Imaging heterogeneity of the crust adjacent to the Dead Sea fault using ambient seismic noise tomography

For the purpose of studying the Earth’s crust by means of tomography, we investigated cross-correlation functions emerging from long-term observations of propagating ambient seismic noise at pairs of broadband stations in Israel and Jordan. The data was provided by the eight permanent broadband stations of the Israel Seismic Network evenly distributed over Israel and the 30 stations of the DESERT2000 experiment distributed across the Arava Fault (South of the Dead Sea basin). To eliminate the influence of earthquakes and explosions, we have applied the methodology of Bensen et al. (Geophys J Int 169:1239–1260, 2007), including bandpass filtering and amplitude normalization in time and frequency domain. The cross-correlation functions estimated from continuous recordings of a few months were used to extract Rayleigh waves group velocity dispersion curves using automatic version of the frequency–time analysis procedure. Subsequently, these curves have been converted into the Rayleigh wave group velocity maps in the period range 5–20 s and S waves velocity maps in the depth range 5–15 km. In these maps, four velocity anomalies are prominent. Two of them are outlined by the previous reflection-refraction profiles and body wave tomography studies, i.e. a low velocity anomaly corresponds to the area of the extremely deep (down to 14 km) sedimentary infill in the Southern Dead Sea Basin and a high velocity anomaly in the Southern Jordan corresponds to the area of the Precambrian crystalline rocks of the Nubian Shield on the flanks of the Red Sea. The two other anomalies have not been reported before - the high velocity zone close to the Beersheba city and the low velocity anomaly in the region of Samaria-Carmel mountains - Southern Galilee. They have relatively low resolution and hence need further investigations for approving and contouring. The highest contrast between the average Rayleigh wave group velocity (2.7 km/s) and the anomalies is 10–13 %, comparable, however, to the level of noise in the data. The results have been verified by modeling the revealed anomalies which showed that all the four zones mentioned above could be detected by the tomography study.