · f 0.8 f max :10 Hz Asperity stress drop (high slip):216 bar Asperity stress drop (int. slip):66 bar Background stress drop:57 bar Simulations are performed for a coarse grid of 12 stations and for an additional site (004) located near Banda Aceh, which suffered great damage from the earthquake and the following tsunami. Preliminary simulation results Acceleration (left) and velocity (right) waveforms and displacement spectra modelled for station 004 located near Banda Aceh in northern Sumatra. Preliminary simulation results Simulated peak ground accelerations (left) and peak ground velocities (right). Accelerations up to 0.5 g and velocities up to 400 cm/s are found for the northwestern part of Sumatra. Eye witness reports from the earthquake indicate an intensity of VII-IX in this region. Converting these intensities to PGA values using emperical relations (Murphy and O’Brien, 1977) gives values in agreement with the above results. Coulomb stress modeling Coulomb stress transfer has been modeled for a simple source model based on the earthquake scenario used for ground motion modeling. The plot shows coulomb stress change in the fault-near region for optimally oriented strike-slip faults. The positive stress lobe in northern Sumatra indicates that stress on the Great Sumatran Fault has increased after the December 26, 2004 event. Additional computations of the coulomb stress change for optimally oriented thrust faults are in agreement with the locations of the large earthquakes which took place on March 28 th 2005 (M=8.7) and April 10 th 2005 (M=6.8). Similar results are also obtained by McCloskey et al. (2005). Location of December 26, 2004 and following large earthquakes in 2005 along the NW part of the Sumatra-Java subduction zone