Relative-amplitude preserving processing for crustal seismic reflection data: an example from western Canada

Abstract

Most crustal-scale seismic reflection profiles are processed using methods that preserve reflection geometries but not true-relative amplitudes, and therefore do not provide direct constraints on the magnitude of impedance contrasts at major crustal boundaries, the effects of attenuation in the crust or the maximum time of signal penetration. Here we describe a new relative-amplitude preserving (RAP) processing scheme tailored for high-fold, land-based acquisition systems that can be used to provide some of these constraints. Our processing philosophy employs a statistical treatment of amplitudes, and is based on the following assumptions: (1) temporal decay of recorded amplitudes is proportional to the product of zero-offset geometrical spreading in a layered earth and attenuation based on a constant- Q model; and, (2) spatial variations in recorded amplitudes are surface-consistent. We apply this methodology to a profile from western Canada, where well logs from oil and gas exploration are available to calibrate observed seismic reflections. Using a spectral-ratio method tailored for noisy data, we obtain an estimate of ∼ 450 for apparent Q in the sub-Phanerozoic crust. Signal penetration is frequency-dependent, and is estimated to decrease from 20 to 13 s over the frequency range of 10 to 40 Hz. Application of our processing methodology leads to a stack section in which lower-crustal reflectivity is much more prominent than reflectivity in the middle crust, compared with conventional processing of the same data which yields a stack section characterized by conspicuous mid-crustal reflections. The P-wave reflection coefficient for the Moho is estimated to be about 0.1.