Page 1 of 1
About theoretical HVSR curves
Posted: Wed Dec 04, 2024 3:10 pm
by chenyu.li
Dear all,
I wonder if there are any forward modeling modules that compute the theoretical HVSR curves. I computed the theoretical ellipticity with 'gpell', but it is very different from my HVSR data.
Best,
Chenyu
Re: About theoretical HVSR curves
Posted: Wed Dec 04, 2024 8:22 pm
by admin
Hi,
What kind of theoretical HVSR? Through the assumption of energy equipartition (
Sanchez-Sesma et al. 2011 and many other papers after them, e.g.
Garcia-Jerez et al. 2016)? No, I've never coded it inside geospy.
Re: About theoretical HVSR curves
Posted: Thu Dec 05, 2024 1:04 pm
by chenyu.li
Hi,
Thanks so much for your reply. I have attached an example of joint inversion with ellipticity from HVTSA and the dispersion curve. In my ellipticity curve, the right flank is hard to see. Is it okay to use the whole curve for inversion here?
Best,
Chenyu
Re: About theoretical HVSR curves
Posted: Thu Dec 05, 2024 1:40 pm
by chenyu.li
I apologize for my previous post not having the attachment.
Re: About theoretical HVSR curves
Posted: Thu Dec 05, 2024 2:28 pm
by admin
Hi Chenuy,
Hobiger et al. 2013 also tried successfully to invert both flanks. Your profiles nicely reproduce the observed ellipticity curve even if it is not perfect. Just one detail, the errors look rather large. I'd like to make sure that there was no error while importing this curve into dinver and the conversion to angular ellipticity. Which columns did you select when you imported the curve produced by HVTFA?
The dispersion curve between 5 and 10 Hz, is it from MASW or ambient vibration arrays?
Best regards,
Marc
Re: About theoretical HVSR curves
Posted: Sun Dec 08, 2024 1:32 pm
by chenyu.li
Hi Marc,
When I import the curve to Dinver, I selected 'Ellipticity (H/V)' for the mean values, and 'H/V stddev (approx)' for the error. I also tried to select 'log (H/V) stddev (approx)', and the error looks similar to the one in maxcurve. However, when I run the inversion, the curve imported with 'log (H/V) stddev (approx)' has a larger misfit than the one imported with 'H/V stddev (approx)'. Could you give me some suggestions on which one is correct?
The dispersion curve I used is not from MASW, it is measured with ambient-noise cross-correlation of a 2D seismic survey line very close to my stations.
Thanks so much for your kind suggestions.
Best,
Chenyu