The Well Tie: From Log to Seismic
The workflow every subsurface team runs and few can defend end to end: velocity from the sonic, a wavelet with its phase pinned down, a synthetic built from the logs, and the honest slide against the seismic. Cross-book on purpose, because the tie lives where petrophysics, modeling, and interpretation meet.
You can turn a sonic and density log into impedance and reflectivity, choose a velocity transform and say why, pick a wavelet and defend its phase, build the synthetic seismogram yourself, name the polarity convention of an unfamiliar dataset, and tie the well knowing which wiggles to trust and which ones tuning has faked.
The logs that matter
Travel time is the one thing the well and the seismic both measure; the sonic is where the tie is born, borehole sins and all.
Wyllie's time-average is the working bridge between porosity and velocity, and the compaction correction is what keeps it honest in young, soft rock.
Wyllie versus Raymer-Hunt-Gardner is a call you make on every well, and secondary porosity is the vuggy exception the sonic quietly ignores.
Half the wells you will tie are missing the shear log or the density log, and impedance needs both; Castagna's mudrock line and Gardner's rule are the honest stand-ins, with error bars you should say out loud.
Velocity times density is impedance, and an impedance contrast is all the seismic ever sees; this is the exact seam where a well log becomes a reflection.
The wavelet
You rarely know the wavelet in your data, so you must know the candidates cold: the Ricker as workhorse, the zoo for everything the processor may have left in.
A phase error moves every picked event by the same amount, a systematic depth error the amplitude spectrum cannot show you; ties die here more than anywhere else.
Sampling sets the frequency ceiling and bandwidth sets the blur; the resolution you promise at the workstation is decided by these two.
Convolution to synthetic
Impedance to reflectivity to trace is a two-step machine; a tie you can defend requires knowing which step made which wiggle.
This is the deliverable: logs in, synthetic out, ready to slide against the seismic. Everything before was preparation for these four tracks.
Real data is noisy and convolution is an approximation; the practitioner's judgment is knowing when the simple model is enough and when it is quietly lying.
Thin beds brighten before they vanish; when the synthetic and the seismic disagree at a pinch-out, tuning is the first suspect to interrogate.
The tie itself
The interpreter's restatement of the physics: contrast in impedance is polarity and brightness on the section, read at the workstation without a calculator.
The well lives in depth, the seismic in time, and the tie is the bridge between them; resolution decides how much weight that bridge can carry.
SEG positive or European, zero-phase or minimum: misread the convention and every amplitude call inverts. The tie is how you catch it on day one.
Convolve reflectivity with a wavelet and watch the section appear: the tie is defensible when you can say what every wiggle is and what none of them can be.