Rupture of the 1997 Umbria-Marche Largest Earthquakes, Interaction between faults

Bruno Hernandez, Oona Scotti, Fabrice Cotton
Institut de Protection et de Sûreté Nucléaire, Fontenay-aux-Roses, France

Michel Campillo
Laboratoire de Géophysique Interne et Tectonophysique, Université Joseph Fourier, Grenoble, France

Françoise Courboulex
UMR Géosciences Azur-CNRS, Valbone, France

Massimo Cocco and Salvatore Stramondo
Istituto Nazionale di Geofisica, Via di Vigna Murata, 605, Rome, Italy

Abstract. Geodetic and seismological coseismic data are used to constrain the spatio-temporal evolution of the slip on the fault planes during the three largest earthquakes of the 1997-1998 Umbria-Marche (Apennines, Central Italy) seismic sequence. More than 20 events of magnitude larger than 4.5, and about 2000 smaller shocks occurred during this sequence. On September 26, 1997 two closely located earthquakes struck the Colfiorito area (00:33, Mw=5.7 and 09:40, Mw=6.0), and on October 14, 1997, (15:23, Mw=5.6) another event occurred in the southeastern section of the seismogenic zone, near Sellano.
We first use GPS data and SAR interferograms to evaluate the coseismic slip distribution on the fault planes in a layered half space. We use different fault geometries and locations constraining the solution which provides the best fit to the data by reducing the rms misfit. We also explore the set of acceptable solutions using a genetic algorithm. We finally invert strong ground motion data to describe the time progression of the rupture.
As far as the 0:33 event is concerned, the slip is mainly located in the northern part of the fault near the hypocenter location. For the 9:40 event, the moment released is mainly located in the first 5 km north to the focus. For this event, the rupture initiated in the southern part of the fault and propagated toward the North-West. A second smaller asperity is observed in the upper part of the fault at about 10 km north of the nucleation zone. We emphasize the importance of performing such combined inversion to constrain the kinematic parameters avoiding the trade-off problem between slip and time encountered in classical source inversion schemes.