GNGTS 2013 - Atti del 32° Convegno Nazionale

Crossing the TRI catalog with the localized events by the network, we retrieved a data set of 319 instrumentally located events with their MAW . As that the discrepancy between a vector length and the arithmetic mean of its projections strongly depends on azimuth, we have retrieved the E-W and N-S components of the original recordings. In fact, knowing the azimuth of the single event with respect to the Trieste station, the MAW vector has been resolved with respect to the horizontal (N-S and E-W) directions and then the correct MAW has been calculated, as the arithmetic mean of the two values estimated independently for the two horizontal components. The average overestimation of the MAW reported in the bulletins with respect to the actual MAW is equal to 0.25. The recovered earthquakes are mainly spatially concentrated in two regions: Friuli (a in Fig. 1a), with azimuth centered in the range 300°–330° N, and Dinarides (b in Fig. 1a), with a predominance of less clustered events with azimuth around 120° N. Overall, 56% of the data is placed in azimuthal bands where the two horizontal components have similar amplitude and, consequently, a minimal discrepancy (equal to 0.15) in magnitude computation (mean vs. vectorial composition) is theoretically expected. To check the validity of the operation done, we selected 30 events of this data set, trying to cover the whole range of magnitudes, we recovered the original photographic sheets with the traces of the earthquakes, and we asked to three experts that were part of the seismological Fig. 2 – Orthogonal regression between MAW s and other quantities: a) the MAW simulated on the BB via the WA filter MLBB ; b) M L published by INGV ( MLINGV ). The earthquakes before April 16, 2005 are reported as red empty dia- monds, the next as orange filled diamonds dots; c) M D provided by OGS from October 31, 2004. 118 GNGTS 2013 S essione 1.1