Dear Marc,

When I process my data with both passive and active methods (f-k tools), I obtain dispersion curves which are somehow the other way around of what it should be. Lower frequency waves are slower than those at higher frequencies, meanwhile following the theory it should be vice-versa. I'm afraid one of the parameters I use may be incorrect. Could you suggest something? where do you think may be the problem?

thanks, Theodora

## Dispersion curves

### Re: Dispersion curves

Be sure to setup the correct parameters when you load curves from ascii files. Mixing up velocity and slowness is a frequent source of error. Tell me if it solves your problem.

### Re: Dispersion curves

Dear Marc,

I don't load any curves from ascii files. The job is done automatically on the waveforms (what I load). Waveforms are in sac format and when imported in geopsy, they look ok. I only do 'set receivers' for the distances. This is why I don't really understand what goes wrong.

Theodora

I don't load any curves from ascii files. The job is done automatically on the waveforms (what I load). Waveforms are in sac format and when imported in geopsy, they look ok. I only do 'set receivers' for the distances. This is why I don't really understand what goes wrong.

Theodora

### Re: Dispersion curves

Sorry I did not understand your question correctly.

Did you set the source coordinates properly to all signals? See the MASW tutorial for comments about source coordinates.

Did you set the source coordinates properly to all signals? See the MASW tutorial for comments about source coordinates.

### Re: Dispersion curves

Dear Marc,

After checking the December updates of the manual, and comparing to test data, I came to the conclusion that my problem could be related to poor data quality. I tried a pair of stations from active experiments where continuously working vibrator (with different frequencies) was recorded. I have data (2 min duration of 200 sps for a pair of stations). Unfortunately I do not have records of more than a pair of stations at the same time. Distance for each pair varies from 8 to 90 meters. Could you, please, comment on best strategy to process this kind of data.

Other type of data that I have it is continuous records (years) of seismic arrays where the distance between stations is about 2 km and the linear array consists of 10 broadband instruments. Is it possible to use this kind of data in geopsy and what is the best approach?

Thank you very much for your help,

Theodora

After checking the December updates of the manual, and comparing to test data, I came to the conclusion that my problem could be related to poor data quality. I tried a pair of stations from active experiments where continuously working vibrator (with different frequencies) was recorded. I have data (2 min duration of 200 sps for a pair of stations). Unfortunately I do not have records of more than a pair of stations at the same time. Distance for each pair varies from 8 to 90 meters. Could you, please, comment on best strategy to process this kind of data.

Other type of data that I have it is continuous records (years) of seismic arrays where the distance between stations is about 2 km and the linear array consists of 10 broadband instruments. Is it possible to use this kind of data in geopsy and what is the best approach?

Thank you very much for your help,

Theodora

### Re: Dispersion curves

What distance varies from 8 to 90 m? Is it the distance between the station pair or the distance of the pair from the source? I assume that the source is aligned in the source direction. You cannot process simultaneously the signals in the FK tools as you would do with 24 channels, even from a vibrator source. In your case, I would compute the correlation function for each pair (Menu waveform/correlation), then gather all functions in the same signal view. Set pseudo receiver coordinates to each correlation function equal to the station inter-distance of the attached pair. Process the correlation functions in the Active Linear FK toolbox like any other 24-channel shot gather. The source coordinate must be 0, 0, 0. Do not use distance normalization options.

I would suggest a try to Passive Linear FK processing for the one-year recordings. But, inter-correlation techniques are certainly more suitable. Signal correlations are implemented in Geopsy but no automatic process has been developed so far.

I would suggest a try to Passive Linear FK processing for the one-year recordings. But, inter-correlation techniques are certainly more suitable. Signal correlations are implemented in Geopsy but no automatic process has been developed so far.