Problem of Vidale method of Polarization analysis

[xichaoqiang]

• Please check whether the bug was already reported or fixed.
• Please provide the following information:
  • ObsPy version, Python version and Platform (Windows, OSX, Linux ...) Obspy 1.1.1, Python3.6, Windows 64bit
  • How did you install ObsPy and Python (pip, anaconda, from source ...) Anaconda
  • If possible please supply a short, self contained, correct example that demonstrates the issue.
• Please take the time to [format your ticket appropriately] I want compare the reuslt of three method offered in polarization analysis. But I met a problem in Vidale. The output size is strang. The length of my fcnt file is 18 min. Sample frequency is 500hz.The fcnt file is in the attachment. The segment in polarization_analysis is 20 s. (https://guides.github.com/features/mastering-markdown/) (e.g. put error tracebacks into code blocks)

  
  st = obspy.read(r"1.fcnt", merge=True)

a=polarization_analysis(st,20.0,1.0,10,12,st.traces[0].meta.starttime,st.traces[0].meta.endtime,False) b=polarization_analysis(st,20.0,1.0,10,12,st.traces[0].meta.starttime,st.traces[0].meta.endtime,False,"flinn") c=polarization_analysis(st,20.0,1.0,10,12,st.traces[0].meta.starttime,st.traces[0].meta.endtime,False,"vidale") print(len(a['azimuth'])) print(len(b['azimuth'])) print(len(c['azimuth'])) Result is here.

55
55
569774

1.fcnt.zip

[ThomasLecocq]

Hi,

I just tested with your file and one thing could be the issue:

In [2]: st = read("1.fcnt/1.fcnt", format="rg16")

In [3]: st
Out[3]:
114 Trace(s) in Stream:

2001.5121.1.DP3 | 2018-05-29T09:42:46.550000Z - 2018-05-29T09:43:16.548000Z | 500.0 Hz, 15000 samples
...
(112 other traces)
...
2001.5121.1.DP2 | 2018-05-29T10:01:16.550000Z - 2018-05-29T10:01:46.548000Z | 500.0 Hz, 15000 samples

[Use "print(Stream.__str__(extended=True))" to print all Traces]

In [4]: st.merge()
Out[4]:
3 Trace(s) in Stream:
2001.5121.1.DP2 | 2018-05-29T09:42:46.550000Z - 2018-05-29T10:01:46.548000Z | 500.0 Hz, 570000 samples
2001.5121.1.DP3 | 2018-05-29T09:42:46.550000Z - 2018-05-29T10:01:46.548000Z | 500.0 Hz, 570000 samples
2001.5121.1.DP4 | 2018-05-29T09:42:46.550000Z - 2018-05-29T10:01:46.548000Z | 500.0 Hz, 570000 samples

Your data file doesn't contain Z, N or E components, I don't understand how the method could actually work (except if you forgot to copy something you do to the code above).

[ThomasLecocq]

but ok, even with the demo file (3000 samples @100 Hz), it gives different lengths:

st = read()
a=polarization_analysis(st,5.0,1.0,8,10,st.traces[0].meta.starttime,st.traces[0].meta.endtime,False)
b=polarization_analysis(st,5.0,1.0,8,10,st.traces[0].meta.starttime,st.traces[0].meta.endtime,False,"flinn")
c=polarization_analysis(st,5.0,1.0,8,10,st.traces[0].meta.starttime,st.traces[0].meta.endtime,False,"vidale")
print(len(a['azimuth']))
print(len(b['azimuth']))
print(len(c['azimuth']))

Out:

4
4
2944

not an expert, but @jwassermann can maybe explain why/if it's normal that the vidale method returns the azimuth for each sample (except some start/end buffer I guess)?

in the code:

if method.lower() == "vidale":
        res = vidale_adapt(stream, var_noise, fs, frqlow, frqhigh, spoint,
                           stime, etime)

doesn't use the wlen and win_frac indeed. If this is normal, it should be reflected in the Documentation.

[obspy-bot]

This issue has been mentioned on ObsPy Forum. There might be relevant details there:

https://discourse.obspy.org/t/polarization-examples/1388/8