A time dependent model of elastic stress in the Central Apennines

Seismicity in the Central Apennines is characterized by normal faulting with dip NE-SW near 45°. We show that if the stress at the hypocenter of the 2016 Norcia (Mw=6.5) and 2009 L’Aquila (Mw=6.3 on the Paganica fault) earthquakes originated only from stress transfer from previous historical events occurred in 1315 and 1461 (L’Aquila), 1703 (Montereale plain) and 1703 (Norcia/Valnerina), then the orientation of the principal stress axes would be inconsistent with the observed tensional regime. The additional contribution of a regional stress is thus required to properly align the principal stress axes to those of the moment tensor, but GNSS geodesy provides only stress rates. We empirically estimate a time multiplier for the regional stress rate, computed with a dense GNSS network, such that the principal stress axes resulting from the sum of the stress transferred by previous events and the regional stress rate multiplied by the empirical temporal scale are consistent with normal faulting, both at the L’Aquila and Norcia hypocenters. Based on a Catalogue of 36 events of magnitude larger than 5.6 we estimate the total Coulomb stress at depths and along planes parallel to those of L’Aquila and Norcia. We provide evidence of an asymmetry of the Coulomb stress leading to a stress concentration near the hypocenter of the two events just prior of the 2009 and 2016  earthquakes. This stress anomaly disappeared after the two events. Similar stress patterns are observed for earlier events which took place in 1461 at L’Aquila, 1703 on the Montereale plain and in 1703 at Norcia/Valnerina. The 1997 sequence of Colfiorito exhibits a similar, anisotropic Coulomb stress pattern. Based on the Database of Individual Seismogenic Sources DISS 3.2.1 of INGV we identify as areas of maximum Coulomb stress at present (>2016) the Gran Sasso , the Camerino and Sarnano areas and the area between the San Pio delle Camere, Tocco da Casauria and Sulmona faults.