Forward modelling

[Leandro]

Hi Marc,

This may be a brush up of theoretical and practical understanding.

If I have 2 clear peaks in the HVSR spectra and I want to forward model using gpsh or gpell, do I account for both peaks as 2 layers above half space in the text file? Or do I use 2 different text files with one layer above half space each to model the 2 peaks?

Thank you very much for your reply as always.

Cheers,
Leandro

[admin]

The origin of the two peaks might be diverse. Effectively, it can be two resonating layers with sufficiently different characteristic frequencies. It can have a body wave origin and hence gpsh simulations are fine. Ellipticity of Rayleigh waves can be also at work, and higher modes may play a role. The best would be to calculate the H/V with codes like Raydec (unfortunately still not in geopsy, only a Matlab code is available) or HVTFA (in geopsy). To invert, you need at least a way to constrain the velocity (a piece of dispersion curve, even a small one from MASW, Vs from literature,...). See Hobiger et al. (2013). The inversion of Sh curve is not implemented in Dinver. In my opinion, but more investigation might prove the contrary, matching the amplitude of the Sh curve may lead to bias in Vs and Vp profiles. The attenuation plays an important role and it is not a well known parameter.