obspy/obspy

ObsPy: A Python Toolbox for seismology/seismological observatories.

Thomas Lecocq

@ThomasLecocq

SAC doesn't rename the components ?

I don't know SAC but did you follow all the commadns to set the orientations of the E and N sensors ? http://geophysics.eas.gatech.edu/classes/SAC/ --> Rotate a pair of data.

79seismo

@79seismo

@vivigb I have a feeling that you're doing something wrong with SAC. As @ThomasLecocq mentioned, your channel codes haven't changed after the rotation. SAC does change these codes. Are you using "rotate to GCP"?

vivigb

@vivigb

@ThomasLecocq @79seismo I think SAC won't rename the components when using the rotate command.
Please see this link
@79seismo yeah i am using " rotate to GCP"

brioglade

@brioglade_gitlab

from obspy.imaging.radpattern import plot_3drpattern
mt = [0.91, -0.89, -0.02, 1.78, -1.55, 0.47]
plot_3drpattern(mt,kind='p_sphere')

ModuleNotFoundError Traceback (most recent call last)

<ipython-input-2-4718903f829c> in <module>
1 import obspy
----> 2 from obspy.imaging.radpattern import plot_3drpattern
3 mt = [0.91, -0.89, -0.02, 1.78, -1.55, 0.47]
4 plot_3drpattern(mt,kind='p_sphere')

ModuleNotFoundError: No module named 'obspy.imaging.radpattern'

hi,how can I fix it?

Elliott Sales de Andrade

@QuLogic

What is radpattern? There's no such thing in ObsPy

brioglade

@brioglade_gitlab

obspy/obspy#1192

Elliott Sales de Andrade

@QuLogic

That is very old, and never released like that, which you can see in that PR

brioglade

@brioglade_gitlab

@QuLogic So what's the right way to write to do seismic source ridiation pattern?I am a new hand.Thank you very much.

brioglade

@brioglade_gitlab

https://docs.obspy.org/master/packages/autogen/obspy.core.event.event.Event.plot.html

What's the CMTSOLUTION file format to write?Do you have an example?Thank you.

Calum Chamberlain

@calum-chamberlain

This first thing that comes up when I google "obspy radiation plot" is this which is what appears to have been the merged change in PR #1192. It looks like you might want to check that page of the obspy docs out and see if that is what you want.

Calum Chamberlain

@calum-chamberlain

For your second question, CMTSOLUTION is a standard format that looks like this.

However, the Event.plot() with defaults to kind=('beachball', 'p_sphere', 's_sphere') calls the obspy.imaging.source.plot_radiation_pattern under the hood, which if you have the moment-tensor and no Event might be simpler to call.

Calum Chamberlain

@calum-chamberlain

I didn't know about this though - they look like really cool plots!

brioglade

@brioglade_gitlab

@calum-chamberlain Thank you very much!

79seismo

@79seismo

@vivigb yes, you are right, KCMPNM has to be changed manually. However, you should read in only the horizontal components for rotate to GCP. The fact that your Z is also changing after the rotation suggests that you're doing something additional to rotating? Rotate to GCP is only a horizontal operation.

Emmanuel D. Castillo Taborda

@ecastillot

Good moorning!
I tried to install obspy with miniconda in google collab. Apparently, it was installed ok. However, when I wanted to use ' obspy.read() ' appears this error: 'TypeError: Unknown format for file /usr/local/lib/python3.6/site-packages/obspy/core/data/BW_RJOB.xml'

can you help me please? Maybe obspy can't be installed with conda in google collab? When I installed with pip, it does his job ok. But for certain reasons I need to install it with conda.

In the next link, you can see the collab notebook that I'm using:
https://colab.research.google.com/drive/1eXNROuqsYeWjHqEyCbIDbZFlgSlPKNMS?usp=sharing

Calum Chamberlain

@calum-chamberlain

Kia Ora, this is the same as obspy/obspy#2547 - the solution then was to use pip and restart the run-time. I don't know if simply restarting the run-time is enough with conda installs, but it is worth a try.

Emmanuel D. Castillo Taborda

@ecastillot

@calum-chamberlain Regards,
I usually use this solution for pip. However, with conda installs are not the same. The error persists.

Thomas Lecocq

@ThomasLecocq

you have to restart the runtime/kernel

the install is likely correct, but obspy doens't find the 'entry points' (=all format readers are defined via entry points) if the kernels are not restarted

Emmanuel D. Castillo Taborda

@ecastillot

@ThomasLecocq have you tried it with conda? Just like I told to @calum-chamberlain , this solution is ok for pip. However, when I restart the kernel after the conda installation, nothing happens.

Thomas Lecocq

@ThomasLecocq

I never used google collab. but it shouldn't be different (and I didn't understand from your answer that it actually worked for you)

and are you working in the right env ?

Emmanuel D. Castillo Taborda

@ecastillot

Yes, I'm agree with you, it shouldn't be different. For this reason I don't understand what is the mistake... I guess that I am working in the right env because
as you can see in the link that I posted, obspy is well imported.

