Abruzzi region (central Italy) producing vast damage in the L'Aquila town and surroundings. In this paper we present the location and geometry of the fault system as obtained by the analysis of main shock and aftershocks recorded by permanent and temporary networks. The distribution of aftershocks, 712 selected events with M L ! 2.3 and 20 with M L ! 4.0, defines a complex, 40 km long, NW trending extensional structure. The main shock fault segment extends for 15-18 km and dips at 45°to the SW, between 10 and 2 km depth. The extent of aftershocks coincides with the surface trace of the Paganica fault, a poorly known normal fault that, after the event, has been quoted to accommodate the extension of the area. We observe a migration of seismicity to the north on an echelon fault that can rupture in future large earthquakes.

We image the rupture history of the 2009 L'Aquila (central Italy) earthquake using a nonlinear joint inversion of strong motion and GPS data. This earthquake ruptured a normal fault striking along the Apennines axis and dipping to the SW. The inferred slip distribution is heterogeneous and characterized by a small, shallow slip patch located up‐dip from the hypocenter (9.5 km depth) and a large, deeper patch located southeastward. The rupture velocity is larger in the up‐dip than in the along‐strike direction. This difference can be partially accounted by the crustal structure, which is characterized by a high velocity layer above the hypocenter and a lower velocity below. The latter velocity seems to have affected the along strike propagation since the largest slip patch is located at depths between 9 and 14 km. The imaged slip distribution correlates well with the on‐fault aftershock pattern as well as with mapped surface breakages.

On 24 August 2016 a magnitude ML 6.0 occurred in the Central Apennines (Italy) between Amatrice and Norcia causing nearly 300 fatalities. The main shock ruptured a NNW‐SSE striking, WSW dipping normal fault. We invert waveforms from 26 three‐component strong motion accelerometers, filtered between 0.02 and 0.5 Hz, within 45 km from the fault. The inferred slip distribution is heterogeneous and characterized by two shallow slip patches updip and NW from the hypocenter, respectively. The rupture history shows bilateral propagation and a relatively high rupture velocity (3.1 km/s). The imaged rupture history produced evident directivity effects both N‐NW and SE of the hypocenter, explaining near‐source peak ground motions. Fault dimensions and peak slip values are large for a moderate‐magnitude earthquake. The retrieved rupture model fits the recorded ground velocities up to 1 Hz, corroborating the effects of rupture directivity and slip heterogeneity on ground shaking and damage pattern.

S U M M A R YIn Italy, the Mercalli-Cancani-Sieberg (MCS) is the intensity scale in use to describe the level of earthquake ground shaking, and its subsequent effects on communities and on the built environment. This scale differs to some extent from the Mercalli Modified scale in use in other countries and adopted as standard within the USGS-ShakeMap procedure to predict intensities from observed instrumental data. We have assembled a new PGM/MCS-intensity data set from the Italian database of macroseismic information, DBMI04, and the Italian accelerometric database, ITACA. We have determined new regression relations between intensities and PGM parameters (acceleration and velocity). Since both PGM parameters and intensities suffer of consistent uncertainties we have used the orthogonal distance regression technique. The new relations are I MCS = 1.68 ± 0.22 + 2.58 ± 0.14 log PG A, σ = 0.35 and I MCS = 5.11 ± 0.07 + 2.35 ± 0.09 log PGV, σ = 0.26.Tests designed to assess the robustness of the estimated coefficients have shown that singleline parametrizations for the regression are sufficient to model the data within the model uncertainties. The relations have been inserted in the Italian implementation of the USGSShakeMap to determine intensity maps from instrumental data and to determine PGM maps from the sole intensity values. Comparisons carried out for earthquakes where both kinds of data are available have shown the general effectiveness of the relations.

We study the 30 October 2016 Norcia earthquake (MW 6.5) to retrieve the rupture history by jointly inverting seismograms and coseismic Global Positioning System displacements obtained by dense local networks. The adopted fault geometry consists of a main normal fault striking N155° and dipping 47° belonging to the Mt. Vettore‐Mt. Bove fault system (VBFS) and a secondary fault plane striking N210° and dipping 36° to the NW. The coseismic rupture initiated on the VBFS and propagated with similar rupture velocity on both fault planes. Updip from the nucleation point, two main slip patches have been imaged on these fault segments, both characterized by similar peak‐slip values (~3 m) and rupture times (~3 s). After the breakage of the two main slip patches, coseismic rupture further propagated southeastward along the VBFS, rupturing again the same fault portion that slipped during the 24 August earthquake. The retrieved coseismic slip distribution is consistent with the observed surface breakages and the deformation pattern inferred from interferometric synthetic aperture radar measurements. Our results show that three different fault systems were activated during the 30 October earthquake. The composite rupture model inferred in this study provides evidences that also a deep portion of the NNE trending section of the Olevano‐Antrodoco‐Sibillini thrust broke coseismically, implying the kinematic inversion of a thrust ramp. The obtained rupture history indicates that in this sector of the Apennines, compressional structures inherited from past tectonics can alternatively segment boundaries of NW trending active normal faults or break coseismically during moderate‐to‐large magnitude earthquakes.