An example from the Northwestern Shelf of Australia
Three very important components of seismic processing are; 1) velocity analysis, 2) normal moveout (NMO), and 3) stacking. We will define and discuss these subjects using data collected from the shelf of NW shelf of Australia. The data was processed using ProMAX software. Before and after examples are included to illustrate the different processes. The importance of these components will be shown through historical perspective and their relation to modern seismic interpretation.
VELOCITY ANALYSIS
Velocity Analysis is the calculation of stacking velocity (Vs) or normal moveout velocity (Vnmo) from measurements of normal moveout (Sheriff, 1991). As the source to geophone distance increases the time for a signal to arrive increases, and when plotted on a time vs distance graph the data forms a hyperbola. There is a certain velocity value, dependent on the properties of the media being looked at, with which one can cause the hyperbolic function to change to a straight line (see NMO section). The determination, by various methods, of this particular value is velocity analysis.
There are several ways in which velocity analysis is usually performed: conventional method; panel method; and picking (Sheriff, 1995). Each of these methods allows one to determine the velocity value needed, with varying degrees of resource expenditure, time investment, and error allotment for the value.
The conventional method is used in areas in which the velocity is believed to increases with depth. Computer anaysis allows one to see peaks in the data which correspond to stacking velocities. These peaks can be represented in several ways depending on the program used. A weighted average velocity or RMS velocity is used to analyze waves that follow paths to receivers that are offset from the source. Computers perform the analysis using a collection of trial velocities that can be manually or automatically picked. ProMAX displays four different tiles or panels for us. Selection of velocity values can be made in the Semblance Tile or the CVS Tile (Constant Velocity Stack). The Semblance Tile is actually a contour map of semblance coefficients. The trial velocity corresponding to a high semblance coefficient is the RMS velocity in the zone above that reflector. A semblance coefficient map can tell us how many reflectors are present and the RMS velocities for waves reflected from them (Robinson, 1988). The CVS tile computes and displays constant velocity stacks for analysis. The other two tiles shown are the Gather Tile and Max Semblance/Interval Velocity. The interval velocity is calculated by the Dix Equation and is updated as soon as those selections are made. The Gather Tile shows the common offset supergathers and the effect of our next subject-NMO. (see Tile Display )
Peaks that are associated with the highest reasonable velocities tend to represent primary reflections while those associated with lower velocities tend to represent multiples (Sheriff, 1995). The velocity measurements from an area may be looked at in a restricted interval window, in which case the velocity analysis tends to be nondefinate, or in a large interval window (using more data), in which the analysis may become better defined but the measurements are then averages over a sizable area (Sheriff, 1995). A compromise must be made between the two.
Velocity panels are simply another way of visually determining which stacking velocity is most reasonable. In this method the data are plotted several times with each plot being based on a different stacking velocity, and then viewed side by side. This method is often used as a check on the interpretation derived from the conventional method.
The picking method involves diminishing the considerable amount of time and resources expended in velocity analysis due to the extensive number of calculations needed. In this method locations to be run are judiciously picked based on the best geologic information available and a few simple rules in order that analyses not be wasted on noisy areas and geologic changes be adequately sampled. The main objective is soley achieving a good stack and since stacking can tolerate significant velocity errors this method can be used and still result in important information not being lost (Sheriff, 1995). Significant potential for error does exists, mainly where the picker knows little about the geology of the area, but errors are becoming less frequent because workstations allow adjacent analyses to be viewed simultaneously as guides for consistency. Picking can be done in ProMax by the selection of velocity values discussed earlier. (see Picks Display )
NMO
Normal moveout or NMO is defined as the variation of reflection arrival time because of source to receiver offset (Sheriff, 1991). The additional time that it takes energy to travel from a source to a flat reflecting bed and back to a receiver ,that is offset from the source point, is shown on a seismogram as a hyperbola. The "moveout" in fact, is the time from the apex of the hyperbola to a point of interest. The normal moveout correction is done to make the seismogram look the same as zero-offset seimograms. The introduction of magnetic tape recording in 1952 and the development of moveable recording heads in 1955 allowed for normal moveout correction. CMP (common mid-point) recording was also made practical by magnetic tape although it was not used extensively until the 1960's (Sheriff, 1995). This will become important when we discuss stacking in a subsequent section. The Gather Tile on ProMAX allows us to display both before and after NMO correction. (see Before NMO and After NMO )
STACKING
Stacking is a composite record made by combining traces from different records (Sheriff, 1991). The development of magnetic tape recording was also important to the stacking process. The full potential of data processing was not realized until the advent of digital recording in the 1960's. The process of stacking is the single most powerful tool for enhancing the quality of seimic reflections (Robinson, 1988). Stacking greatly improves the signal to noise ratio due to the fact that noise is reduced by destructive interference as traces are combined. Wavelets on the other hand interfere constructively to produce a stronger signal. CDP stacking is used for normal moveout correction. (CDP stacks are also known as CMP stacks for common-midpoint) This method is very effective in attenuating several kinds of noise (Robinson, 1988). The tiles from ProMAX display various types of stacks such as constant velocity and CMP stacks. We can see the effect of stacking in the figures that show pre-stackand post-stack sections. Velocity analysis and NMO correction can be viewed as forms of stacking as well. Other types include weighted stacks, diversity stacks, and Simplan stacks. These types go beyond the scope of our discussion but additional information can be found by using the reference list.
CONCLUSIONS
The revolutions of magnetic tape and digital recording have greatly improved our ability to understand the subsurface. Through velocity analysis, NMO, and stacking the reflections of ancient sedimentary layers and structural features are interpreted. These are only a handful of the existing techniques that are used everyday by geophysicists. New techniques such as tomography are being developed in an ongoing process to use digital and computer technology to help us better understand the subsurface.
Futher information can be found from the following sources:
Robinson, E.S., and Coruh, C., 1988. Basic Exploration Geophysics, John Wiley & Sons, New York. p163-210.
Sheriff, R.E., 1991. Encyclopedic Dictionary of Exploration Geophysics, Society for Exploration Geophysicists, Tulsa, OK.
Sheriff, R.E., and Geldart, L.P., 1995. Exploration Seismology, Cambrige University Press, New York, NY. p306 - 322.
ProMAX online Help and Reference Manuals
ACKNOWLEDGEMENTS
Data used in this study was collected from the NW shelf of Australia and generously supplied to us by Dr. Juan Lorenzo. Computer time was supplied by the Louisiana State University Subsurface Laboratory. Computer hardware and software supplied by Landmark.
by Don Rehmer and