Multiple prediction without prestack data - an efficient tool for interpretive processing

Multiple-attenuation during data processing does not guarantee a 'multiple-free' final section. Although a great deal of effort has been given to the problem of multiple-suppression (Berryhill and Kim 1986; Wiggins 1988; Verschuur et al. 1992; and others), perfect solutions still do not exist. When the subsurface structure is complex, the remaining multiples will be difficult to recognize, especially after the data have been migrated. In addition, many of the available multiplesuppression techniques are restricted to two-dimensional geometries and do not attempt to handle interbed and complex multiples (which consist of a combination of interbed and surface-related multiples). These restrictions further increase the possibility of obtaining undesired multiple energy in the final sections. Generally speaking, multiple prediction is used as an initial step for multiple suppression. The predicted multiples must be very accurate, both kinematically and dynamically, since the suppression operation involves subtraction of multiple energy from the recorded data. Therefore, standard multiple-prediction algorithms require manipulation of prestack data or an accurate forward-modelling technique based on a given interval velocity model of the subsurface. While interpreting data from areas where significant multiple energy has been recorded, the interpreter must rely on the success of the multiple-suppression operation. When it is suspected that a certain event is a multiple, it is difficult to verify this using standard interpretation tools. We believe that multiple prediction and identification can play an important role in seismic interpretation. The main obstacle preventing multiple-prediction procedures from being routinely used by interpreters is the necessity to access prestack data. Velocity analysis in an interpretive process that may become difficult to carry out in the presence of multiples (Gasparotto and Lau 2000). It is often the analyst's decision to distinguish between primary and multiple energy while picking velocities. If the multiple-prediction procedure can provide information on the velocity that will optimally stack the multiple energy, it could be used as a guideline during velocity analysis to help in avoiding erroneous velocity picks. This study is based on the assumption that if the prediction is not aimed at providing input to multiple-suppression algorithms, then it can become a purely kinematic procedure. Furthermore, without prestack data, interactive algorithms, highly suitable for interpretive work, can be developed on the basis of the proposed procedure. In the following, we briefly describe the concept of the prediction method. We then show how the need to access prestack data is avoided, thus permitting fast interactive prediction after stack and/or migration. Finally, we suggest a practical workflow to promote prediction during interpretation and standard velocity-analysis sessions. All the above procedures will be illustrated using synthetic and field data examples.