GNGTS 2016 - Atti del 35° Convegno Nazionale

232 GNGTS 2016 S essione 1.2 To better understand the kind of movement, we calculated for any single station the horizontal component in this frequency band along directions from N to N165°E, spaced 15° to find its maximum amplitude (Rossi et al. , 2013).The displacement is initially upward, except for GSR1, with a slight tilting parallel to the locally prominent tectonic trend (Fig. 1b). Later, the opposite behavior is observed. The amplitude of the horizontal signal varies from a minimum of 0.25 mm to maximum of 1.0 mm. The vertical displacement amplitude ranges from a minimum of 1.9 mm to a maximum of 4.3 mm. Source location and tomographic inversions. The methods used, a novelty in this context, include earthquake location techniques and tomographic inversion of the arrival times. A first estimate of the transient’s source location and origin time was done by applying the classical circle intersection method, based on the P-arrival (vertical component) times only, using an average propagation velocity. We refined the velocity model as well as the source location and origin time estimate in a tomographic approach, by minimizing the time residuals (Rossi et al. , 2016). The 3D model extends about 210 km in longitude and 125 km in latitude, from NE-Italy to W-Slovenia, and from the topographic surface to the Moho surface. It includes three surfaces at the average depths of 2, 4 and 11 km, respectively, (Scarascia and Cassinis, 1997; Brückl et al. , 2007). The tomographic method adopted enables to limit the uncertainty of the results even for few data as in our case (Böhm et al. , 1999; Vesnaver and Böhm, 2000). The transient source was finally located close both in the space (about 6 km to the NW, at 9 km depth) as in time (three months before) to the M w =5.2 event occurred in July 2004 in Bovec (Slovenia). The propagation velocity and the signal characteristics suggest that the physical process at the origin of the transient can be the spread of a solitary/porosity wave (Rossi et al. , 2016). These waves are packets of fluid-filled cracks, hydraulically interconnected, the upward propagation of which is conditioned by the permeability heterogeneities and by the tectonic stresses (Spiegelman, 1993a, 1993b; Connolly and Podladchikov, 1998, 2013; Miller, 2013). To confirm such a hypothesis, we invert the arrival times of the transient through hydraulic tomography, to obtain hydraulic diffusivity values within the region (Brauchler et al ., 2013). Fig. 1 – a)The strain transient recorded at various GPS stations in the frequency band 1/3.5 yr<f<1./1.5 yr. The GRACE-corrected vertical components (dashed line) and the horizontal component (solid line) in the direction of the maximum amplitude of the transient. We applied constant shifts of 2 mm to the curves to enable comparison between the sites. Gray dashed lines indicate the years. X-axis: time (days since 1/1/2002 (modified from Rossi et al. , 2016b). b) The arrows indicate the direction along which the transients acts in the horizontal plane for the various sites. Amplitude roughly proportional to the one observed.