A spatially compact source designature filter

作者:Agudo Oscar Calderon*; Caprioli Philippe; van Manen Dirk Jan
来源:Geophysics, 2016, 81(2): V125-V139.
DOI:10.1190/GEO2015-0259.1

摘要

Source arrays are usually tuned for optimum primary-to-bubble ratio of the output. In addition, they may also be designed to mitigate source ghost effects, but such configurations introduce directivity effects as the generated downgoing wave-field varies with the source angles. We have investigated a practical source signature deconvolution operator, valid for any marine source array, which is able to remove source ghosts, residual bubble, and directivity effects from seismic data. The designature operator was designed to be a small wavenumber approximation of the inverse of the far-field signature of a source array, which involved the array geometry, the notional source signatures, and the firing synchronization of the sources. We have found that such an approximation leads to the design of spatially compact source designature (CSD) filters able to correct for shot-to-shot perturbations of the sources' output and reduced their smearing due to the limited spatial aperture of the filters. They also had the potential to account for some source directivity and source ghost effects when applied to common receiver gathers. Our method was then further studied for a particular synchronized multilevel source array and azimuthal variations were neglected. First, tests on synthetic data were performed and the benefits and limitations of the approach were analyzed. Second, the method was tested on a field data set under further approximations that simplified the processing flow (1D earth and average notionals) and which enabled the direct application to common shot gathers. Our results have determined that the expected small - but nevertheless observable - improvements of source designature even at the target level improved event continuity, dephasing, and sharpening of the wavelet. Additionally, the CSD filter was used in conjunction and compared with the current best-practice data processing techniques.

  • 出版日期2016-4
  • 单位ETH, Zurich