geopsy.org

Hello
I am trying to determine Vs profiles using F-K analysis. I used a linear array of 5 sensors spaced at 22m distances to record the MAM data. Then after processing I tried to pick a dispersion curve from the max file using gpviemax. But I cannot identify a clear dispersion curve. Is this due to the linear array? Will the dispersion curve be better if I use triangular configuration. Also what will the dispersion curve look like in the image I attached.

I also don't understand the function of gphistogram. What exactly is it used for and what factors should be considered? Is there any detailed guides on gphistogram and gpviewmax?

Sincerely
Farhan

Dear Farhan,

Effectively, a 2D layout is better to capture ambient vibrations. With a line, you are never sure that the measured velocity is not an apparent velocity. Theoretically, for long recordings you can retrieve the true velocity if the wavefield is uniformly distributed.

With 2D layouts, regular geometrical shapes (triangles, hexagons, square grids,...) are usually not the best choice. The best is usually a random placement that covers the space uniformly. The external shape must be more or less circular. There is no previous knowledge of the ambient noise directions, the array must have the same response in all directions.

gpviewmax was developed after gphistogram. gphistogram does the same job for a single plot at a time: from pairs of values (frequency, velocity) it draws one histogram per frequency, normalized and shown with a colour scale. gphistogram allows the user to plot the density and the cumulative curves for each frequency. It can be used to plot any kind of results, not just dispersion curves. gpviewmax is more oriented towards dispersion curve results. It shows histograms for various values at the same time (velocity, ellipticity, incoherent noise, azimuth, power). Cleaning of curves can be viewed simultaneously from various points of view (velocity, ellipticity,...).

For your particular case, there is visibly no dispersion curve to pick. How long were the recorded signals? Changing the geometry might be the first option to test. You can eventually try SPAC over your linear array. You may recover better results.

Thank you so much for your reply!!

The signals were recorded for 30 mins. I used F-K because I thought that SPAC required a circular configuration to give good results. I have tried processing using SPAC as well, but no clear dispersion curves were visible.

Should I increase the recording time? Also, I only have one seismometer and my study area is quite small. So, I have to use an array of 5 sensors (I will spread them out following your advice) instead of an array of recorders with small sub arrays of sensors. Is this also going to affect the results?