Seismic imaging for crustal velocity and attenuation structure

Michael P Matheney, Purdue University

Abstract

In this study, seismic attributes are used to determine seismic attenuation and velocity structure of the subsurface. The seismic attributes considered include instantaneous frequencies, amplitudes and travel-times of selected phases. Complex trace analysis is used to estimate the instantaneous frequencies. Instantaneous frequency matching is then used to obtain the differential t* values between a reference pulse and the observed pulses. The differential t* values computed using instantaneous frequency matching, along with travel-time and amplitude information, are then utilized in a simultaneous inversion of seismic attributes for velocity and attenuation structure. Uncertainties in the model are estimated using covariance calculations and checkerboard resolution maps. The method is then applied to seismic refraction data along line A of the GLIMPCE Lake Superior experiment to determine velocity and attenuation structure. The inverted velocity structure was found to be similar to that found in previous studies. The prominent features include a central rift basin and a smaller northern basin along with an increase in velocity near the Isle Royale fault. The inverted attenuation model indicates that the basin structures are more attenuative than the surrounding areas. In the shallower portions of the model, microfracturing may have masked any compositional variations between the basins and the rocks along the Isle Royale fault. The instantaneous frequency matching procedure is then applied to seismic reflection data from line A of the GLIMPCE Lake Superior experiment. The computed differential t* values are converted to apparent ${\cal Q}\sp{-1}$ layers by a fitting procedure that simultaneously solves for all the interval ${\cal Q}$ values using non-negative least squares. The bootstrap method is used to estimate the uncertainties in the computed models. Comparison of the ${\cal Q}\sp{-1}$ profiles obtained from the seismic reflection data shows a close correlation with the results obtained from the seismic refraction data. This correspondence suggests that the effects of wave propagation and scattering on the apparent attenuation are similar for the two data sets. The attenuation model from the seismic reflection data is then related to the interpreted reflectivity structure. The highest interval ${\cal Q}\sp{-1}$ values ($>$0.01) were found near the surface, corresponding to the upper Keweenawan sedimentary rock sequence. Low ${\cal Q}\sp{-1}$ values ($<$0.0006) were obtained beneath the central rift basin. The surrounding crystalline rocks had ${\cal Q}\sp{-1}$ values similar to those found for the basaltic flows which included some interflow sedimentary rocks.

Degree

Ph.D.

Advisors

Nowack, Purdue University.

Subject Area

Geophysics

COinS