Quantitative Interpretation: Rocks from Seismic
From amplitude to pore fluid, with the error bars attached. Rock physics, AVO, inversion, and the calibrated probabilities that let a seismic volume speak about reservoir quality out loud.
You can carry a seismic amplitude through rock physics, AVO, and inversion to a facies probability, defend what the processing did to amplitudes upstream, and state the uncertainty like someone who will still be in the room when the well comes in.
The elastic rock
Acoustic impedance was the interpreter's whole vocabulary; the four moduli and Vp/Vs are the sentences QI actually speaks.
Gassmann is the one equation that lets you ask what the seismic would look like if the fluid were different, which is the whole question; the Rock Physics course supplies the idea in full, with its assumptions and its failure modes attached.
The workflow only becomes yours when it runs on real logs: recover the dry frame from the brine sand, substitute gas, and watch impedance drop and the top-sand reflection flip polarity.
AVO
Amplitude versus offset is the only pre-stack rock signal you get; Aki-Richards is how you read it without lying to yourself.
The intercept-gradient crossplot turns AVO into a map you can screen a basin with; the classes are its compass directions.
A synthetic seismogram is your handshake between wells and seismic; no loop closed, no QI story believed.
Processing that preserves amplitudes
Every gain, decon, and migration upstream either preserved relative amplitudes or quietly destroyed your rock signal; know which happened before you interpret it.
Q compensation, offset conditioning, and gather hygiene decide whether simultaneous inversion sees rock or residual moveout.
The QI workflow
Attributes stop being pictures and start being evidence the moment they are tied to rock properties; this is that moment.
The workflow is a chain of custody from seismic to rock properties, and templates are how you read elastic space without getting lost; the Rock Physics course builds the template itself, mesh line by mesh line.
The template earns its keep on points it never modeled: drop inversion points onto the mesh and read porosity and fluid together, splitting pairs that share one impedance.
Inversion turns reflectivity into impedance, and reading its products honestly is a skill the colour bar will not do for you.
A facies probability is a promise about frequencies; calibration is whether you keep it, and reviewers can check.
Facies probabilities at every voxel close the QI loop into a reservoir model, and uncertainty is part of the deliverable, not an apology.
Advanced QI
Azimuthal attributes read fractures and stress, and FWI velocity models are becoming QI inputs; both extend the toolkit past isotropic comfort.
Neural networks classify facies at scale; the QI discipline you built is what keeps their confidence honest.
Case files
A Girassol-style fluid prediction end to end: the case where calibrated AVO either pays for the survey or teaches humility.
Subsalt QI inherits every imaging compromise made above it; Thunder Horse is the masterclass in knowing what survived.
Ten percent of gas makes almost the whole anomaly: fizz water mimics pay because Wood's average collapses at the first whisper of gas. Amplitude proves gas is present, not abundant, which is the sentence that saves dry holes.