What can you really see in seismic?

Module 00  ·  Start here

How the earth changes as you go down

Start here if you are new to seismic. Questions about what can and cannot be resolved depend on the properties of the earth the wave travels through. Build a column below, from the sea surface down to five kilometers, and three things change together as depth increases.

Three things happen as you go down

Rocks compact. Pile sediment on top of sediment and the grains pack tighter, water is squeezed out, and the rock gets both faster and denser. Velocity climbs roughly in a straight line with depth, and density follows it. That is the background trend, and most of the earth obeys it.

Time is not depth. Seismic arrives in two-way time. Because velocity keeps rising, a second of time near the surface covers far less rock than a second of time at four kilometers. Every thickness you read off a section in milliseconds has to be converted with a velocity that is itself changing.

The earth absorbs energy. It takes the high frequencies first, so the pulse that started sharp arrives at depth long and smooth. This is attenuation, measured by a number called Q, and it is why deep data usually looks softer than shallow data.

What that does to resolution

All three trends act on the same quantity. How much detail seismic can show depends on its wavelength, and wavelength is velocity divided by frequency: λ = V / f. Going down, the velocity on top of that fraction gets bigger and the frequency underneath it gets smaller. Both push the wavelength up, and a longer wavelength means less detail.

The thinnest bed you can resolve therefore grows steadily with depth, often from ten meters near the surface to fifty or more at four kilometers. The cause is the earth itself rather than the processing or the interpretation.

Drag the depth marker down and watch the last panel. It gives the thinnest bed that could be resolved at any depth, for the data you have. The λ/4 it uses is the standard rule of thumb; module 02 measures the number directly for a specific wavelet and gets something closer to λ/5. The trend with depth is the same either way.

Start from

The compaction trend

Density follows the velocity through Gardner's relation, and shear velocity through the mudrock line. Both are rules of thumb, and both are close enough to reason with.

An overpressured zone

Attenuation

A high Q means the rock absorbs little. Hard rock is 200 and up; soft, gassy, shallow sediment can be under 50.

The column you just built

Click any panel to move the depth marker

1 The earth shaded by impedance

2 Velocity Vp and Vs

3 Density

4 Impedance Vp × density

What that means for your data

Depth becomes time, the pulse loses its high frequencies, and the thinnest bed you can resolve grows.

5 Depth to time

6 Frequency left after the earth absorbs

7 The pulse itself at the surface and at your marker, both scaled to the same height

8 Thinnest bed you can resolve λ/4, the rule of thumb

Depth at the marker— m
Two-way time there— s
Vp / Vs— / —
Vp/Vs ratio—
Density— g/cc
Impedance—
Peak frequency left— Hz
Wavelength there— m
Thinnest bed you could resolve— m
Compared with the shallow section—

Try this — and what to watch for

  1. Drag the depth marker from top to bottom and watch the last stat cell — the thinnest bed you could resolve. The point: it is a single number you can look up for any target depth. The thinnest resolvable bed grows several times over between the shallow section and the deep one, with no change in the acquisition. A target that gives a clean pair of reflections at 1.5 s can be a single merged event at 4 s.
  2. Find the overpressured zone in panel 2. It is the kick to the left in the velocity curve, where the rock stops getting faster. The point: overpressure is a drilling-safety problem, and seismic velocity is one of the few ways to see it ahead of the bit. Compaction reverses because the pore fluid carries some of the load the grain framework would otherwise carry, so the rock stays less compacted than its depth suggests. Velocity drops and density drops with it. On seismic the zone appears as a velocity inversion.
  3. Watch what it does to impedance in panel 4. There is a strong negative reflection at the top of the zone and a positive one at its base, because impedance falls and then recovers. That pair is often the clearest thing on the section, and it is a fluid-pressure boundary rather than a rock-type boundary.
  4. Turn the water depth down to zero. The whole time axis shifts up by the two-way time through the water, and the shear velocity curve now starts at the surface instead of at the seabed. Water carries no shear wave at all, which is why marine streamer data records only P waves.
  5. Take Q from 300 down to 50 and watch panels 6, 7 and 8 together. The point: bandwidth lost on the way down cannot be fully recovered later. The pulse broadens and the resolvable bed thickness grows. Gassy or very soft shallow sediment absorbs strongly, and everything imaged beneath it is degraded, because the loss happens during propagation rather than in the recording.
  6. Push the velocity gradient from 0.3 up to 1.0. A fast-compacting section reaches high velocity sooner, so it runs out of resolution sooner. Slow, undercompacted sections keep their resolution deeper. Two basins at the same depth can be very different targets.

If you remember nothing else

  • Velocity and density climb together with burial, and impedance climbs faster than either, because it is their product.
  • A millisecond is not a fixed thickness. It is worth more rock the deeper you go, because velocity keeps rising.
  • Overpressure reverses the trend, and shows up as a velocity and impedance inversion — one of the few places seismic tells you about fluids rather than rock.
  • Q describes how much energy the rock absorbs. High frequencies go first, so bandwidth narrows steadily on the way down.
  • Deep resolution is worse for two reasons at once, higher velocity and lower frequency, and both act on the same wavelength.
Method, assumptions, and where to read more

Velocity below the seabed follows a linear compaction trend, V = V₀ + k·z. Density comes from Gardner's relation, ρ = 0.31·V0.25 with V in m/s, which is the metric form of the original ft/s expression. Shear velocity uses Castagna's mudrock line, Vs = (Vp − 1360)/1.16. Water is 1500 m/s, 1.03 g/cc, and carries no shear wave. All three are rules of thumb that hold for clastic sections and fail for carbonates, evaporites and anything unusual — they are here to give the right shape, not the right number for your field.

The overpressured zone multiplies the velocity trend by a smooth raised-cosine trough. Density falls with it through Gardner, which gives a drop of a few percent for a substantial velocity reversal, and that is about what is observed.

Two-way time is integrated down the column as Σ 2·dz/V. Attenuation is constant-Q: the amplitude spectrum is multiplied by exp(−π f t*) where t* = Σ dt/Q along the path, with water treated as lossless. For a Ricker source the surviving peak frequency then has a closed form, which is checked here against a numerical search of the attenuated spectrum and agrees to better than a tenth of a hertz.

What this does not include, and it matters: processing. Inverse-Q filtering and spectral whitening recover part of the lost bandwidth, so real processed data at depth usually looks better than the raw physics here suggests. What is shown is what the earth does before anybody tries to correct it. Also left out are velocity dispersion, which accompanies attenuation, and the anisotropy that makes real velocities depend on direction.

Next up

01 · Where a seismic trace comes from

You now have an earth with velocities and densities in it. The next module takes a handful of those layers and follows them all the way to a seismic trace, one step at a time, with nothing hidden: rock to impedance, impedance to reflections, reflections to a wiggle.

Open module 01 → All modules
← All modules Next: where a trace comes from →