GG450 April 22, 2008

1. GG450 April 22, 2008 Seismic Processing

2. Reflection Processing We've been talking about acquisition of reflection data - spreads, frequency and resolution considerations, etc. Now we talk about how these data are processed to obtain reasonable profiles of the sub-surface "geology". Purpose: To enhance the reflection record so that noise is minimized and desired reflections are enhanced, and corrected to provide the clearest possible representation of the structures below.

3. STATIC CORRECTIONS: • "Static" implies that the same identical correction is made to a trace at all times. For example, topographic corrections and weathering corrections add or subtract a fixed time from a trace depending on the elevation,. thickness, and velocity of the upper low velocity layer. WHY? So that reflectors that are actually flat will appear flat in the final profile. • Think about this: If there is a large hill in the middle of your profile, the change in elevation will add travel time to all the traces shot on the hill. This extra time will affect the “flatness” of all layers below, and –if we ignore it- our reflection stacking process won’t work. We need to correct the traces so that all shots and receivers appear to be at a constant elevation with a constant velocity directly below.

4. In the section above, the “reflectors” are actually flat, but, because of changes in the elevation, they appear to be folded. The elevation static correction fixes this. A similar correction is applied to correct for changes in the thickness of any low-velocity surface (weathered) layer, since the extra travel time generated by changes in this thickness will distort the shape of reflections below: Thus, static corrections correct the effects of variable surface layers so that the geometry of deep reflectors is correct.

5. In the case above, the elevation correction has been applied on the left, but changes in the thickness of the surface low-velocity layer distort the reflections below. • Static corrections are applied by specifying a "datum", or level where each trace will be referred to, and determining the amount of time to add or subtract to each trace depending on the thickness and elevation of the low velocity layer. • Alternatively, one could apply a weathering correction by just subtracting the time differences to the water table – assuming that the water table is flat.

6. This figure shows static correction applied to real data.

7. Frequency filtering: Some noise is at different frequencies than the signal. It can be removed by applying a filter to the data to remove noise at both higher and lower frequencies. This process is often done by looking at the SPECTRUM of the data and determining which part is signal and which is noise. Alternatively, one can look at a FILTER PANEL, where the signals have been passed through various narrow-band filters. The final filtering is determined based on the bands where the reflections stand out most clearly. (SEE FIGURE BELOW)

8. This is a filter panel showing how the reflections change as the frequency changes. The panel to the right shows the sum over all frequencies.

9. DYNAMIC CORRECTIONS: These corrections stretch or squeeze the time scale of each trace. Again with the intent of making the resulting profile as close to geologic reality as possible. AGC: Automatic gain control increases the size of small reflections and decreases large ones until all reflections are roughly the same size. VELOCITY filtering: Since surface waves and some other arrivals are generated by the source, they usually have about the same frequency as the reflections, and they are thus difficult to remove using frequency filtering. The apparent velocity of reflections, however is usually much greater than the apparent velocity of these other arrivals, and there are methods to remove arrivals that have an apparent velocity much different than that of the reflections.

10. NMO will “flatten” the reflections (one example of a reflection is the red line), and velocity filtering will at least partly remove surface waves (blue regions). This gather extends on both sides of the shot . This gather has had AGC applied, otherwise the arrivals at later times would be very small.

11. NMO (Normal Move-Out) Correction: Each gather is processed by subtracting the appropriate move-out corrections and then STACKING all the traces in the gather to produce ONE trace, which would be equivalent to the trace obtained at x=0, directly above the common depth point. If the NMO corrections have been applied correctly, primary reflections should be enhanced, and all other arrivals (diffractions, 2nd reflections, refractions, surface waves) should be decreased in size.

15. MIGRATION: migration attempts to correct “errors” caused by reflecting points that are not directly below, as is assumed by the above processing. Migration "collapses" diffraction hyperbolas, moving reflected energy not only in time but from trace to trace. • A simple geometric method of migration is possible. If a profile of a seismic reflector is available as shown, then circles drawn from each shot point with a radius equal to the time to the reflector must be tangent to the location of the true reflecting point. By drawing a series of such circles from many shot points, an image of the true reflector can be formed:

16. The black curve is the true reflector, but the seismic signals make it appear to be along the blue and orange curves.

17. Seismic reflections (green curves) make structures (black curves) appear wider than they actually are. This can be very important in estimating the volume of a reservoir.

18. Migration also helps to reduce the effects of DIFFRACTIONS that are caused by arrivals from corners and sharp edges:

19. HOMEWORK Due Thursday, April 24. The green lines above show seismic reflection times. Graphically migrate these reflections and delineate the actual locations of the reflections. Assume that the plot is scaled to yield circular wave fronts and that the sources and receivers are on the top line.

20. DEPTH SECTIONS: The profiles we have been working with up to now have been TIME SECTIONS, in that the Y - axis is travel time since the shot. Thus, it is possible, in fact LIKELY, that there will be considerable distortion in our profiles from a real geologic section since the velocity changes with depth. Consider the figures below:

21. MORE HOMEWORK FOR Thursday: The reflection time section to the left is observed. Plot the equivalent depth section in the grid below the time section.

22. Both of these contain the same information, but by plotting vs. depth, the relative thickness’ of the layers becomes apparent. Low velocity layers appear much thicker and more pronounces on time sections than on depth sections. • Another common distortion is VERTICAL EXAGGERATION. Seismic sections are often relatively flat, and they are plotted with a considerable vertical exaggeration to emphasize structure. Consider the two sections below:

23. The exaggeration badly distorts angles of faulting and slopes, but without it, the structures would be very hard to see.

24. REVIEW OF VELOCITIES: • We've talked about many types of velocities, and some clarification is needed: • PARTICLE: du/dt, where u is the displacement of a particle in the ground as the wave passes by. • PROPAGATION: The speed at which the wave moves through the ground, also called phase velocity • GROUP: The speed at which the energy in the wave travels. When waves are dispersive (velocity depending on frequency) the phase and group velocity are different (this is the case for swell in the ocean and seismic surface waves) • APPARENT: The speed at a wave travels across the ground (=1/P) • AVERAGE: In reflection: the velocity used when ignoring refraction (Green's method) • RMS: In reflection: the velocity obtained from the slope of the reflection in an x2t2 plot. Used to obtain interval velocity. • STACKING: in reflection: the velocity that will provide the best normal moveout correction • INTERVAL: in reflection: the velocity of a particular layer between two reflectors