Will the 4D Signal Show?

A Signal Is Only Useful If You Can See It

Every scenario in this part predicts a change in the rock, and prediction is not the same as detection. A time-lapse survey is two surveys subtracted, and the subtraction never lands perfectly. Sources and receivers are not in exactly the same place, the near-surface has changed, the noise is not identical. What is left after the best processing is a residual, the 4D repeatability noise, and any real reservoir change must rise above it to be read. The feasibility question is blunt and it comes before the money is spent: is the impedance change the reservoir will produce larger than the noise the survey will leave behind?

A Working Threshold

Put a number on the noise as a rule of thumb. Good modern time-lapse acquisition and processing leave a repeatability residual of a few percent in impedance, so a working detectability line of about 4 percent is a reasonable place to draw it: an impedance change comfortably above 4 percent is likely visible, one below it is likely lost in the residual. This is a rule of thumb, not a law, and a particular survey sets its own line from its own repeatability metrics, but 4 percent is a fair working value for judging feasibility on paper. Now take the three impedance changes from the pressure-versus-fluid section and lay them against it. Gas breakout changed the P-impedance by 26.9 percent, far above the line, loud and unmistakable. Depletion changed it by 3.2 percent and injection by 3.7 percent, both sitting at or below the 4 percent line. The fluid signal shouts; the pressure signals whisper.

Designing for the Answer

That ranking is the honest close of the pressure story. A gas or water front sweeping through a reservoir is one of the most detectable things in geophysics, which is why 4D earns its keep watching fluid movement. A pure pressure change, depletion or pressure support with no fluid front, produces a smaller impedance shift that can fall below the repeatability floor, so pressure-only 4D is often marginal and sometimes not worth shooting. Three levers can lift a marginal case: choose a rock that sits where the velocity-pressure curve is steep so the pressure change buys more velocity, spend on repeatability to lower the noise floor, or carry the shear information so a small change still separates pressure from fluid even when its impedance is faint. All of this is a design calculation done before the survey, not a lament after it. And it closes Part 8. Pressure and time have changed the rock in place; the last thing left to change is the rock itself, its layering, its shales, and the frequency of the wave that probes it. Part 9 takes up Backus averaging, velocity dispersion, and attenuation, the scale-and-frequency effects that decide what a seismic wave even sees.

References

• Lumley, D. E. (2001). Time-lapse seismic reservoir monitoring. Geophysics, 66(1), 50-53.
• Landro, M. (2001). Discrimination between pressure and fluid saturation changes from time-lapse seismic data. Geophysics, 66(3), 836-844.