Project Abstract
|
This work aims to develop and implement a stochastic description of fine-scale velocity structure in the SCEC Community Velocity Models (CVMS and CVMH). The trend of numerical wave propagation studies to higher frequencies (> 2 Hz) has been facilitated by sustained increases in computational power, and has created a demand for higher resolution velocity models. Building such models is challenging, in part because geologic and seismologic data indicate that fine-scale elastic inhomogeneities can be strong in sedimentary basins and have spatially anisotropic statistical distributions. While we have local measures of fine-scale velocity structure (down to meter scales) along boreholes with sonic logs, there is not a sufficient density of such samples to facilitate the development of a deterministic regional model. Thus, we are developing a statistical description of fine-scale velocity structure, informed by these local observations and geological correlations, to enhance the community models so that they can support higher-frequency simulations. Our current results define the standard variation (6.5%) in velocity that is represented in the wells but not the current CVM-H, and show that this variation is markedly non-Gaussian. In addition, we define vertical and horizontal correlation lengths for velocity structures within the sedimentary basin of 80 and 2000 meters, respectively. Our goal is to use these characteristics to implement a fine-scale velocity structure in the CVM-H to support wave propagation and strong ground motion simulations to higher frequencies. |