Thomas Lecocq

@ThomasLecocq

So when you restart the kernel, do you actually reinit the conda env? Could you deactivate and activate, evne if it is the base env?

uky-jps

@uky-jps

MLMC.20200928T0000-20200928T0059.spectrogram.png

I am creating 1x3 plots with each component (1,2,Z) hour long spectrograms using Obspy, but I am creating each hour long plot to be compared to other hour long plots. I want to set the cmap scale to be based on a min/max value that I hard code as opposed to the "clip=[0.0-1.0]" scaled option that is currently available, because that "clip" for the cmapping range is based on the min/max values calculated from that hour specifically (as far as I can tell). Is there a way with Trace.spectrogram() that I can set the values that cmap is based on, as opposed to the min/max from the current hour? any help here is appreciated! ( https://docs.obspy.org/packages/autogen/obspy.imaging.spectrogram.spectrogram.html ). As an aside, here is a current example of my plot. above.

my current code that plots one of these individual spectrograms looks like
stm[ii].spectrogram(show=False, axes=ax[ii], clip=[0.35, 0.95], dbscale=True, cmap='jet')

Emmanuel D. Castillo Taborda

@ecastillot

@ThomasLecocq Thanks a lot!, with your comment I could recognize that if I restart the kernel, collab doesn't reinit the conda env.
With this help https://datascience.stackexchange.com/questions/75948/how-to-setup-and-run-conda-on-google-colab . I could find the way to correct it, but only it's ok if I run all my python code in one cell, you can see the way in the next link: https://github.com/ecastillot/notebooks/blob/master/test_obspy_condainstall.ipynb

Due to run all my python notebook code in one cell is uncomfortable, then I tried to type "!conda deactivate" or "!conda activate" for deactivate or activate the env . However, this error appears :
"CommandNotFoundError: Your shell has not been properly configured to use 'conda deactivate'.
To initialize your shell, run

$ conda init <SHELL_NAME>

Currently supported shells are:

bash
fish
tcsh
xonsh
zsh
powershell

See 'conda init --help' for more information and options.

IMPORTANT: You may need to close and restart your shell after running 'conda init'."

Emmanuel D. Castillo Taborda

@ecastillot

For now I am satisfied with the solution that I showed. If someone finds a better way that would be great!

Thomas Lecocq

@ThomasLecocq

Top!!??

Top!!!

Tobias Megies

@megies

generic question related to "best practice": how far away from each other could you have two sensors sharing net+sta , but not loc ids ? I imagine on a field with up to 100m apart, logged on a single unit? or borehole strings with different sensors at different depths ?

I think SEED manual says something about it, pretty sure officially its less than 100m, more like 20m or something, but personally I wouldnt be too picky, most important what fits your workflow

Robert Barsch

@barsch

https://imgs.xkcd.com/comics/26_second_pulse_2x.png

Thomas Lecocq

@ThomasLecocq

Trueeeeeee

Thomas Lecocq

@ThomasLecocq

did some loading time tests in #2748 ; I initially thought that the matplotlib stack was the culprit, but it's interesting to see that it's not !

vivigb

@vivigb

Hi ,everyone
Me again with the doubts related to obspy rotate command.
Why I can't rotate the files with component names such as BH1 & BH2 ? it seems like there is no change after using rotate command on such files!! Also there is no change in component's name after applying rotation!

Thomas Lecocq

@ThomasLecocq

did you apply the changes (1,2,etc) shown in the doc ?https://docs.obspy.org/packages/autogen/obspy.core.stream.Stream.rotate.html

Tobias Megies

@megies

have a look at "components" kwarg here: https://docs.obspy.org/packages/autogen/obspy.core.stream.Stream._rotate_to_zne.html

it will get passed down from your stream.rotate() call

vivigb

@vivigb

@ThomasLecocq @megies No I didn't try those changes shown in the docs.I am little confused in that part .
Should I first rotate to ->ZNE and then use rotate to'NE->RT'?
and I am little confused with the syntax also.
st.rotate('->ZNE',back_azimuth=st[0].stats.sac.baz,["Z", "1", "2"])

Thomas Lecocq

@ThomasLecocq

dunno, I never used 1,2 traces in rotations

but it would make sense, Z12->orientation_from_response->ZNE , then ZNE->baz->ZRT

Tobias Megies

@megies

@vivigb if you want to go from unoriented to RT, indeed you need to do ->ZNE first, which will solely rely on the sensor orientation specified in your Inventory. Your Traces should then be rotated and relabled to ZNE. Then do a second rotate call with NE->RT, specifying the back azimuth.

vivigb

@vivigb

@megies Thank you very much... <3

vivigb

@vivigb

Is there any option to mass download stationXML files alone? (since I already have the IRIS data and need statioXML files to rotate BH1 BH2 components to BHR and BHT components)

Where communities thrive

People

Repo info

Activity