Page 1 of 1

Flat and floating datum corrections, why and when?

Posted: Thu Nov 19, 2020 6:31 am
by jefry123
Hi every body, I hope you all are safe and healthy in this Covid19 situation!

I may have a problem in understanding the application of flat and floating datum corrections in seismic data processing.

I know that, we use the flat datum correction because the hyperbolic assumption is true if sources and receivers be on a flat surface. So, in case of topography, we compensate this elevation differences at sources and receivers positions by moving them to a flat datum. On the other hand, we use the floating datum, a smooth version of topography, for processing, velocity analysis and migration.

My question is that why we choose the floating datum for processing instead of the flat datum, while the hyperbolic assumption is valid for the flat datum?

For the interpretation step, the final stack sections to be used, are on flat datum, right? If so, why is that? Don't they need the true topography? The floating datum isn't a better approximation of the topography?

Re: Flat and floating datum corrections, why and when?

Posted: Thu Nov 19, 2020 9:21 pm
by GuyM
They key thing here is we are trying to use vertical static shifts on traces to "fix" the elevation corrections, when the ray-paths we are dealing with are not vertical. So statics are a bit of an approximation.

If the static shifts are small, this is not too bad. It certainly gets us closer to the solution, and we can then use residual statics to mop up the rest.
However if the shifts are large (lets say > 8-12ms or so TWT) then the approximation is not good. So we address this by using a floating datum (a smoothed version of the elevation) so that the static shifts are small.

You'll also find issues in hard rock areas; that's because the velocity variations can be big (when the hard rock is weathered or fractured, then not) so it can give rise to much bigger static shifts.

So - it's all down to vertical static shifts being an approximation, which gets worse the larger the shift.

Re: Flat and floating datum corrections, why and when?

Posted: Fri Nov 20, 2020 12:13 pm
by jefry123
Thank you Guy, I learn a lot from your answers.

So, the smoother the topography, the more closer the two corrections using flat and float datums, right?

What should be the datum for the final output of the processing (stacks and gathers)?

Re: Flat and floating datum corrections, why and when?

Posted: Sat Nov 21, 2020 2:57 am
by GuyM
So, the smoother the topography, the more closer the two corrections using flat and float datums, right?
Um, not quite - it's not just how rough the topography is but the how much it varies from the highest to the lowest point that counts. That's what gives the size of the static shift.

A linear ramp from sea level to 500m altitude is very smooth, but if you have use a constant, flat datum you will have big static shifts. if you slammed the flat datum through at 250m you'd have big distortions to the reflection hyperbolae at either end of the line.
What should be the datum for the final output of the processing (stacks and gathers)?
You always correct to a final, flat datum before delivery and final displays. It can be useful to consider both for QC's while processing of course.
Thank you Guy, I learn a lot from your answers.
Thanks!

I'm out of geophysics professionally now (perhaps forever!) and while I have a great job I'm really enjoying it's nice to have a chance to help those still carrying the torch.