Support with H/V calculation and data import

[Faisal]

Hello Marc,

I am a PhD student at CAU in Kiel and have done passive seismic measurements over a period of 14 days. For a specific day, I calculated the H/V ratio using MATLAB. Unfortunately, I was unable to import my H/V data into Geopsy or Dinver for inversion. is that impossible?

My second question: i would like to calculate the H/V ratio using Geopsy. On my specific day, I recorded 22,501 values for each component (Vertical, North, East) on an hourly basis, resulting in a matrix of dimensions 22,501 x 24 for each component. How can I import these signals into Geopsy?

Thank you very much for your help!

Best regards,

[admin]

Hi,
Under which format did you try to import the H/V curve into Dinver? Can you provide an example? Note that Dinver invert H/V curves assuming that they are Rayleigh ellipticity curves. If the H/V curve is computed in the classical way, it includes body waves and Love waves as well which leads to modifications of the amplitude. A better way to get an ellipticity curve is through HVTFA (in geopsy) or Raydec (Hobiger et al. 2009).

It is possible to import .mat files into Geopsy like any other signal file, have you ever tried?
Best regards,
Marc

[Faisal]

Hi Marc,

Thank you for the response.

_ I was able to import my signal. However, I would like to point out that my signal only contains positive amplitudes after processing. Does this cause any problem with the H/V calculation? In fact, I started the calculation, but I did not receive any results.

_ In my Signal both the variables AZ, AN, and AE have the dimensions (22501, 888).
22501 rows: This corresponds to 22501 different frequency bands over which the amplitudes have been calculated. This frequency range spans from 0 to 50 Hz and is divided into fine frequency intervals. Sampling Frequency = 200Hz.
888 columns: The analysis was conducted over a period of 888 time points. Each of these time points corresponds to a time window of 3600 seconds (1 hour).

_ I think a matrix cannot be imported into GEOPSY, therfore I intend to average each signal, resulting in a matrix for each component with dimensions (22501, 1).
In fact, I started the calculation, but I did not receive any results.

Best regards,
Faisal

[admin]

_ I was able to import my signal. However, I would like to point out that my signal only contains positive amplitudes after processing. Does this cause any problem with the H/V calculation? In fact, I started the calculation, but I did not receive any results.

As far as there is a signal, even it is shifted positively or negatively, HV computation can be achieved. After splitting the signal into time windows, the average value of the signal is subtracted to remove the baseline offset.

_ In my Signal both the variables AZ, AN, and AE have the dimensions (22501, 888).
22501 rows: This corresponds to 22501 different frequency bands over which the amplitudes have been calculated. This frequency range spans from 0 to 50 Hz and is divided into fine frequency intervals. Sampling Frequency = 200Hz.
888 columns: The analysis was conducted over a period of 888 time points. Each of these time points corresponds to a time window of 3600 seconds (1 hour).

It is unclear if your signals are spectra or time series. It is possible to provide spectra but not through Matlab files.

_ I think a matrix cannot be imported into GEOPSY, therfore I intend to average each signal, resulting in a matrix for each component with dimensions (22501, 1).
In fact, I started the calculation, but I did not receive any results.

That's possible. I can't remember if there are heuristics implemented for matrices (e.g. the longest dimension is the time axis). This is done this way for HDF5 files. It could be modified in the same way for Matlab files.