Implications of Seismic Attribute Computations From Depth-Migrated Data

Abstract

Abstract

Prestack depth migration (PSDM) is a highly popular imaging process because it enhances seismic images of structurally complex subsurface, by handling both vertical and lateral velocity variations. Thus, in principle, seismic attributes extracted from depth-migrated data are more reliable than those from time-migrated data. However, there are significant conceptual differences in the way seismic attributes are calculated from depth-migrated data. For instance, vertical sampling is no longer in milliseconds but in meters. Attribute calculations are no longer in frequency (cycles/s) but in wavenumber (cycles/m). Using constant windows to compute such attributes is no longer valid due to the wavelet stretching produced by the rapid velocity changes accounted for during PSDM. A common solution to circumvent these issues is to convert depth-migrated data to time and only then compute seismic attributes, which is valid if seismic attributes are used for qualitative interpretation but not for quantitative interpretation. In this work, we discuss the computational implications of extracting seismic attributes from depth-migrated data and to what extent the interpreter can rely on seismic attributes calculated directly from PSDM data. To illustrate the implications for attribute extraction we present examples of time- and depth migrated synthetic and field data. Not surprisingly, frequency-based attributes are the most affected ones and corrections for steeply dipping interfaces need to be implemented. We believe geoscientists can benefit from this discussion given the increasing availability of depth-migrated data.

AAPG Datapages/Search and Discovery Article #90260 © 2016 AAPG/SEG International Conference & Exhibition, Cancun, Mexico, September 6-9, 2016