ms-stretch

A Stretching Method (plot) plugin for MSNoise

MSNoise possesses the possibility of computing relative velocity variation curves, but a way to plot the results in the 'standard MSNoise way' was missing.

This plugin contains a command to plot the stretching data in the same format as for the MCWS method. Moreover, some other scripts are included to plot the stretching data along with forcings like precipitation, temperature, pressure, etc. The dvv curves can also be plotted with their respective correlation coefficients to quickly check for low correlation parts of the curve.

Examples can be found here.

For questions, comments and bugs please contact [email protected].

Installation

• Follow the MSNoise installation instructions (here)
• Install ms-stretch. One way to do it is to copy it into the folder where msnoise and other python packages are saved, type cd ms-stretch and then run python setup.py install.

Running ms-stretch

After installing the plugin make sure to add "ms_stretch" to the plugin parameter in the web admin. Alternatively msnoise config set plugins=ms_stretch works too. Note in both cases that an underscore is needed, not a dash.

To just plot the stretching data in the typical MSNoise fashion with all the defaults simply type: msnoise p stretch plot dvv. This command can also plot (multiple) pairs and I tried to keep it as close to the original msnoise plot dvv command as possible.

Plotting multiple filters

While working with the stretching method one often tries out different lag time windows (LTW) for the same filters. To better compare this results they can be plotted together by simply calling the filter argument multiple times.

• msnoise p stretch plot ndvv -f 1 -f 2 plots the dvv curves corresponding to the two filters with otherwise default values (ndvv stands for new dvv).

• msnoise p stretch plot ndvv -f 1_2_4 -f 1_4_8 plots two dvv curves corresponding to the same filter, BUT for different LTW. In this case, 2s-4s and 4s-8s.

Setting up LTW data

MSNoise outputs by default the stretching data in a folder called STR, where the results for the different filters are saved. If you already have computed some STR data, but still want to have plots with various LTW then you can change the folder names of the filters from e.g. STR/01 to STR/01_2_4 or similar. Naturally, you have to remember in which LTW you computed the data.

Another option is to run the command msnoise p stretch compute stretching. A script is executed that is identical to the original MSNoise one with the only difference that the output folder is changed to the format mentioned above, i.e. STR/filterid_startlag_endlag. This also has the (nice) side effect that if data for the same filter with another LTW is computed, it does not overwrite the previous data.

Plot dvv curves with forcings

This plugin also supports the possibility of plotting forcings like precipitation, temperature and pressure alongside the dvv curves. To get started one must at first "install" another table called DefaultStations.

The command msnoise p stretch plot install takes care of that. Now this table should appear and be editable in the MSNoise web admin (run msnoise admin).

As a default a precipitation entry is given. Further forcings can be added at will. The columns of the table should be self-explaining*. The following commands extract information from that table so it is advised to regularly check the settings.

• msnoise p stretch plot forcing -f 1 plots a dvv curve with the default of plotting the first forcing in the DefaultStations table. Multiple filters or different LTW are possible.

• msnoise p stretch plot mforcing -f 1 is the same as the command before only that multiple forcings are possible. Check out the examples to get an idea.

*Note: In the column Default Station 'all' can be passed and this takes the average of all the stations in the forcing folder. Otherwise, write the name of the station without the '.csv'-ending. Additionally, click on examples above to find out more about the expected file and folder structure for the forcing commands to work properly.

Other commands

• msnoise p stretch plot corr -f 1 plots the dvv curve with defaults and adds a subplot with the corresponding correlation coefficients.

• msnoise p stretch plot uninstall deletes the DefaultStations table from the database.

Miscellaneous

If questions arise it can be helpful to use the --help argument after the command to see which parameters are needed or optional. Furthermore, the Python Code has generally a docstring for each function so that sometimes extra information can be found there. Otherwise don't hesitate to contact me under the E-mail address given before.

Outlook

Some further modifications are planned for this plugin including:

• an alternative stacking method to improve noisy dvv curves in some cases

• corresponding plots

• computing and plotting Signal-to-Noise ratios

Citing MSNoise

In the spirit of the MSNoise documentation: If you use MSNoise, even a small part of it, for your research and publications, please consider citing it:

Lecocq, T., C. Caudron, et F. Brenguier (2014), MSNoise, a Python Package for Monitoring Seismic Velocity Changes Using Ambient Seismic Noise, Seismological Research Letters, 85(3), 715‑726, doi:10.1785/0220130073.

Furthermore, if you find this plugin useful, consider mentioning or even citing it:

J.F. Kühn (2020), ms-stretch: A MSNoise plugin, GitHub repository, https://github.com/janfer95/ms-stretch