Lateral variations of the upper mantle structure in southeastern Europe and Asia Minor as derived from travel-time data

Abstract

To s tudy lateral velocity variations in the upper mantle in southeastern Europe, Asia Minor and the eastern Mediterranean we used travel times of P waves as reported by the International Seismological Centre for stations in Europe, Asia Minor and northern Africa from earthquakes in the region under investigation for 1970-1981. The earthquakes have been grouped into 18 localized source zones, each zone containing more than 20 events and being of a size less than 1 ° × 1 ° × 40 km (depth). The averaged travel-time residuals (with respect to Jeffreys-Bullen tables), as well as s tandard deviations, have been calculated at each station for each source zone. For inversion we used only those averaged residuals for which s tandard deviations did not exceed 0.25 s. Such data have been obtained for 270 paths. To estimate the 3-D pattern of velocity corrections to Jeffreys' v d curve, we applied the method based on Backus-Gi lber t formalism, which was used previously by Gobarenko and Yanovskaya (1983) for determining the upper mantle structure in the Al ta i -Sayan region. A disadvantage of this method is that a solution for relative velocity corrections, 8v(O, A, z)/vo(z), is obtained in the form of a sum of three functions, each of them depending on one coordinate only. But, in practice, the solution fairly reflects the main peculiarities of the real velocity pattern if a horizontal axis coincides with the direction of the strike of the geological structures. Since the strikes of the structures in the region under investigation are of different directions, we divided the region into three partly overlapped suhregions: Carpa th ian-Balkan zone and Aegean Sea, Eastern Mediterranean and west Turkey, and Asia Minor and Black Sea. In each of the subregions, a principal direction of strike could be chosen (Gobarenko et al., 1986). The number of the travel-time residuals used for the inversion was equal to 100 for each subregion. In addition, some data for the two first subregions have been combined with the data for large epicentral distances, corresponding to penetration of the rays at depths up to 300 km. This data set was used for determining the velocity distribution at a depth of 250-300 km. The results of the inversion for all subregions have been combined to construct the 3-D velocity distribution for the whole region over the depth range 50-300 km (Gobarenko et al., 1987). Within the overlapped areas the velocity corrections have been averaged. A pattern of the velocity corrections at 100 km depth is shown in Fig. 1. At the depths of about 200 km, a general decrease of the velocity is observed for almost the whole region, with the exception of the east Mediterranean and Asia Minor, indicating the existence of the low-velocity asthenosphere in southeastern Europe. A comparison of the velocity anomalies in the upper mantle with the regional isostatic anomalies and the heat flow pattern is presented in Fig. 2. It shows a striking correlation of the high (low) velocities with negative (positive) gravity