Master Thesis in Seismic Data Processing ?

[liath Fahad]

Hi all I'm Master student in geophysics (seismic reflection method). I'm preparation my thesis trying to use Seismic unix software to process seismic data. What seismic processing topics can be made using seismic unix software or what is the potential uses of seismic unix in seismic data processing .?

[GuyM]

Seismic Unix is a processing tool - so you can do pretty much anything you want to.

I'd suggest the more important main constraints you will have are :

- The amount of storage you have available on disc.
Pre-stack seismic data is BIG, and you want to have at least 5-7 times your input dataset size available on disc before you start, and plenty of detachable drives for backups. You'll need backups. If you are not a Linux/Unix expert (by which I mean 3+ years day-in day-out working in that environment) then the chances are you will delete or overwrite something important at some stage. Even if you are an expert, you might do this.

- The amount of CPU cores you can use.
An MSc has a finite time, and processing on a large scale takes time. A 3D reprocessing project needs tens of Tbytes of disk space and tens of thousands of core hours to complete. If you have a huge cluster for use (100-200 cores) then that's great, but if you are constrained, then this is a consideration

- How much experience you have in processing and Linux.
Practical seismic processing and Linux (at the command line) both have learning curves that are more like a cliff, especially if you are not used to them. You will generate a lot of files (processing flows, meta data like velocities, and seismic data) that you will need to track and manage, and at any time using the wrong file will be a problem you *might not spot.* Experienced people in general have made mistakes, and have adopted habits that minimize those mistakes, but more critically can look at their data and *know what they have done wrong*, - which means they can fix it easily and efficiently.

The data you have available
You can't really do a project comparing acquisition geometries and their effectiveness unless you have multiple seismic surveys/shots in exactly the same area collected in different ways. It's hard to do a project on multiple attenuation techniques on a land seismic line, or on refraction statics if all your data is marine.

The choice of processing tool is (to me) secondary to these, and perhaps the most important question is how much use/support of this package there is available. If you are on your own learning the tool, how to do processing, and Linux at the same time, you will make more mistakes and proceed more slowly, limiting what you can achieve.

So - My MSc was on comparing shallow acquisition sources for P-wave and S-wave generation, as well as the ease/speed at which data could be collected; the packages I used for analysis were much less important than being able to access multiple source types and a suitable test area with a good shallow reflection to "chase."

[liath Fahad]

Thank you Guy M