SCEC Award Number 12037 View PDF
Proposal Category Collaborative Proposal (Integration and Theory)
Proposal Title Iterative Improvement of CVM-H 11.9 Using Spectral-Element and Adjoint Methods
Investigator(s)
Name Organization
Carl Tape University of Alaska, Fairbanks John Shaw Harvard University Andreas Plesch Harvard University Emanuele Casarotti Istituto Nazionale di Geofisica e Vulcanologia (Italy)
Other Participants
SCEC Priorities 6a SCEC Groups CS, USR, CME
Report Due Date 03/15/2013 Date Report Submitted N/A
Project Abstract
 The multi-year project objectives are to perform an iterative tomographic inversion with CVM-H 11.9 as a starting model. Our emphasis this year was to document the spatial variability of the numerical resolution of our model. This is a critical initial step for any comparisons with other synthetics or with data. We are able to highlight the most challenging regions of CVM-H 11.9 to resolve.
Intellectual Merit Our results quantify the numerical resolution of our 3D seismic velocity model and mesh. Typically this information is only roughly known. We are using CVM-H 11.9, thereby promoting and hoping to improve one of SCEC's primary community products.
Broader Impacts We continue our solid collaboration between geologists at Harvard and seismologist(s) at University of Alaska Fairbanks. Our overall emphasis in on the seismic structure of the uppermost 10 km, notably basins. By better characterizing these structures, we will be able to better model ground motion in these regions.
Exemplary Figure Figure 3. Quantification of numerical resolution for the coarse mesh used for \cvm\, based on the vertical component of synthetic velocity seismograms. The source is a north-striking, vertical strike-slip event at 8~km depth. (Top) For each station (and each component), we construct a (h_{\rm dur})$ curve, from which we estimate the station-specific numerical resolution, plotted in the histogram.
(Bottom) Simulation-based numerical resolution as a function of space. We see that the poorest resolved regions (highest {\rm dur}$) are associated with sedimentary basins, as expected.