Earthquake Ground Motion and 3D Georgia Basin Amplification in Southwest British Columbia: Shallow Blind-Thrust Scenario Earthquakes

Abstract

Finite-difference modeling of 3D long-period (> 2 s) ground motions for large (M w 6.8) scenario earthquakes is conducted to investigate the effects of the Georgia basin structure on ground shaking in Greater Vancouver, British Columbia, Canada. Scenario earthquakes include shallow blind-thrust North America (NA) plate earthquakes, simulated in locations congruent with linear clusters of shallow seismicity, that is, potential active faults. A slip distribution model of the M w 6.7 Northridge, California, blind-thrust earthquake, with the hypocenter modified to 5 km depth, is used to characterize the source rupture process. Two sets of simulations are performed for a given scenario earthquake using models with and without Georgia basin sediments. The ratio of predicted peak ground velocity (PGV) for the two simulations is applied here as a quantitative measure of amplification due to 3D basin structure.

Basin Resonance and Seismic Hazard in Jakarta, Indonesia

“…SEM is very effective at achieving high accuracy even for realistic earth models, and is therefore applicable for a wide range of applications in seismology (Komatitsch & Vilotte [16]). This approach to scenario ground modelling is similar to that used by Molnar et al [40] and Molnar et al [41] to study seismic wave interaction with 3D structure of the Georgia Basin, British Columbia, Canada.…”

Earthquake Ground Motion and 3D Georgia Basin Amplification in Southwest British Columbia: Deep Juan de Fuca Plate Scenario Earthquakes

“…Shorter period ground motions are not resolved, limited by the grid spacing and minimum V S chosen for the 3D basin model according to a ≥ 5 node per minimum shear wavelength rule-of-thumb commonly used for fourth-order FD schemes. Overall the work presented here (and in Molnar et al, 2014) represents an important step toward quantifying the effect of the Georgia basin on earthquake ground motion in southwest British Columbia.…”

“…In the base elastic 3D model, a northeast-trending velocity contrast occurs beneath Greater Vancouver, which is not supported by geological and structural information, but rather results from extrapolation of the 1 km gridded V P model of Ramachandran et al (2006) to the surface. When the base elastic 3D model is used in FD simulations of the Nisqually earthquake, good agreement is obtained between synthetic and empirical waveforms in the Seattle basin region (Molnar et al, 2010), because significant effort had gone into validating the 3D model (Frankel and Stephenson, 2000;Hartzell et al, 2002;Pitarka et al, 2004;Frankel et al, 2007Frankel et al, , 2009). However, synthetic waveforms overpredict Nisqually waveform amplitudes in the Georgia basin region by a factor of 2.1 (Molnar et al, 2010).…”

“…When the base elastic 3D model is used in FD simulations of the Nisqually earthquake, good agreement is obtained between synthetic and empirical waveforms in the Seattle basin region (Molnar et al, 2010), because significant effort had gone into validating the 3D model (Frankel and Stephenson, 2000;Hartzell et al, 2002;Pitarka et al, 2004;Frankel et al, 2007Frankel et al, , 2009). However, synthetic waveforms overpredict Nisqually waveform amplitudes in the Georgia basin region by a factor of 2.1 (Molnar et al, 2010). Therefore, in order to update V P in the upper 1 km of the base elastic 3D model in the Georgia basin region for the modeling work conducted here, all nonconfidential government geological and geophysical datasets, as well as the higherresolution (600 m gridded) tomographic regional model of Dash et al (2007), were collected, converted to V P estimates (if required), and merged (details in Molnar, 2011).…”

“…The shallow interplate event causes very large ground motions (frequency bandwidth of 0.2-4 Hz) along the path from the coast to the Mexican interior due to interference of multiple crustal arrivals (Lg phase), whereas ground motions for the deeper inslab normal-faulting event demonstrate simple attenuation with increasing distance, that is, the Lg phase is too small. This paper presents FD simulations of long-period (> 2 s) ground motions computed for scenario deep JdF plate earthquakes in southwest British Columbia in a regional 3D velocity model of the Georgia basin, Molnar et al (2014) deals with shallow (5 km) crustal North America plate scenario earthquakes (Molnar et al, 2014). This research provides the first detailed investigation of 3D earthquake ground motion for a sedimentary basin in Canada.…”