471 Posts in 231 Topics by 67 members

Noise Attenuation

Forum » Noise Attenuation » Removing linear noise with Tau-p domain

If this is your first visit, you will need to register before you can post. However, you can browse all messages below.

Moderators: Andy Juniper , Guy Maslen , Keleigh Jones

Page: 1 Go to End
Author Topic: Removing linear noise with Tau-p domain 972 Views

Removing linear noise with Tau-p domain Link to this post

Dear Globe Claritas moderators,

I have a question related to removing noises using Tau-p domain. I tried to build a flow to interpolate my data, remove noises in tau-p domain, my job flow is also based on the job file "07_Taup_decon.job" in Globe Claritas tutorial. But I cannot run my job and I do not know what the issue is. I have following questions:
- I have sent you my job file and a presentation describing parameters of my dataset and my job file. Could you please review this and give me comment to edit my job flow (in detail if possible).
- Do I always have to interpolate my seismic dataset when applying taup module ?
- Could you please provide other documents or job flows related to Tau-p domain muting and it combination with deconvolution as mentioned in the Globe Claritas's blog by Dr. Guy Maslen ? In the 2D marine tutorial, it just mentioned the job but did not explain modules in the job flow.
Thank you so much. I am looking forward to hearing from you.

Attached Files

Re: Removing linear noise with Tau-p domain Link to this post

Thanks for attaching your jobflow and a screenshot of the messages seen in the logfile while trying to run your job.
It looks like the data doesn't have the COORD_SCALE header set - this value is used to scale the REC_X, REC_Y, SOURCE_X, SOURCE_Y, CDP_X, CDP_Y and OFFSET headers, if the value is positive then the integer values in the header are to be multiplied by COORD_SCALE to get the real values; if negative then COORD_SCALE is to be used as a divisor. The jobflow is assuming a value of -10 (the GLOBE Claritas default), so if this is correct for the dataset then this message can be ignored, otherwise you will need to correctly set the COORD_SCALE header. If COORD_SCALE is set to -10 then an offset of 63.4m will be stored in the OFFSET trace header as 634.
A question about the offsets in the data -> (far offset - near offset)/receiver interval seems to give 215.312 rather than the number of channels -1?
The RANGE module can be used to output the range of the data's offset (and other) trace headers.
Using OFFREG to double the number of traces in the gathers may require that the far offset is increased from it's initial value - use ((far offset - near offset)/interval)+1 to calculate the number of traces that will be output.
The OFFREG module is used to interpolate the shot records prior to the Tau-P transform to remove spatial aliasing effects (in the tutorial jobflow this is aided by coupling the OFFREG module with a K-fliter). The screenshot of the example shot you attached in your powerpoint doesn't look particularly aliased - so in this case the OFFREG/DEOFFREG modules may not be required in the flow.
The DEOFFREG module can be used to return the data to it's original offset distribution.
The TAUP module transforms the data to the Tau-p domain, while the RHO and TAUPINV modules return the data to the X-T domain.
Currently the jobflow isn't applying a mute (or anything) while the data is in the Tau-p domain. You may find that setting the velocity range in the TAUP module to the more common -1500 to 1500 and applying a mute (via the SMUTE module) while in the Tau-p domain might be a more effective way of attenuating noise.
A good way of trying to track down where an issue might be in a jobflow is to flip off all the modules then add them back in one at a time, running the jobflow after each addition until it fails with the error you were seeing.
If you like, send us an email at claritas.support@gns.cri.nz and we can let you know how you can send us a sample of your dataset so we can try to replicate the issue here and take a look at what might be going on.

Re: Removing linear noise with Tau-p domain Link to this post

Dear Globe Claritas moderators (Ms. Keleigh Jones),
Thank for your recommendations, I could solve my problem. Besides, I have 2 more questions related to muting in TauP domain and parameters in decon. I put it in 2 slides presentation, Could you please see that and give me some comments.

Attached Files

Re: Removing linear noise with Tau-p domain Link to this post

Thanks for the powerpoint.
Usually you would test a range of Tau-P domain mutes, say a wide, mid, and narrow one to start with, to try to obtain the best result. The current mute looks fairly severe and it's possible that the waterbottom may have been removed - you will need to look at shot gathers before and after the Tau-P domain mute to check on the results.
In the *.dwf file used in the DECONW module the TD_on and TD_off columns are the start and end of the design window, and the TA_on and TA_off columns are the start and end of the application time window. The filters are designed within the design time gates, and are applied over the application time gates.

Go to Top