Sensitivity study of seismic reflection/refraction data

Abstract

S U M M A R YThe reliability of information obtained from surface seismic measurements is an important factor to be considered before one attempts to depth migrate or invert seismic data. The uniqueness of the retrieved model can only be achieved if the observed seismic data contain information on all the wavelengths of the Earth structure.In this study we investigated the sensitivity of model parameters obtainable from elastic multi-offset seismic data. We perturbed a background model, which was defined in terms of longitudinal and shear velocities and density, with quasi-sinusoidal perturbations. The reference and perturbed synthetic seismograms were calculated in the intercept time and slowness (7-p) domain. The seismic waveform misfit between these two synthetic seismograms computed in a least-squares sense was used as a quantitative measure of the sensitivity of model parameters. We found that middlewavelength variations of model parameters are sensitive to critical-angle seismic data. This was further confirmed by using a 1-D waveform inversion. These results suggest that all wavelengths of model parameters can be retrieved from waveform analysis of seismic reflection/refraction data: the long wavelengths from traveltime data and the short to medium wavelengths from waveform analysis. The non-uniqueness associated with the information gap for middle wavelengths, as previously proposed for seismic reflection data, can therefore be avoided.

Full waveform inversion in the time lapse mode applied to CO2 storage at Sleipner

“…To avoid getting trapped into local minima, a data windowing inversion strategy is applied, as suggested by Shipp and Singh (2002). In order to reduce the lobes it is also important to include long offset data in the inversion (Neves and Singh 1996). We first focus the inversion on a window associated with the upper half of the model for 5 iterations.…”

“…This is because only the medium to long scale of the overburden velocity is required for accurate inversion of the reservoir structure (Jannane et al 1989;Neves and Singh 1996). Test inversions of these data show that the locations and magnitudes of velocity changes in the reservoir are not altered by changes in the small-scale features of the overburden velocity significantly.…”

Imaging of intrabasalt and subbasalt structure with full wavefield seismic tomography

“…Following procedures set out by Jannane et al [1989], Neves and Singh [1996], and Primiero [2002] we have analyzed the sensitivity of the seismic wavefield to these perturbations. Sensitivity measures the least squares difference across shot and receivers between wavefields generated from the starting W ref and perturbed W pert according to S varies between 0 and 1, with a value of 0 indicating that the wavefields are identical, and with high values of S indicating that the wavefield is sensitive to the perturbations.…”

Two-dimensional full wavefield inversion of wide-aperture marine seismic streamer data

“…The final model from this far offset inversion was then used as the starting model for a mid offset range inversion (4 to 8 km offsets). For this particular example, within this offset range most of the critical energy is located and this provides strong constraint on the medium scale model features (Neves & Singh 1996) as shown by the final inverted model (Fig. 1e).…”