Project Abstract
|
Cybershake-based ground motions are being considered for engineering application in the greater Los Angeles area and nationally. One of the principal motivations for these applications is a belief that Cybershake-based estimates of basin effects on long-period ground motions in Southern California and elsewhere are more accurate than those from semi-empirical ground motion models (GMMs). This belief is, in fact, a hypothesis that we seek to test. Doing so is a long-term research objective. This contract supported the first year of work in which a set of initial simulations was performed and preliminary data analyses were conducted. Simulations were performed for 13 events with a magnitude range of M3.99 to M5.39, each in southern California in the vicinity of one or more well-recorded recent earthquakes.
The data analyses involve computing residuals relative to a regionally unbiased ground motion model. Mixed effects analyses of the residuals allow estimation of so-called “site terms”, which represent the average site response implied by the data. As part of related (and separately funded) research, we use these site terms to analyze basin effects from empirical data. These results challenge some aspects of current models, particularly for shallow soil sites. They also provide a baseline against which to evaluate the simulated ground motions. Computed ground motions from the simulated events are analyzed in the same manner as recordings, with site terms evaluated. Results to date indicate some general similarities in the scaling of site amplification with depth in basins between data and simulations. |