Rock-physics templates: reading rocks in elastic space

Why Ip vs Vp/Vs?

The crossplot axes are chosen to maximize rock-class separation:

• Ip (acoustic impedance): ρ · Vp. The fundamental property of a rock’s response to a compressional (P) wave. The full seismic amplitude at normal incidence is controlled by the CONTRAST in Ip across a boundary. Low Ip = fast/dense rock; high Ip = slow/light rock.
• Vp/Vs ratio: the ratio of P-wave to S-wave velocity. Sensitive to FLUID: gas dramatically lowers Vp but barely changes Vs, so Vp/Vs drops. The ratio is also sensitive to lithology: sandstones typically have Vp/Vs around 1.7–1.9, shales around 1.9–2.1, carbonates 1.8–2.0.

Together, Ip and Vp/Vs capture most of what matters for QI: density, P-velocity behaviour, and the fluid-sensitive ratio. Other RPT axes exist (λρ vs μρ, Ip vs Is, φ vs Kφ) but Ip vs Vp/Vs is the workhorse.

Exercise — move the sand

• The widget opens with a SHALE cluster (green) fixed at Ip ≈ 7500, Vp/Vs ≈ 2.0, and a SAND cluster (blue) at porosity 22% with brine pore fluid. Notice how close the two clusters are — they nearly overlap. This is the BRINE SAND scenario: sand with water is hard to distinguish from shale.
• Drag the Porosity slider to 30%. Watch the sand cluster move to the LEFT (lower Ip) and DOWN (lower Vp/Vs). Higher porosity = more pore space = lower density and velocity = lower Ip. At 30% the brine sand is still somewhat close to shale but now distinguishable.
• Switch the Fluid to oil. The sand cluster shifts further down-left — oil lowers both Ip (less dense than water) and Vp/Vs (compressible pore fluid softens the frame). The oil-sand cluster is now clearly separated from shale at 30% porosity.
• Switch the Fluid to gas. Dramatic shift — the sand cluster drops much further in both dimensions. This is the classic GAS ANOMALY of AVO interpretation: gas produces a huge impedance and Vp/Vs change that seismic can see clearly.
• Drag porosity back to 10%. Even with gas, the sand cluster is now tight against shale again. Low-porosity sands don’t give the fluid-sensitivity leverage that QI needs. This is why tight sands are generally NOT QI-favourable plays.
• Study the GHOST POINTS (faint circles of the other two fluid colors): they show where the sand WOULD sit for the other two fluid options at the current porosity. The arrows connecting current to ghost are the GASSMANN FLUID-SUBSTITUTION TRAJECTORIES — what §5.3 taught you in elastic-space form.

The rock classes you need to know

• Shale: high Ip (7000–9000), high Vp/Vs (1.9–2.1). Typically anchored at the top-right of the crossplot. The dominant background facies in siliciclastic basins; your main job is to distinguish other rocks from it.
• Brine-saturated sandstone: moderate Ip (6500–8500), moderate Vp/Vs (1.80–1.90). Often OVERLAPS WITH SHALE, especially at low porosity. QI struggle case.
• Oil-saturated sandstone: lower Ip than brine (by ∼300–1000), slightly lower Vp/Vs (by ∼0.05–0.10). Moves the sand down-left from the brine baseline.
• Gas-saturated sandstone: much lower Ip (by ∼1000–2500) and much lower Vp/Vs (by ∼0.2–0.4). Produces a distinctive low-Vp/Vs cluster — the classic gas "bright spot" region.
• Tight / cemented sandstone: very high Ip (10000+), moderate Vp/Vs (1.8–1.95). Upper-right of the crossplot. Often confused with carbonates.
• Carbonate: variable. Clean limestones can plot around Ip 11000–14000, Vp/Vs 1.85–1.95. Dolomites extend to higher Ip. Porous carbonate reservoirs move toward the sand clusters.
• Coal: very low Ip (3000–5000) and low Vp/Vs (1.6–1.8). Dramatic outlier at lower-left; usually easy to identify.
• Salt: high Ip (10000–12000), low Vp/Vs (∼1.85, because halite has Vp ≈ 4500 m/s but also substantial Vs ≈ 2450 m/s). Cluster near upper-left of the crossplot.

The QI-favourable vs unfavourable map

An RPT is a MAP OF QI FEASIBILITY. By plotting the target facies on the template before any inversion is run, you can predict whether the project is feasible:

• QI favourable: target facies plots in a region WHERE THE CLUSTERS SEPARATE. Example: high-porosity gas sand at φ > 20% — plots clearly below and to the left of shale. Seismic amplitudes will be diagnostic.
• QI marginal: clusters OVERLAP but the target lies near the edge of the shale cluster. You may be able to resolve it with CAREFUL inversion + good wavelet + multiple attributes, but confidence is lower.
• QI unfavourable: target plots INSIDE the shale cluster. No inversion can separate rocks that have identical elastic properties. The project should not be run with QI — use conventional interpretation + direct drilling instead.

Making this determination before committing to a QI project is perhaps the most valuable use of an RPT. A clean RPT showing clear separation is the GREEN LIGHT for a serious QI investment; an RPT showing overlap is the warning flag to spend the budget elsewhere.

Building an RPT from well data

RPTs are built from WELL LOG DATA — the only place where you know both the rock properties and the elastic properties simultaneously. The workflow:

• Gather Vp, Vs, and ρ logs from every available well. If Vs is missing (common), predict it with an empirical relationship (Castagna mudrock for shale, Greenberg-Castagna for sand).
• Compute Ip = ρ · Vp and Vp/Vs at each depth sample.
• ASSIGN each sample to a rock class based on GR (shale vs sand) and resistivity (hydrocarbon vs water).
• PLOT each sample on the RPT, color-coded by rock class.
• Compute cluster centroids and confidence ellipses (typically 2σ) for each class.
• RUN GASSMANN fluid substitution (§5.3) on each sample to generate the brine / oil / gas variants. Plot ghost clusters for the alternative fluids.
• Validate: does the RPT make geological sense? Does the gas-sand cluster lie where it should, given the reservoir conditions?

For a new basin or a new target facies, the RPT takes weeks to calibrate properly. For an existing basin with established rock-physics standards, a few days. Time spent on the RPT is the single best investment in a QI project.

Pitfalls

• Wells in the wrong facies. If your calibration wells are in distal shale but your target is a proximal channel sand, your RPT doesn’t represent the target. Mitigate: cross-check against analog basins, invest in offset-well data if needed.
• Anisotropy. Real rocks are often transversely isotropic (TI) — Vp along bedding differs from Vp across bedding. Using isotropic Vp/Vs overestimates overlap at the edge of clusters. Mitigate: use VTI rock physics where data warrants.
• Wrong fluid model. Gassmann assumes quasi-static fluid response. High-frequency seismic can deviate (Biot-Gassmann, dispersion). For most conventional QI applications Gassmann is adequate; for heavy oil or very porous rocks, consider dispersion corrections.
• Confusing elastic overlap with impossibility. Overlap on Ip vs Vp/Vs doesn’t mean there’s NO elastic distinction — it means the 2D projection doesn’t show it. Try λρ vs μρ or other axes; some rocks separate in 3D elastic space even when they overlap in 2D.
• Ignoring mineralogy. Clean quartz sand and clean carbonate sand have very different elastic properties. An RPT calibrated on one can’t be used for the other. Always calibrate per-basin, per-lithology.
• Over-interpreting a single well. A cluster from one well may be too narrow (under-sampling) or may reflect local conditions that don’t generalize. Use at least 3-5 wells per rock class if at all possible.