14. Seismometer Correction/Simulation

14.1. Using a PAZ dictionary

The following script shows how to simulate a 1Hz seismometer from a STS-2 seismometer with the given poles and zeros. Poles, zeros, gain (A0 normalization factor) and sensitivity (overall sensitivity) are specified as keys of a dictionary.

from obspy.core import read
from obspy.signal import cornFreq2Paz

paz_sts2 = {
    'poles': [-0.037004 + 0.037016j, -0.037004 - 0.037016j, -251.33 + 0j,
              - 131.04 - 467.29j, -131.04 + 467.29j],
    'zeros': [0j, 0j],
    'gain': 60077000.0,
    'sensitivity': 2516778400.0}
paz_1hz = cornFreq2Paz(1.0, damp=0.707)  # 1Hz instrument
paz_1hz['sensitivity'] = 1.0

st = read()
# make a copy to keep our original data
st_orig = st.copy()

# Simulate instrument given poles, zeros and gain of
# the original and desired instrument
st.simulate(paz_remove=paz_sts2, paz_simulate=paz_1hz)

# plot original and simulated data
st_orig.plot()
st.plot()

[source code, hires.png, pdf]

../../_images/seismometer_correction_simulation_1.png

For more customized plotting we could also work with matplotlib manually from here:

import numpy as np
import matplotlib.pyplot as plt

tr = st[0]
tr_orig = st_orig[0]

t = np.arange(tr.stats.npts) / tr.stats.sampling_rate

plt.subplot(211)
plt.plot(t, tr_orig.data, 'k')
plt.ylabel('STS-2 [counts]')
plt.subplot(212)
plt.plot(t, tr.data, 'k')
plt.ylabel('1Hz Instrument [m/s]')
plt.xlabel('Time [s]')
plt.show()

[source code, hires.png, pdf]

../../_images/seismometer_correction_simulation_2.png

14.2. Using a RESP file

It is further possible to use evalresp to evaluate the instrument response information from a RESP file.

from obspy.iris import Client
from obspy.core import UTCDateTime
from obspy.core.util import NamedTemporaryFile
import matplotlib.pyplot as plt
import numpy as np

# MW 7.1 Darfield earthquake, New Zealand
t1 = UTCDateTime("2010-09-3T16:30:00.000")
t2 = UTCDateTime("2010-09-3T17:00:00.000")

# Fetch waveform from IRIS web service into a ObsPy stream object
client = Client()
st = client.getWaveform('NZ', 'BFZ', '10', 'HHZ', t1, t2)

# Download and save instrument response file into a temporary file
with NamedTemporaryFile() as tf:
    respf = tf.name
    client.saveResponse(respf, 'NZ', 'BFZ', '10', 'HHZ', t1, t2, format="RESP")

    # make a copy to keep our original data
    st_orig = st.copy()

    # define a filter band to prevent amplifying noise during the deconvolution
    pre_filt = (0.005, 0.006, 30.0, 35.0)

    # this can be the date of your raw data or any date for which the
    # SEED RESP-file is valid
    date = t1

    seedresp = {'filename': respf,  # RESP filename
                # when using Trace/Stream.simulate() the "date" parameter can
                # also be omitted, and the starttime of the trace is then used.
                'date': date,
                # Units to return response in ('DIS', 'VEL' or ACC)
                'units': 'DIS'
                }

    # Remove instrument response using the information from the given RESP file
    st.simulate(paz_remove=None, pre_filt=pre_filt, seedresp=seedresp)

# plot original and simulated data
tr = st[0]
tr_orig = st_orig[0]
time = np.arange(tr.stats.npts) / tr.stats.sampling_rate

plt.subplot(211)
plt.plot(time, tr_orig.data, 'k')
plt.ylabel('STS-2 [counts]')
plt.subplot(212)
plt.plot(time, tr.data, 'k')
plt.ylabel('Displacement [m]')
plt.xlabel('Time [s]')
plt.show()
[source code]

Exception occurred rendering plot.

A Parser object created using a Dataless SEED file can also be used. For each trace the respective RESP response data is extracted internally then. When using Stream/Trace‘s simulate() convenience methods the “date” parameter can be omitted (each trace’s starttime is used internally).

from obspy import read
from obspy.xseed import Parser

st = read("http://examples.obspy.org/BW.BGLD..EH.D.2010.037")
parser = Parser("http://examples.obspy.org/dataless.seed.BW_BGLD")
st.simulate(seedresp={'filename': parser, 'units': "DIS"})

Table Of Contents

This Page