Diffraction sensitive to velocity

[Tingpingbing]

Why is diffractions sensitive to velocity? And why can we build a improved velocity models from diffraction?

Hope you can help

[GuyM]

Is this question tied to depth conversion at all?

One of the approaches you can use is:

- interpret horizons in two-way-time on a pre-stack time migration
- develop a layer based interval velocity model using the V(rms) velocities from Pre-stack time migration and well control
- depth convert layer by layer (ideally via image ray map migration to give a true-vertical depth model)
- ray trace this model using normal rays to give you back the horizons at their on a zero-offset normal-ray section (stack)
- compare to the seismic stack

This is useful because you only get diffractions at (velocity) discontinuities, and the long tails of the diffractions means that even a small velocity error gives you a large difference between the modelled event and the observed event on the seismic.

So it's really the size/length of the diffraction tails that gives you the sensitivity and a way to cross-check your models.

[Tingpingbing]

Thank you so much

What is the difference between time and depth when it comes to diffraction and sensitivity to velocity?

Do you also know what this sentence means below:
"As a stacking operator for diffractions, Double Square Root operator provides an exact traveltime for a point diffractor
in a homogenous medium"

[GuyM]

What is the difference between time and depth when it comes to diffraction and sensitivity to velocity

?

Well you'd never normally depth convert a stacked (normal ray) section, so it doesn't really come up. Or rather, I'd be asking what your goal is when you do it(!) Mostly depth conversion is for the picked horizons rather than the seismic data when it comes to deeper seismic. The problem is that in depth, the layer depth depends on the interpretation you make - change the interpretation of a key velocity boundary and the depth changes...

"As a stacking operator for diffractions, Double Square Root operator provides an exact traveltime for a point diffractor
in a homogenous medium"

Um, its a nice model but doesn't work for real rocks??

So joking aside (real rocks tend not to be very homogeneous - they have velocity gradients from compaction and structure) its saying that the double square root operator is not too bad of an approximation if the velocities are not varying too much, and can be used to collapse a diffraction back to a point source.

[Tingpingbing]

Thank you for making the sentence easier to understand