Array analyses of seismic noise at Mt. Vesuvius Volcano, Italy

The properties of the seimic noise wavefield at Mt. Vesuvius Volcano, Italy, are investigated using a dense array of short-period seismometers deployed on the south-west flank of the volcano and operated on 1997–1998. The array had an aperture of 500 m and average station spacing of 40 m. The main goal of this deployment was to investigate the possible presence of correlated signals, which could be attributed to subsurface magmatic and/or hydrothermal activity (volcanic tremor). The spectral properties of the noise wavefield are investigated using data from the arrays and from the summit stations of the permanent seismic network operated by the Vesuvius Observatory. The seismic noise is characterized by three main spectral peaks at frequencies <1, 1–1.5, and 3–4 Hz, which are sustained over the different periods of observations. Measurements of slowness and azimuth indicate that the low-frequency (<1 Hz) signals propagate from the S–SW direction, and are attributed to the marine microseismic noise. At higher frequencies (around 1.2 Hz) the wavefield is characterized by waves whose azimuth of propagation changes with time, overimposed to a sustained signal, which propagates from the east direction. No significant preferential directions of propagation are observed at higher frequencies. The dispersive properties of the wavefield are investigated using the correlation technique first proposed by Aki for the analysis of stochastic signals. The good fit we obtain indicate that the 1–10 Hz noise wavefield is dominated by surface waves. The dispersion characteristics of Rayleigh waves are then inverted for a shallow velocity structure beneath the array, which is in agreement with previous results from both large- and small-scale seismic prospections. Comparison between the dispersion curve for Rayleigh waves and the apparent velocities obtained from frequency-slowness analyses, allows to attribute the analyzed wavefield to the action of shallow sources. The shallow nature of the noise wavefield is also confirmed by polarization analyses, which depict a dominance of horizontal components in the particle motion ellipsoid. This evidence implies that the seismic noise cannot be related to the action of the magmatic reservoir, whose depth has been inferred to be at least 6 km below sea level. The daily changes of spectral power show that the noise signal in the 1–4 Hz frequency band has maximum energy during hours of intense human activity. It is therefore reasonable to attribute the noise origin in this particular frequency band to artificial sources related to the densely urbanized areas surrounding Mt. Vesuvius. Notwithstanding that, the concomitant action of additional natural sources cannot be rejected a priori.