/qopen_ridgecrest

Qopen scripts for community stress drop study

Primary LanguagePython

Qopen Ridgecrest

This repository contains results and the source code to reproduce results for the Community Stress Drop Validation Study using the Qopen method.

The Qopen method fits observed seismic envelopes (energy density) of the shear wave onset and coda with a synthetic Green’s function accounting for scattering, intrinsic attenuation and geometrical spreading plus source and site terms. Intrinsic attenuation and scattering strength are estimated from the full S wave envelope and are relatively well constrained. The fitting is done in different frequency bands independently. Therefore scattering, intrinsic attenuation, spectral source energy and site amplification are obtained as a function of frequency.

References: https://github.com/trichter/qopen#references

Robust source spectra using Qopen, BSSA, doi:10.1785/0120210200 [pdf]

Results

Attenuation

Site amplification

Geometric mean of site amplification of stations CI.MPM,CI.WBM,CI.WMF is fixed to 1

sites

Source spectra of 8 events

sds8

Source spectra of 55+1 events

See pdf

Earthquake parameters

eqparams

Tables

  • Mw, fc, stressdrop in files data/eqparams*.txt.
  • _cv -> used velocity from 2d model instead of 3200 m/s to calculate M0 from W and to calculate stress drop
  • _freen -> optimize for M0, fc and n instead of optimization for M0, fc with fixed n=2.58

Load source spectra or other results

import json

with open ('qopen/04_source_nconst/results.json') as f:
    results = json.load(f)

freq = results['freq']
evid = '38445975'
sds = results['events'][evid]['sds']
print(freq, sds)

Energy site amplifications are located in the file qopen/02_sites/results.json, key 'R'.

Run code

Preparation

  1. Prepare the code environment by installing the relevant python packages: qopen>=4.3 obspy>=1.3.1 obspycsv>=0.4, e.g.:

    conda --add channels conda-forge
conda create -n eqspec obspy statsmodels
conda activate eqspec
pip install qopen obspycsv

  2. Download the data from https://scedc.caltech.edu/data/stressdrop-ridgecrest.html.

  • Unpack waveforms to events folder

  • Copy StationXML to stations folder:

    cp data/events/38472279/*.xml data/stations/
cp data/events/38472279/*.txt data/stations/

  • Optionally, recreate the event CSZ files

    • Download the phase files RC_phase_*.tar, unpack these into data/phase
    • Downlaod the catalog files to data
    • Change to scripts folder and run python prepare_events.py to prepare a catalog in CSZ format readable by ObsPy

Run Qopen

  • Change to qopen folder
  • Check out configuration in conf.json. I chose mean S wave velocity of 3200 m/s estimated from velocity model
  • Run run_qopen.sh script to recreate results
  • I have not yet done it, but the scripts can be used to invert for spectra of all ~13,000 earthquakes. Just repeat the last two steps (qopen source and qopen recalc_source) for the full event catalog

Calculate stress drops and plot figures

Change to scripts folder and run the remaining scripts.