ABSTRACT

About 10 years ago, a theory was developed in global seismology to account for the finite-frequency nature of seismic data from which delay times relative to a 1D reference Earth model are derived, known as finite-frequency traveltime tomography (FFTT). By taking frequency into account, a more accurate estimation of velocity is possible because the physics of wave propagation is more accurately being taken into account. FFTT is generally not applicable to controlled-source data because there is no requisite reference velocity model of the crust or near-surface that is capable of yielding realistic synthetic waveforms that are close enough to the recorded seismograms to estimate a meaningful delay time. As a result, an inverse method that uses a frequency-dependent traveltime calculated along the total path, as opposed to a delay time, is needed in order to model controlled-source (picked) traveltime data. I will develop a frequency-dependent form of traveltime tomography (FDTT) that has the same objective as the FFTT theory used in global seismology, namely more accurate velocity estimation. FDTT is based on a new method of forward calculation called wavelength-dependent velocity smoothing (WDVS). Although applicable to all controlled-source data, FDTT's most important application will be in near-surface studies (upper ~100 m) since these data often have long seismic wavelengths relative to the length scale of subsurface heterogeneities, and hence the greatest potential for ray (infinite-frequency) theory to be invalid. FDTT has already been developed and tested for the case of first-arrival times and 2D models. This award will allow for further testing on synthetic and real data, extension to 3D models, treatment of reflection times, a full-wavefield check and calibration of the WDVS algorithm, improved computational speed, and creation of a software package for free academic distribution.

Seismic traveltime tomography is used to image the Earth?s interior by determining the spatial variation in the velocity of elastic waves from the surface to the inner core. In the 1970's, seismologists adapted tomography from the medical community and until recently, the forward modeling component of traveltime tomography has always been ray theory, an infinite-frequency approximation of wave propagation. This is because ray theory is computationally efficient and applicable to arbitrarily heterogeneous media. In the past 10 years, global seismologists have shown that by exploiting the frequency content of the recorded seismograms, it is possible to obtain a more accurate estimation of the 3D velocity structure of the Earth?s mantle. Controlled-source data, as opposed to earthquake data, are acquired in applied studies of the Earth?s crust and near-surface in tectonic, petroleum, environmental and engineering studies. A different theory is needed to exploit the frequency content of controlled-source data since the crust and near-surface are much more heterogeneous than the Earth?s mantle, and therefore there is no reference model known in advance from which small perturbations would be sufficient to match the real data. The new method and corresponding software developed under this award will have its greatest impact in the societally relevant neo-tectonic (risk), environmental and engineering fields. This will include imaging shallow faults for earthquake hazard assessment, assisting in the remediation of a groundwater contamination site, delimiting the boundaries of unmapped waste disposal sites, and assessing the competence of the subsurface before large structures such as bridges and dams are constructed. The methodology will be made freely available as software to the academic community.

Please report errors in award information by writing to: awardsearch@nsf.gov.