Seismic atributes and their classification by baroes104


Seismic atributes and their classification

More Info
									                Seismic attributes and their classification
                                                      ALISTAIR R. BROWN
                                                 Consulting Reservoir Geophysicist
                                                           Dallas, Texas

   A ttributes have proliferated recently      information is time, amplitude, fre-         ability. Most attributes are derived from
   with different selections available on      quency, and attenuation, and these form      the normal stacked and migrated 3-D
   different workstations. What do they all    the basis of the attribute classification    data volume but variations of basic
   mean? When do we use one and when           shown in attached hierarchy (Figure 1).      measurements as a function of angle of
   another? The answers to these questions        As a broad generalization, time-de-       incidence (and hence source-to-receiver
   are not easy but the first step is to un-   rived attributes provide structural infor-   offset) provides a further source of in-
   derstand what our options are, and          mation, amplitude-derived attributes         formation. The principal examples of
   herein lies the purpose of this article.    provide stratigraphic and reservoir in-      these prestack attributes are AVO.
       An attribute is necessarily a deriva-   formation. Frequency-derived attributes         Poststack attributes can be extracted
   tive of a basic seismic measurement. All    are not yet well understood but there is     along one horizon or summed over a
   the horizon and formation information       widespread optimism that they will pro-      window. The latter provides the concept
   attributes available are not independent    vide additional useful reservoir infor-      of a formation attribute. In some cases
   of each other but simply different ways     mation. Attenuation is not used today        the window is a constant flat time inter-
   of presenting and studying a limited        but there is a possibility that in the fu-   val so that the display is effectively a
   amount of basic information. That basic     ture it will yield information on perme-     thick time slice, sometimes termed a
                                                                                                       stat (statistical) slice. The
                                                                                                       window may be of a constant
                                                                                                       time interval but hung from
                                                                                                       one structurally-interpreted
                                                                                                       horizon so that the window
                                                                                                       properly follows a reservoir.
                                                                                                       The window may also be the
                                                                                                       interval between two struc-
                                                                                                       turally-interpreted horizons,
                                                                                                       for example the top and the
                                                                                                       base reservoir reflections.
                                                                                                       Within the window, values
                                                                                                       may be summed to produce a
                                                                                                       gross attribute measurement,
                                                                                                       only some of the information
                                                                                                       may be extracted to provide a
                                                                                                       selection attribute, or the vari-
                                                                                                       ation of the attribute within
                                                                                                       the window may be measured
                                                                                                       to assess a distribution.
                                                                                                          I know people have tried
                                                                                                       other classification ap-
                                                                                                       proaches but I recommend
                                                                                                       this one. Do you find this use-
                                                                                                       ful? What novel applications
                                                                                                       have you put attributes to re-
                                                                                                        cently? Do you feel the list
   Figure 1. Attribute classification.                                                                 has significant omissions?


To top