Abstract
Diffractions can provide information about small scale heterogeneities. However, in many cases, the diffracted energy is masked by the stronger reflections. Thus, techniques for separating diffractions from reflections are desired in seismic processing. In this thesis, different approaches to diffraction enhancement are discussed, but with special emphasis put on the Double Square Root (DSR) operator employed here. A complete workflow has been established for diffraction separation and imaging based on standard seismic reflection data. In this work, controlled data sets of various complexity have been employed to test out this procedure. Data from two simpler models were calculated using a finite-difference code from the NUCLEUS package of PGS. In addition, controlled data associated with the well-known SIGSBEE 2A model was employed. This latter model represents a complex subsurface including a salt body. For all models investigated, diffraction separation seemed to work well employing the proposed workflow.