Dear you!
I am calculating wavelength of motion of a earthquake. The steps are as follow:
Wavelength = Vs/fmax = Vs x T = Vs x 2@t
where:
Vs: shear velocity
T: period of wave
fmax: maximum frequency
@t: recording interval
I try a example for Kobe motion
is this way right?
Thank in advance!
Hi Nguyen,
A new, small privatissimo for you. One must first distinguish between primary and secondary waves. In addition, between types of waves: space waves, surface waves, Love waves, etc.The propagation medium and its properties is very important. For a description of an earthquake can one relate always only to the Seismogram.
Arrangement see below
Good luck
Michael Lersow
Dear Nguyen,
I guess that by @t you mean the sampling interval. In this case, you cannot use it for the computation because you would overestimate the frequency of the wave and so underestimate the wavelenght. Indeed the inverse of 2*sampling interval gives you the Nyquist frequency (i.e. the higher un-aliased frequency) which is for sure higher than the maximum frequency of the signal.
In addition, be aware that even using the maxim frequency of the signal (I suppose S wave as you use Vs) could be too high. I suggest to use the dominant frequency (period) which could be where the amplitude spectrum of the proper wave has its maximum. Or you can compute the dominant period looking at the number of time cycles in your seismomgram.
Regards,
Lorenzo
Dear Lorenzo,
the algorithm which is described by Nguyen, leads on a wrong track and should not be pursued. The approach has as basis a harmonic wave in which also the phase shift missing. I can not recommend the path to pursue continue. General procedure see above.
Propagation velocities and wavelengths in different rocks for comparison can you find in literature: For (P) elastical space waves: (5,000 - 7,000) m/s - Granite (Gneiss) 5,000 m/s; (S): (3,000–4,000) m/s - Granite (Gneiss) 3,000 m/s.
From a existing Seismogram it is possible with a statistical program easy to calculate the average period time or the average frequency. The Nyquist sampling theorem determines the frequency the correct with a sampling rate can be described.
Regards and good luck
Michael Lersow
Dear Michael Lersow!
Thank you for your helps. You have many good comment for me.
Dear Lorenzo Pastore Barzaghi,
I think you are right. I am also using predominant period to calculate wavelength. It is much larger compare to my previous method.
Thank you so much for your help!
- Badges
- Science topic