Structural interpretation - 3-D Seismic Interpretation - page 71

below the well bore, which can be compared directly with an arbitrary line selected
from the surface seismic dataset. Note, however, that this technique does not work for
a horizontal borehole; if there is no vertical separation between levels, it is not easy
to separate the upgoing and downgoing wavefields, as there is no difference between
them in the time-depth plot.
Although the VSP is often the best way to establish the tie between the surface
seismic and the well information, it does have one disadvantage compared to the well
synthetic. This is that it gives no real insight into how reflections are caused by changes
in velocity and density values from one formation to another. This knowledge is vital,
as we shall see in chapter 5 , if we want to understand the likely causes of changes in
seismic reflection amplitude or character from one part of the survey to another, which
may allow us to predict lithology, reservoir quality or porefill.

Workstation interpretation

3.2

Having identified some horizons that are significant for understanding the geology and
prospectivity of an area, the next task is to map them across the survey. Before carrying
out any detailed work, it is useful to inspect the volume as a whole to get a general
impression of structural and stratigraphic features of interest. This can be done by using
the volume visualisation techniques discussed in chapter 7 . Increasingly, interpretation
is being carried out in this environment (Kidd, 1999) . However, much detailed work is
still carried out by picking the two-way reflection time to various horizons on some or
all of the traces of the survey. These picked horizons are fundamental to the attribute
measurement work discussed in chapter 5 .
In the earliest days of 3-D survey, horizons were picked using methods carried over
from interpretation of grids of 2-D lines. These were presented to the interpreter as a
stack of paper prints; he would mark up the horizons of interest on a line through a well
location and then follow them along the line to intersections with other lines, where the
picks would be transferred to the crossing lines. By working round a loop of intersecting
lines, it would be possible to get back to the starting point, where it could be checked
that the interpretation was consistent around the loop. Interpretation would proceed by
following the horizons round a series of such loops until they had been picked on all
the lines (see McQuillin et al ., 1984 , for a more detailed explanation with examples).
The analogous method for 3-D data was to make paper displays of all the inlines and
all the crosslines, with the idea of interpreting them in a similar way. However, it is easy
to see that the effort needed for a full manual interpretation is prohibitive. Suppose we
have a quite small rectangular survey consisting of 500 lines, each with 1000 traces.
Then there would be 500 inline sections and 1000 crossline sections. The number of
intersections to be checked would be 500 000. Assuming the data were easy to pick,
it might be possible to verify the intersections at a rate of, say, 2000 per working day,