24th IUGG General Assembly, Perugia, Italy, July 11-13, 2007
Analisa Romeo, Raffaella De Matteis, Giuseppe Pasquale, Giovanni Iannaccone, Aldo Zollo
The Southern Apennines is one of the Italian areas characterized by most intense geodynamic activity; as evidenced by the seismic catalogues in historical time it was strucked by several destructive earthquakes (Boschi et al.,1998). The most recent one occurred on 1980, 23 november, M 6.9, producing more than 3000 casualties and extended damages in all the region. Nowadays the seismic activity is characterized by low-moderate earthquakes.
In this work we present a three-dimensional P-wave velocity crustal model of Southern Apennines region using travel time of local earthquakes. In this model we re-localize the earthquakes and then we compute the focal mechanisms in order to estimate the stress field acting in the area.
In order to obtain a 3D velocity model we merged the data collected from 1988 to 2003 by Istituto Nazionale di Geofisica e Vulcanologia (INGV) network (Italian National Seismic Network) with arrival time of aftershocks of 1980 M 6.9 Irpinia earthquake recorded by a temporary network. Only the earthquakes which were recorded at least eight stations were considered and the final database consist of 1196 earthquakes with 15500 P and 7000 S arrival time readings.
We used the linearized, iterative tomographic approach proposed by Benz et al. (1996), which allows inversion of local earthquakes first-arrival travel-time to solve simultaneous velocity model parameters and hypocentral parameters.
The iterative technique takes into account the nonlinearity of the problem but in each iteration the method is based on a linear approach. This means that the starting model will influence the inversion process. In order to avoid falling in a local minimum, the final velocity model was obtained as the mean of the 250 velocity values assumed by each cell in the 3D inversions performed with different 1D reference models randomly generated in a selected range. Uncertainty associated to the velocity value for each cell has been analyzed and the model resolution has been evaluated through a standard checkerboard test.
In order to give an improved representation of the seismicity pattern, we have re-localized the 1196 earthquakes in the 3D final velocity model using a probabilistic, non-linear, global-search earthquake location method (NonLinLoc code, Lomax et al., 2000). The relocated seismicity is shifted eastward caused by a low velocity zone in the eastern part of the investigated area. It is clustered around the system fault of the 1980 Irpinia earthquake and have a depth ranging between 0-20 km.
For the computation of best fit double-couple focal mechanisms we used the fault-plane fit grid-search algorithm (FPFIT) of Reasemberg & Oppenheir (1985). Restrictions are place on the minimum number of first arrival polarities (six polarities), maximum acceptable RMS residual (≤ 0.5 s) and maximum values for ERX (≤ 0.8 km), ERY (≤ 0.8 km), ERZ (≤ 0.8 km) and gap ≤ 180. The FPFIT algorithm computes the orientation of the P and T axes, but there is no correspondence between P and T axes and the principal stress orientations. With the purpose of investigating the stress field in this region, we applied the Michael (1984) procedure to our data set of fault-plane solutions. It computes the best uniform stress field for the dataset and meaningful confidence regions using a statistical tool known as bootstrap resampling. Stress inversion shows a nearly horizontal NE-SW minimum compressive stress axis (3), while maximum compressive axis (1) is vertical. This result reveals that Southern Apennines is generally ongoing through NE-SW extension, as evidenced by the fault-plane solutions of major earthquakes occurring along the Apenninic Belt.