I see that there are a lot of misunderstandings about the topic.

From a physical point of view, it is impossible to retrieve the Rayleigh wave ellipticity curve from an old H/V curve. The H/V curve is measured using ambient seismic vibrations, which consist of all different types of waves, especially Rayleigh and Love waves. As the Love waves are only present on the horizontal components, you will overestimate the horizontal energy on almost the entire frequency range. Therefore, the measured H/V curve will be higher than the Rayleigh wave ellipticity (except for singularities). If you know the contribution of Love and Rayleigh waves, you can correct for this. In general, you don't know it and it is impossible to correct an H/V curve that was measured long time ago without reanalyzing it (e.g. using Article Single station determination of Rayleigh wave ellipticity by...

Dinver can invert the Rayleigh wave ellipticity, but not the H/V curve (what is written in the drop-down menu just means that the ellipticity is defined as the ratio between horizontal and vertical component; some people define it the other way round). Another mistake that is often done: You should never ever invert an ellipticity curve without additional information. The ellipticity (or H/V) curve is a dimensionless ratio, it contains information about the shape of a velocity profile, but not the absolute values. You need additional information (dispersion curves, layer thicknesses, etc.) to constrain the velocity profiles (see e.g. Article Ground structure imaging by inversions of Rayleigh wave elli...

Other methods exist that model and invert the full H/V curve, assuming equipartitioning of the wave field, but still there, you will need additional information to really constrain the velocity profile.

It is indeed possible to get ellipticity curves directly from the H/V curves. This option is available in Denver. As input, you have to enter the resonant frequency along with your desired band of frequency. For details, have a look at the following paper.

Article Shear Wave Velocity Estimates through Combined Use of Passiv...

Dear Dr Biswas,

Thank you for your response.

As you know there is an 'HVTFA' tool defined in geopsy which creates probability density ratios for all frequencies as a *.max file,

which will later converts to ellipticity curve using Max2curve tool.

the output of max2curve which is ellipticity against frequency is needed for inversion in Dinver. I cant find any option in dinver to define H/V directly as the input for inversion.

I looked the paper you mentioned. what I understood was that you employed the same method i mentioned above. my problem is that i dont have the motion component, so I cant employ HVTFA in Geopsy to create *.max file. I need a method to convert the "H/V" curve to "H/V ellipticity" curve.

I would be grateful if you can show me the way.

With dinver from geopsy it is possible, but you should consider the H/V cruvi FROM the main peak on.

You can also take a look on what we did some years ago:

Article H/V spectral ratio analysis and Rayleigh modelization in Eas...

Best regards :)

Dear Hortencia, thanks for your help.

Can you please take a look at the attachement?

is that the part you mean? There is no simple H/V option.

As far as I know the "H/V ellipticity" option needs ".max" file from Geopsy which I don't have, because motion components are needed to create it.

I have no idea which option i can use. I just have a "txt" file containing Frequency against H/V.

Thank you.

here's all I've got here.

I would be grateful if you can show me what to do. thank you so much.

Hi,

Have u found the solution?

Thanks.

Hey Shahram,

I think it's too late to respond to your question, but in case you still need the answer.

If you have the processed H/V data, then why do you need the raw three components recordings? From your H/V curve, you can either put the ellipticity peak along with a misfit weighting factor (0.2 suggested) or just input the whole H/V curve for the joint inversion in Dinver. But, be cautious with the weighting factor assigning to Dinver.

But, if you want to construct the low frequency part of a dispersion curve from the H/V measurements, it is a different story and I can give you some documents to read through that could help.

Best,

I see that there are a lot of misunderstandings about the topic.

From a physical point of view, it is impossible to retrieve the Rayleigh wave ellipticity curve from an old H/V curve. The H/V curve is measured using ambient seismic vibrations, which consist of all different types of waves, especially Rayleigh and Love waves. As the Love waves are only present on the horizontal components, you will overestimate the horizontal energy on almost the entire frequency range. Therefore, the measured H/V curve will be higher than the Rayleigh wave ellipticity (except for singularities). If you know the contribution of Love and Rayleigh waves, you can correct for this. In general, you don't know it and it is impossible to correct an H/V curve that was measured long time ago without reanalyzing it (e.g. using Article Single station determination of Rayleigh wave ellipticity by...

Dinver can invert the Rayleigh wave ellipticity, but not the H/V curve (what is written in the drop-down menu just means that the ellipticity is defined as the ratio between horizontal and vertical component; some people define it the other way round). Another mistake that is often done: You should never ever invert an ellipticity curve without additional information. The ellipticity (or H/V) curve is a dimensionless ratio, it contains information about the shape of a velocity profile, but not the absolute values. You need additional information (dispersion curves, layer thicknesses, etc.) to constrain the velocity profiles (see e.g. Article Ground structure imaging by inversions of Rayleigh wave elli...

Other methods exist that model and invert the full H/V curve, assuming equipartitioning of the wave field, but still there, you will need additional information to really constrain the velocity profile.

Dear Manuel

I have another question related to the topic.

As you mentioned in your papers we can invert ellipticity curves without singularities not only in right flank but also in whole curve ( of course with some additional information like small arrays or ... )

The question is, How much is reliable the results from such curves in the basins with no velocity contrasts?

I know that we should have velocity contrasts to be able to using ellipticity equation.

do you have any new idea or expriment about reliability of using the this kind of curves for inversion?

Dear Saeed,

if the ellipticity curve does not have a singularity, this is also a valuable information to use, as it shows that there is no strong velocity contrast in the structure. In general, such an ellipticity curve still has peaks, which are related with velocity contrasts, but if the peaks (and troughs) are not singular, the velocity contrast is not that strong. In any case, the ellipticity curve carries important information on the structure.

Best regards,

Manuel

Dear Manuel Hobiger ,

I applied RayDec on my case, calculating H/V from single measurement of microtremor. I wrote the script on python with the help of Obspy module, but not sure whether I'm doing it right. Anyway, the result seems promising, even though I have to be careful in narrow-filtering the signal. I work with signals with peaks around 0.2 Hz. Do you have something that I can take a look to confirm what I have done?

Regards,

Rexha

Dear Rexha,

I can send you the matlab code, if you want to compare it. At low frequencies, the width of the filter may be an issue.

Best regards,

Manuel

Dear Manuel Hobiger ,

I am trying to develop a code using obspy for single station H/V of ambient noise. I am interested in low frequency peaks below 0.1 Hz or even lower. What caution should I take about the width of the filter as you mention? I am using bandpass filter with range from 0.01 to 2 Hz. Or which filter is recommended?

Best regards,

Farazi