Item

Abstract

The CALIXTO (Carpathian Arc Lithosphere X-Tomography) experiment offers a dense, high-quality data set to study the lithospheric/asthenospheric system underneath SE Romania, an earthquake-prone region in SE Europe. To increase the image resolution of structures in the uppermost mantle, the application of crustal traveltime corrections by a priori information before the teleseismic traveltime inversion has become a well-accepted procedure. For such a correction we present a regional 3-D crustal seismic velocity model that serves as the basis for a high-resolution teleseismic tomography (forthcoming paper by Martin et al.). Our 3-D crustal model is based on recent research in the region. We collect new results from two long-range seismic refraction lines, 3-D refraction tomography and teleseismic Ps conversions. Adding previously published models of the sediment distribution, Conrad and Moho depths, as well as crustal seismic P-wave velocities, we compile a 3-D crustal model for SE Romania. This 3-D model does not contain shallow small-scale heterogeneities (<10 km), but it reflects the large-scale structures such as variations in sediment thickness, average seismic velocities and 3-D Moho depth. It is well suited for the correction of teleseismic traveltime residuals, a prerequisite for a high-resolution teleseismic tomography study: for example, traveltime delays of up to 1.3 s are caused by the almost 20-km thick layer of sediments in the Focsani Basin. Such delays are comparable to or larger in size than the expected upper mantle traveltime residuals. We study the significance of 3-D crustal traveltime corrections relative to 1-D station corrections and show that the complex basin structures in SE Romania require a 3-D approach to reduce the smearing of crustal anomalies into the mantle. By modelling synthetic mantle structures with a slab, as it is expected for SE Romania, we also investigate how to adapt the inversion strategy, if crustal corrections are applied. Significant improvements are found by including the already corrected crustal layers in the inversion procedure, thereby enabling the inversion algorithm to project still remaining uncertainties in the less-resolved upper crustal layers. However, the fixing of the upper layers during the inversion due to the a priori knowledge of the crustal velocity anomalies clearly leads to smearing of uncorrected anomalies that are possibly located close to the crust–mantle boundary.