Derecke Palmer, 2010. "Characterizing the Near Surface with Detailed Refraction Attributes", Advances in Near-surface Seismology and Ground-penetrating Radar, Richard D. Miller, John H. Bradford, Klaus Holliger
Download citation file:
The tomographic inversion of near-surface seismic-refraction traveltime data is fundamentally nonunique. It is possible to generate tomograms, which range from the geologically improbable to the very detailed, all of which accurately satisfy the traveltime data. A comparison of the starting models with the final tomograms demonstrates that refraction tomography usually does not improve lateral resolution significantly. Therefore, if important geotechnical features are to be delineated, it is essential that they be included in the starting model, especially zones with low seismic velocities. Suitable detailed starting models for both traveltime and full-waveform inversion can be derived using a suite of parameters, generally known as seismic attributes. Refraction attributes can be computed readily from all near-surface seismic-refraction traveltimes, amplitudes, and waveforms, using the generalized reciprocal method (GRM) and the refraction convolution section (RCS). Furthermore, refraction attributes can be employed as a priori information to resolve nonuniqueness before the acquisition of any a posteriori information, such as borehole or other geophysical data. Narrow and wide zones with low seismic velocities are delineated with detailed attribute-based tomograms and are consistent with other refraction attributes derived from head-wave amplitudes and the RCS. Those zones are not detected with refraction tomograms which use low-resolution starting models, such as the smooth vertical velocity gradient. Additional models of the near surface, such as scaled density ratios and the P-wave modulus, can be computed from combinations of the refraction attributes. The use of a suite of attributes and combined attributes as well as the seismic velocity facilitates derivation of more comprehensive quantitative models of the near surface and thus more effective integration of seismic with borehole and other geotechnical data, using either multivariate geostatistics or full-waveform inversion.