Remove response and stationXML for accelerometers

Dear all,

I have some questions regarding removing responses at Slovenian seismic station network.

Some stations have seismometers installed together with accelerometers, for example CRES station.
Here is the link to waveforms from seismometer and accelerometer and also corresponding stationXML file.

https://www.dropbox.com/s/km20duk5ty9zkzl/CRES_HHZ?dl=0
https://www.dropbox.com/s/f7xvhpxze816ya1/CRES_HNZ?dl=0

https://www.dropbox.com/s/p9hgh8t591sd89f/CRES.xml?dl=0

We want to simulate displacement from both sensors and both should look similar.
Here is the code, where I implement everything

Hi Miha,

I try to answer your two questions.

(1) The function of remove_response for calculating displacement from accelerometers has no problem, if you give proper parameters. The problem you have is the improper water level. If you revise water level of HNZ from 60 to 120, you will see much similar waveform between HHZ and HNZ. The reason is that accelerometer compensates more energy than velocity seismometer in lower frequency band. The detail explanation can be found in http://docs.obspy.org/tutorial/code_snippets/seismometer_correction_simulation.html

(2) You could try the following command to get digitizer delay in XML (net=inv[0] ; sta=net[0] ; cha=sta[0] ; cha.response.response_stages )

The calculation of group delay for FIR can refer to https://www.mathworks.com/matlabcentral/newsreader/view_thread/58747 and https://www.mathworks.com/help/signal/ug/compensate-for-the-delay-introduced-by-an-fir-filter.html

Hope that helps!

Best,

Youngman

Hi!

  1. Yes you are quite right the function works well with proper parameters. The water level 60 was enough for seismometers but it was not enough for accelerometers.

  2. After looking deeper into theory we managed to calculate the delay, which is already corrected by our Quanterra digitizers, so the time difference between
    original and corrected waveform is apparently due to the instrument response itself.

Thanks for the effort