GNGTS 2019 - Atti del 38° Convegno Nazionale

GNGTS 2019 S essione 1.1 103 the velocity around the source position is assumed to be constant, thus the relative time delay between nearly identical waveforms allows for an estimation of the slightly different location in space of the two events. Calling the relative hypocenter coordinates (Δx, Δy, Δz, ΔT 0 ), positive to the E, N and up, of an event related to the master event and φ and θ the azimuth and take- off angles at the k-th seismic station, respectively, the time delays Δt N are used to compute the inversion for the hypocentral coordinates of the current event with respect to the master event as follows: (1) Equation (1) may be written in matrix form as Gm = d , which is a standard mathematical problem whose solution is given by m = [G T G] -1 G T d . We estimated the relative time delays from the cross-correlation of filtered waveforms. We adopted different bandpass filtering and window length of analysis to check the stability of the results. For each cluster the distribution of source positions obtained from the relative location was fitted by a plane. The orientation of the best fit plane was compared with that of both fault and auxiliary planes obtained from the focal mechanism of the master event, estimated through the software FOCMEC (Snoke et al. , 1984; Snoke, 1989). The high number of seismic stations available in the area allowed for the estimation of focal mechanisms of earthquakes characterized by very low magnitude, down to M L 1.2. Taking into account the focal mechanism, we computed the synthetic seismograms at all stations that recorded the events ( Computer programs in seismology , Herrmann R.B., 2013), and compared the polarity of P wave with that of the recorded signals in order to be sure about the reliability of the estimated source kinematics. We finally applied the relative location technique to the fault patches obtained from the single cluster re-locations to enhance the imaging of the seismogenic volume. Since the master events do not have similar waveforms among them to estimate the time delay through the cross-correlation, we a slightly different procedure. Seismograms were bandpass filtered in a high frequency band (10 Hz - 20 Hz), then they were manually cut retaining only the first 3 pulses, and those starting with a negative impulse were multiplied by -1. The cross-correlation of these signals gave an estimate of the time differences to be used for the relative location applying the same method previously used for any clusters. We also performed the absolute location of all the events of the clusters using a self-made code, for comparison with the relative locations (Fig. 2). Fig. 2 - The 432 earthquakes of the 27 clusters. Different colors show the catalog location (blue), absolute location performed in this work (green) and relative location of all the events of the clusters (red).