A novel statistical analysis, which augments conventional interpretation of deep seismic reflection/refraction data, is applied to the DOBRE 2000/2001 reflection profile. The goal is to improve delineation of the lithosphere in terms of lithotectonic units and to compare these to existing interpretations. After a successful validation of the method using synthetic data, stochastic parameters estimated from windowed average lateral autocorrelations in the DOBRE data are compiled in a two-dimensional profile of von Karman-type causative reflectivity. These estimates in terms of lateral correlation lengths and power law exponents are accompanied by associated uncertainties. Given its low uncertainties, the lateral correlation length a(x) turns out to be a robust delineator. The profile of a(x) reveals systematic spatial variations in the lithospheric fabric below the Donbas Fold Belt. As in earlier interpretations based upon conventional processing of reflection/refraction data, both the sedimentary basin and Moho discontinuity stand out clearly, as well as a region indicating massive vertical intrusion in the crust from upper mantle sills and ultramafic underplating of the lower crust. Notable differences with conventional interpretations include the number and extent of supra-Moho rift pillow structures, the lack of imaging a crust-cutting dislocation feature and a laterally disturbed uppermost mantle. Von Karman lateral correlation length provides new independent information at a scale between velocity models from wide-angle reflection/refraction data and line drawings from near vertical reflections and provides new insights and understanding of lithospheric evolution.
- Dniepr-Donets Basin
- seismic velocity heterogeneity
- crustal heterogeneity