Estimating Tp for records with pulse period larger than 4

[ebrahimi.sahar]

Dear Seismosignal Users,

I would like to inquire if Seismosignal is capable of accurately estimating the pulse period (Tp) for records with a pulse period greater than 4 seconds.
I attempted to input several pulse-type records with Tp>4, as mentioned in Baker's 2007 paper, into Seismosignal. Upon extracting the pulse in each record, I observed that the Tp was consistently calculated as 3.98 seconds. I speculated that this might be due to the default maximum period of 4 seconds being used for creating spectra. Consequently, I adjusted the maximum period to 8 seconds and reattempted the analysis, but unfortunately, there was no change in the results.

Could anyone help me on this issue?

[seismosoft]

There is no intentional limitations in the period Tp and this is not related whatsoever with the maximum spectral period.
Possibly the result that you get is a coincidence. You could also refer to the corresponding publication:
Kardoutsou V., Taflampas I., Psycharis I.N. [2017] " A New Pulse Indicator for the Classification of Ground Motions", Bulletin of the Seismological Society of America, Vol. 107, No. 3, doi: 10.1785/0120160301.

Seismosoft Support

[maddie.heidari]

I have the same exact problem. For example record RSN8164 of PEER database has Tp=10.05. but using seismosignal I am getting 3.9!

[hung0937]

I’ve encountered a similar issue. Is this a known constraint of SeismoSignal?”

[seismosoft]

No, there are no constraints in the pulse identification algorithm of SeismoSignal.
In some cases, there are inconsistencies between the NGA database parameters and the SeismoSignal calculations with records that are NOT pulse-like. In these cases, since the algorithm employed is different (Kardoutsou et al. vs Shahi & Baker) this can lead to discrepancies. However, I re-iterate that these discrepancies have been found in Non pulse-like records, i.e. in cases when there is not a clear identifiable pulse to be found.
Seismosoft Support