vera
if you are using verion 10, filter is not used. but in the later versions
it is. when you start automag, the first output is e.g.
Ml constants: 1.000000 0.910000 0.000870 -1.670000
Only 4 pole filter for wa, changed
Only 4 pole filter for wa, changed
WA filter constants: 2.00 4 20.00 4
Give input S-file
so you can see which constant are used. i suppose you do not get this ?
if you do, i have to check the code more.
jens
On Fri, 26 Feb 2016, Vera Schlindwein wrote:
Hi Jens,
I am working with Florian on this problem.
I meanwhile checked the following: automatic amplitude picking in mulplt
works fine.
We are using an extra filter (ML Low cut 5 hz, high cut 20 Hz as defined in
MULPLT.DEF) because most of our events are high frequency and low magnitude.
My feeling is from inspecting the automag source code that this filter is not
read. Automag determined amplitudes have typically periods of 1-2s whereas
running mulplt we obtain periods of typically 0.15 s. Could you have a look,
if that is potentially a problem? That would be great. If needed I'll send
you some example files.
Regards
Vera
florian
i have seen it a couple of times,so not enough to do something about it.
might be the automatic amplitude routine
that fails. we are testing an alternative routine. if you only ise a 5 s
window, it should not be possible to pick amplitude later, sounds like a
bug. when using the automtic amplitude pick in mulplt, it is the same
routine that is used so you could check how it happens there. in any case,
i would like to look at a specific example that does not work. for that i
need the s-file, the station file, MULPLT.DEF, the waveform file and the
response file for the station that does not work.
jens
On Fri, 26 Feb 2016, Florian Schmid wrote:
> Dear Seisan community,
>
> we tried to employ the AUTOMAG routine (Seisan version 10) for
> amplitude picking in a large dataset of local events.
> However, about half of the automatic amplitude picks occur in
> spurious places, away from the
> maximum amplitude of the Wood-Anderson filtered seismogram in
> MULPLT.
>
> We do not use the spectral model (typing a '0' after AMI in EEV)
> and tried to adjust the wa window in AUTOMAG, in order to tie the
> automatic amplitude
> pick to the P- and S-phase picks. However, that does not show an
> effect on the automatic pick location,
> which still might occur 15s after the phase picks, even if a
> wa window of 5s was set.
>
> Has anyone encounter such problems?
> Did You find a solution to it?
>
> Thanks a lot in advance!
>
> Regards
> Florian
>
> _______________________________________________
> seisan mailing list
>[email protected]
>http://mailman.uib.no/listinfo/seisan
>
--
Alfred-Wegener-Institut, Helmholtz-Zentrum für Polar- und Meeresforschung
*PD Dr. Vera Schlindwein*
*Senior Scientist*
Section Geophysics
Am Alten Hafen 26
27568 Bremerhaven
Phone: +49 471 4831 1943
http://www.awi.de
Helmholtz-Zentrum
_______________________________________________
seisan mailing list
[email protected]
http://mailman.uib.no/listinfo/seisan