- Copyright © 2000 Society of Exploration Geophysicists
Editor's note: This series is designed to update the classic Pitfalls in Seismic Interpretation by Tucker and Yorstan, first published by SEG in 1973. Contact Steve Henry (email@example.com) for information about contributing to this series.
The fundamental value of horizon autopicking is that it quickly and accurately performs straightforward event correlations (in the words of C. Hewitt Dix, those correlations that “almost anyone can understand”), thereby allowing an interpreter to devote greater attention to more difficult and complex problems. In “Modern 3-D seismic interpretation” (TLE, 1998), Dorn thoroughly describes autopicking (also known as autotracking) as one of the most important tools available to an interpreter for efficient and accurate interpretation of a 3-D volume. He remarks on the importance of knowing the type of autopicker being used (feature versus correlation tracking) and the path it follows through the data. An interpreter's lack of attention to either of these elements of autopicking, or failure to assess the suitability of a data set for autopicking, can give rise to serious interpretation pitfalls. The pitfalls discussed in this paper are proper selection of the input control grid for autopicking, the “directional” behavior of autopickers, autopicking in areas with low signal/noise, and recognition of elements of geology that are not suitable for interpretation by auto-picking.
Control grid for autopicking
The first and most obvious pitfall associated with autopicking is failure to optimize the input grid of control points. This grid can be either too broad (the spatial frequency of control points is insufficient for unique sampling of the picked horizon) or too tight (the spatial frequency of control points is greater than needed for unique sampling of the picked horizon). In the former situation, erroneous picks in the output horizon can be caused by failure to satisfy autopicking criteria between control points—said another way, in the absence …