Dr. Nori Nakata of Stanford U. presents "Data-driven characterization of subsurface structure using Large-N array" at the MIT Earth Resources Lab on February 12, 2016.

"Earth's seismic velocity structure is heterogeneous at all scales, and mapping that heterogeneity directly contributes to the accuracy of understanding subsurface geology and geodynamics, and simulating linear path effects for ground motion prediction. Very dense receiver networks, such as a few thousands receivers within 10x10 km2, provide a unique opportunity to understand the local structure at scales that are relevant to high-frequency strong ground motion. I use body waves hidden in continuous ambient-field records of the dense arrays to create tomographic images of 3D velocity model. Compared with conventional surface-wave ambient-field tomography, body-wave tomography provides a 3D velocity model with higher spatial resolution. We find that the velocity model at Long Beach, California, has strong anisotropic heterogeneity in which the horizontal correlation length is five times greater than the vertical scale length. These small-scale heterogeneities are important for high-frequency motion prediction for earthquakes. I also demonstrate the importance of Large-N arrays for volcano monitoring to extract body waves propagating through magma bodies."