9416451 Somerville The objective of this study is to develop an empirical model for the prediction of ground motions around thrust faults, including differences between ground motions on the hanging wall and the foot wall, using strong motion data from the 1994 Northridge earthquake and other thrust earthquakes. Ground motions are known to be significantly larger on the hanging wall than on the foot wall of dipping faults. This was observed in the ground motions of the 1994 Northridge earthquake, where the peak accelerations in the northern San Fernando Valley, which lies on the hanging wall of the Northridge rupture, were larger than those recorded at equivalent distances on the foot wall side, which lies in the San Gabriel Mountains to the north. The peak accelerations from the Northridge earthquake were much larger (e.g. 40-50% larger at distances less than 30 km)than predicted by current empirical attenuation relations, even accounting for the 20-30% larger motions expected for thrust events. Knowing whether or not the ground motions from the Northridge earthquake are representative of ground motions from the Northridge earthquake are representative of ground motions expected in future thrust and/or blind thrust earthquakes has important implications for the development of empirical ground motion attenuation relations, and on the revision of seismic building codes. The project uses strong motion data from the 1994 Northridge earthquake and other earthquakes to quantify the effect of earthquake mechanism on horizontal and vertical response spectra for periods of 0 to 4 seconds, distances of 0 to 50 km, and magnitudes of 6 to 7.5. It uses an empirical analysis of recorded motions from thrust and oblique earthquakes, supplemented by numerical modeling. These two approaches compliment each other and will lead to more robust results than would result from using either approach by itself. The empirical model for the style-of-faulting factor considers the distance and magnitude dependence of the effect of earthquake mechanism as well as differences between ground motions on the hanging wall and foot wall. Numerical ground motion modeling is used to help understand the causes for the effect of earthquake mechanism on the ground motions and to guide the selection of the functional form used in the regression analysis to develop the empirical model. It permits the influences of various conditions, such as location of the site on the hanging wall or on the foot wall, the depth of rupture, stress drop, and rupture directivity to be tested separately. Numerical modeling is also used to evaluate whether the ground motions from the Northridge earthquake, which are larger than predicted by existing ground motion attenuation relations for thrust earthquakes, are anomalous for thrust events in general, and for blind thrust events in particular, or if they are representative of ground motions expected in future thrust and/or blind thrust earthquakes. ***
