To what extent the accuracy of slip models helps explain the aftershock distribution using CFF : The Mw 5.9, 1999 Athens earthquake case study.
D. Baumont (1), O. Scotti (1), F. Courboulex (2), C. Nostro (3), and N. Melis (4)
(1)Institut de Protection et de Sûreté Nucléaire
(2)UMR Geosciences AZUR-CNRS, Nice, France
(3)Istituto Nationale di Geofisica e Vulcanologia, Roma, Italy
(4) National Observatory of Athens, Institute of Geodynamics, Greece
European Geophysical Society XXVII General Assembly - Nice, 21-26 avril 2002.
On September 7, 1999, a moderate magnitude (Mw 5.9) earthquake struck thé City of Athens. Source processes of moderate earthquakes are not well understood and yet they represent thé main hazard for moderate seismicity regions such as Europe. Using regional seismic and SAR data, it was possible to characterize the complex kinematic rupture of this event. Apparent Source Time Functions (ASTF), obtained using an Empirical Green's Function (EGF) method, reflect a rather complex rupture with a source directivity primarily pointing towards Athens.
To further characterize the faulting, we inverted the ASTFs for the kinematic history of the
rupture using a damped-least square inversion scheme with inequality constraints. The rupture area was found to be mainly confined into a 10 km along-strike x 20 km along-dip zone with a mean slip amplitude equal to 25 cm which corresponds to a relatively low average stress drop of about 1 MPa. The slip concentrates in two elongated zones with amplitudes reaching locally 60 cm. The total rupture lasted between 5 and 6 s. SAR data inversion led to significantly different slip distributions especially on the NW part of thé fault. These differences could be explained by an important contribution of after slip to the surface deformation recorded by the SAR (i.e. 10-to-30% of the total slip).
This study is part of the European project PRESAP (Practical, Real-time Estimation of Spatial Aftershock Probabilities) aiming to a better assessment of seismic hazard in a near real time condition. Computing the Coulomb Failure Function on the 3D optimal planes for all the slip distributions, we showed that heterogeneous slip models better predicts the spatial aftershock distribution than homogeneous slip models (80 versus 30 % of success, respectively). However, the whole stressed areas did not undergo aftershock activity. This may suggest that the "optimally CFF" hypothesis alone is not valid and it should not be decoupled from a structural analysis.