GNGTS 2014 - Atti del 33° Convegno Nazionale

58 GNGTS 2014 S essione 1.1 2003), where the background seismicity delineates the edge of the locked portion of the major thrusts. Thus, the distribution of interseismic coupling and its comparison with instrumental seismicity suggests that a significant amount of tectonic deformation is stored as elastic deformation at seismogenic depths and available to be released in future earthquakes. Seismic moment accumulation and implications for seismic hazard. To compare the interseismic strain accumulation with seismic release, we calculate the rate of seismic moment accumulation. From the computed pattern of interseismic coupling and long-term slip rate, we estimate a rate of seismic moment accumulation of 1.3±0.4 × 10 17 Nm/yr (assuming a shear modulus of 30 Gpa). Our estimated value of moment accumulation is quite robust with respect to the assumed model parameterization (fault dip angle, degree of smoothing of the solution and eastern Alps sliver pole). The estimated rate of moment accumulation should be balanced on the long-term by seismicity release or by alternative aseismic mechanisms. To evaluate the rate at which historical earthquakes have released elastic deformation we select the seismic events that fall within the investigated area from the CPTI11 catalogue. To compute the cumulative earthquake frequencies we use the magnitude binning and the completeness time intervals proposed by Stucchi et al. (2011). Under the assumption that the geodetic moment rate distributes into earthquake sizes that follow a Gutenberg-Richter distribution, truncated to an assumed maximum moment earthquake, the recurrence time of earthquakes of moment ≥ M is (Molnar, 1979): where b is the slope of the Gutenberg-Richter relation and α accounts for the fraction of total seismic moment released by earthquakes. Using the proposed maximum magnitude ( M max = 6.7) for the investigated area of the most recent version of the Italian seismic hazard map (Stucchi et al. , 2011), assessed from the estimated magnitude of the largest historical event, it immediately appears (Fig. 3) that although the relative earthquake distribution is well reproduced with a Gutenberg-Richter relation, the historical seismicity accounts only for a small fraction (α = 0.3-0.5) of the total seismic moment rate. The parameter α can be smaller than 1 because of transient aseismic creep events (post-seismic or spontaneous events), but none was observed during the period covered by the geodetic observations. Because the most likely reason for α < 1 would be post-seismic aseismic creep, that typically releases around 20-25% of coseismic slip (e.g., Perfettini and Avouac, 2007), then α should be in the range between ~0.7 and 1. Alternatively, removing the maximum magnitude constrain from historical seismicity, we can reproduce the observed earthquake frequency distribution respecting a full seismic release of tectonic deformation (α = 1) with a larger M max . This second option requires the extension of M max to ~7.5 to satisfactory reproduce the observed frequencies (Fig. 3). Because the recurrence interval of such “large” ( M > 6.7) earthquakes is > 1000 years, certainly larger than the interval for which we have reliable historical information (Stucchi et al. , 2011), the absence of large earthquakes in the historical record cannot be considered as an anomaly. The effects of a rare M W ~7.5 earthquake would be substantially different from the recent, moderate 1976 Friuli earthquake sequence in several important respect. In particular, a large earthquake would cause strong ground shaking over a much larger area and would have a much longer duration. As regard an intuitive understanding of the sensitivity of probabilistic hazard assessment to changes in the maximum magnitudes and their recurrence intervals, according to the work of Young and Coppersmith (1985), increasing in the upper bound magnitude while keeping a constant seismicity rate, results in an imperceptible reduction of the recurrence rate for smaller earthquake as the number of larger events added represents only a very small fraction of the total number of earthquakes. Given the rate of historically observed earthquakes,