Network covariance matrix spectral width: natural case with synchronization

This example shows how to synchronize traces in the stream before to calculate the spectral width of the network covariance matrix. It makes use of 6 hours of real seismic data acquired the 2010.10.14 (between 10:00 and 16:00 - GMT) by the Piton de la Fournaise seismic network. This dataset contains seismovolcanic signals linked to a pre-eruptive seismic swarm and to the beginning of a co-eruptive tremor starting at 15:20 (GMT). This basic example does not apply any pre-processing.

import covseisnet as csn
import matplotlib.pyplot as plt
from obspy import UTCDateTime
from obspy.clients.fdsn import Client

# download data from the YA Undervolc seismic network with RESIF Seismic data portal
client = Client("RESIF")
signal_duration_sec = 6 * 3600
t = UTCDateTime("2010-10-14T10:00:00.00")
list_stations = [
    "SNE",
    "FOR",
    "UV01",
    "UV02",
    "UV03",
    "UV04",
    "UV05",
    "UV06",
    "UV07",
    "UV08",
    "UV09",
    "UV10",
    "UV11",
    "UV12",
    "UV13",
    "UV14",
    "UV15",
]
stream = csn.arraystream.ArrayStream()
for sta in list_stations:
    st = client.get_waveforms("YA", sta, "00", "HHZ", t, t + signal_duration_sec)
    stream.append(st[0])

# downsample data to 25 Hz
stream.decimate(4)

# synchronize data
stream = stream.synchronize(start=t, duration_sec=signal_duration_sec, method="linear")

# calculate covariance from stream
window_duration_sec = 20
average = 15
times, frequencies, covariances = csn.covariancematrix.calculate(
    stream, window_duration_sec, average
)

# calculate spectral width
spectral_width = covariances.coherence(kind="spectral_width")

# show network covariance matrix spectral width
fig, ax = plt.subplots(1, constrained_layout=True)
img = ax.pcolormesh(
    10 + times / 3600, frequencies, spectral_width.T, rasterized=True, cmap="viridis_r"
)
ax.set_ylim([0, stream[0].stats.sampling_rate / 2])
ax.set_xlabel("2010.10.14 (hours)")
ax.set_ylabel("Frequency (Hz)")
plt.colorbar(img).set_label("Covariance matrix spectral width")