Evaluation of Regional Travel-Time and Location Improvement Along the Tethyan Margin Using a New Three-Dimensional Velocity Model

摘要

For purposes of studying the lateral heterogeneity as well as predicting seismograms for this region, we construct a new 3-D S-velocity model by jointly inverting a variety of different seismic data. We jointly invert regional waveforms, surface wave group velocity measurements, teleseismic S and SKS arrival times, and crustal thickness estimates from receiver functions, refraction lines, and gravity surveys. These data types have complementary resolving power for crust and mantle structures, vertical and lateral variations, shallow and deep mantle features, local and global structure. Therefore, a joint inversion of these data sets might help unravel the complexity of this tectonically diverse area. These measurements are made from a combination of MIDSEA, PASSCAL, GeoScope, Geofon, GSN, IDA, MedNet, national networks, and local deployments throughout the study region which extends from the western Mediterranean region to the Hindu Kush and encompasses northern Africa, the Arabian peninsula, the Middle East, and part of the Atlantic Ocean for reference. The Moho depth result is broadly consistent with CRUST2.0, except in mid-northern Africa, where the crust from our joint inversion is about 5 km thinner. Fast velocity anomalies are found beneath the West African Craton, the Hellenic trench, the Apennines, the East European Platform, and the Arabian Platform at a 75:150 km depth, whereas low-velocity anomalies are located along the plate boundaries such as the mid-Atlantic ridge, Afar, the Anatolian Plateau, Iran, Afghanistan, western Mediterranean Sea, and the Red Sea. Based on the assumption that P-wave velocity anomalies in m/s scale are very close to S-wave velocity anomalies in m/s scale in our study area, we convert our S-velocity model to a P-velocity model with use of teleseismic P arrival times. Finally, our model is validated by performing travel-time predictions with a dataset of ground truth events. Our model generally produces better travel-time predict.