A source study of the Thessaloniki (northern Greece) 1978 earthquake sequence

Abstract

The Thessaloniki (northern Greece) earthquake sequence appears to have occurred along faults forming a graben structure. This graben, situated in the border region between the Serbomacedonian massif and the Vardar zone, is bounded to the south-west by clearly exposed north-west striking north-east dipping normal faults. Relative hypocentre determinations, fault-plane solutions, surface faulting and the aftershock distribution suggest that some of these faults have been reactivated during the 1978 earthquakes. The source parameters of the mainshock (mb=6.1, Ms=6.4) were determined by computing body-wave synthetic seismograms in the time domain and comparing them with the observed. This modelling constrained the orientation of faulting determined by the P-wave fault-plane solution. It also constrained the source depth to 6± 2 km. Similar depths were calculated by a relative relocation method for the other three large events of the sequence. The dislocation time-function required for the far-field had a total duration of 9± 1.5 s, the body-wave moment was 5.2± 1.8 × 1025 dyne cm. For a fault length of 35 km and a fault width of 17 km (both estimated from the aftershock distribution) the static stress drop was found to be 4 bar. An accelerograph record of the mainshock shows two distinct events, 3–4 s apart. These two events were unresolved by the long-period data in the far-field but the unusually long duration of the mainshock time-function suggests an overall slow energy release which probably occurred as a sequence of events close in space and time. Seismic energy released in this fashion can account for waveform complexities observed at some stations.