Seismosoft | Earthquake Engineering Solutions

Could we spectral matching for near fault? Is it ok?

Hi beiraghii@yahoo.com. It is possible to use spectral matching for near fault ground motion records but you do have to be aware of the pulse-type character often found in near-field records. You could use spectral matching and then examine the before and after records to see if the pulse has been preserved in the matching process.

There are utilities available (see Stanford University, Baker Research Group at http://web.stanford.edu/~bakerjw/resear ... ation.html, for one example, and PulseSnapper at https://nees.org/resources/pulsesnapper, for another).

Also note that spectrum-matched records are useful in estimating mean response, but not in estimating the dispersion of response.

As a final note, I believe the proper place for this topic would have been in the Analytical Capabilities subsection rather than the Referencing and License Issues Section.

Best of luck beiraghii@yahoo.com.

Dear huffte,

As usual, your responses are very thorough and helpful, for all of us - thank you.

And yes, you are correct, this topic was initially posted in the wrong section, but fortunately the Seismosoft Support team has now moved it to the right place.

Thanks again,

Rui

Rui, this brings to mind a possible enhancement to SeismoMatch (and SeismoSignal, and SeismoSpect for that matter). I wonder if it might be possible to add an algorithm which would

(a) identify the presence or absence of pulse-type characteristics in ground motion records

(b) if a pulse is identified, then report the pulse period, Tp.

For SeismoMatch, this would enable users to check whether or not the matching process has preserved the pulse characteristics or not.

The identification of pulses can be a little tricky, I suppose, so it may not be possible. It might be worth exploring though.

Thanks very much, Rui. All the best.

Dear Tim,

Thank you for your very interesting suggestion. We will definitely explore the possibility of introducing (or developing ourselves?) such an algorithm in future releases of the seismological Apps.

Best Regards,
Seismosoft Support

Thanks you for your help.

How would the fictitious response spectrum be defined, rich in the low frequencies (high periods), believed to be representative of a near-field event?

Navid, this would probably require special studies.
You should probably better discuss this with your supervisor.
Seismosoft Support