geopsy.org

Hi, Marc,

Thanks for dinver snapshot. The ground profile of Ellipticity inversion has only Vp profile, so can I have other way to see the Vs profile?

What is difference between a sub-layer and a new layer, I supposed there is velocity jump to a new layer, but sub-layer is velocity smoothing, is that right?

Thanks again

Be careful that ellipticity inversion is correct only if your ellipticity curve is effectively the Rayleigh ellipticity. DO NOT TAKE the H/V ratio as the ellipticity, you may found any kind of results, quite different that the real profile. This option in Dinver has been added in parallel to recent developments we made to extract the ellipticity (still testing and validation).

What is safer is to invert for the peak frequency only. But in this case there is a huge trade-off between depth and velocity. If you fix the velocity profile or the depth then you will retrieve the other value. Better to use it in combination with a dispersion curve (even at high frequency from SASW).

If you cannot see the Vs profile, check that the number of profiles to plot (entered in the dialog box that appear just after selecting the menu item) is greater than 2. In fact, this number of profiles may vary from 1 to 5 (Vp, Vs, Rho, Vp2, Vs2). The last two are for refraction, a future implementation. You'll be able to combine surface waves and first arrival information. Rho is generally fixed for whole column, hence the classical value is 2 (Vp and Vs).

Sub-layers are used only for non-uniform layers (linear or power law). For the parameterization, you have only one layer for which you can set the velocity at the top and at the bottom. Behind, to compute dispersion by Thomson-Haskell techniques only uniform layers are accepted (implementing the propagation through heterogeneous layers would have been much more harder). Hence, the functional variation is simply discretized by a certain number of uniform sub-layers. The default value is 5, with modern CPUs 10 or more doesn't hurt and it gives smoother and nicer results. More than 10 may slow down the inversion.

I See. Thanks a lot.