University of Bergen

Researcher in seismology; national project coordinator EPOS, Department of Earth Sciences

University of Bergen

Thesis Title: Deterministic seismic hazard assessments through hybrid ground motion simulation. Case study for Wenchuan, China and Izmir, Turkey

About

In areas with potential for large earthquakes, ground motion simulation based on earthquake rupture scenarios serve as an important tool to quantify the seismic hazard in terms of expected ground shaking. A kinematic hybrid broadband frequency simulation technique, combining deterministic low frequency modeling with stochastic high frequency modeling, is applied in a retrospect and a predictive study in order to assess the seismic hazard. The geographical focus in this thesis is on the region of the Mw 7.9 2008 Wenchuan, China earthquake and the city of İzmir, Turkey.

On 12 May 2008, a devastating earthquake of Mw 7.9 occurred in the Sichuan Province of China. The earthquake had disastrous consequences and took more than 80,000 lives. The rupture occurred along a 300 km long fault dipping northwest along the Longmen Shan fold-thrust belt, which separates the Tibetan Plateau in the northwest from the Sichuan Basin in the southeast. The earthquake is simulated using three available finite-fault slip models as input in the earthquake rupture scenarios. Acceleration values of the order of 1 g are simulated, which is close to what was recorded during the event. The simulation results reveal large variations in ground shaking due to the rupture complexity in terms of a variety of factors, including the location of asperities and the width of the fault plane. However, the simulated and observed ground motions are comparable in terms of ground shaking level and frequency content. Furthermore, from a field reconnaissance trip in October 2008, it is evident that extensive damage occurred over a wide area due to the shear size of the earthquake rupture combined with poor building practices. The combination of several factors, including mountainous landscape, strong ground shaking, extensive landslides and rock-falls, has exacerbated the human and economic consequences of this earthquake.

West-ward migration and counter-clockwise rotation of the Anatolian microplate in the Aegean Sea result in reactivation of several faults along the west coast of Turkey, which produce destructive earthquakes. İzmir, the third largest city in Turkey, has been destroyed by earthquakes several times in recent history, latest in 1778 and the seismic hazard for the city is assessed in three parts. The first part concerns identification of the faults that control the seismic hazard in the area. Nine different earthquake rupture scenarios along recognized faults in the area are defined, based on existing knowledge of source parameters from earthquakes in similar tectonic regimes. From the ground motion simulations the largest peak ground motions in İzmir are associated with two faults, the İzmir fault (normal) which lies underneath the city and the Tuzla fault (strike-slip) which lies southwest of the city.

The level and distribution of simulated ground shaking from ground motion simulation studies based on earthquake rupture scenarios is highly dependent on the input parameters used in the calculations. The second part of the hazard assessment for İzmir concerns the variability of the simulated ground motions, due to uncertainties in the input parameters. In this part the uncertainties of the input parameters are considered by calculating ground motions for a number of scenarios, while varying the input parameters. The level of ground motion is found to be most sensitive to the velocity model, seismic moment, rise time and rupture velocity. Stress-drop and attenuation control the change in ground motion levels to a large extent, if they are changed sufficiently. Low frequency ground motion is mostly affected by the location of rupture initiation, seismic moment, fault depth, rise time and crustal velocity model. The largest variability in the simulated ground motion is found close to the fault plane, coinciding with the center of İzmir. The standard deviation of all the simulations exceeds 200 cm/s2 and 20 cm/s for peak ground acceleration and peak ground velocity, respectively.

İzmir, being located on thick sedimentary deposits, is vulnerable to local site effects during strong ground shaking and it is therefore important to consider these when assessing the seismic hazard. In the last part of this study, local site effects are considered by evaluating transfer functions obtained from H/V spectral ratios of ambient noise. The ground motion, taking into account the soil layer, is then calculated as a convolution of the simulated waveform with the transfer function. Additionally, for the Karşıyaka district in İzmir, a soil column is modeled by theoretically calculating the soil response and comparing this to the obtained H/V spectral ratios. Considering the soil layer in the calculated ground motions suggests an increase in ground shaking due to the soil response. Furthermore, potential of soil deformation, such as liquefaction, during an earthquake is expected for the northern and innermost parts of İzmir Bay. Larger amplification of the seismic waves is calculated for a thicker soil column.

Although assessing the seismic hazard through ground motion simulations gives valuable information, such as expected ground shaking level and frequency content of the seismic waves, it is important to keep in mind that the question of probability of occurrence of the simulated earthquake is not considered in such a seismic hazard assessment. The applied methodology successfully reproduces the strong ground motion distribution and frequency content of seismic waves in retrospect studies and the simulation method is found appropriate in order to obtain realistic ground motion estimates for predictive studies. The importance of considering several possible rupture scenarios along the same fault when assessing the deterministic seismic hazard is shown, since uncertainties in the input parameters result in large variability of the simulated ground motion.