Empirical mode skeletonization of deep crustal seismic data: Theory and applications

Seismic skeletonization is a pattern recognition technique used to decompose reflection seismic data into events or reflectors and their seismic attributes, and to store the results in an “event file” for an analysis toward seismic interpretation. The extraction of instantaneous frequency from the event file for any statistical analysis of the seismic attribute analysis is currently not possible, thus precluding the complete parameterization of the reflected seismic wave field in terms of its decomposition products. Empirical mode decomposition of reflection seismic data has uniquely addressed the question of computing the instantaneous frequency from its decomposition products using the Hubert transform. Although the decomposition products of the seismic skeletonization and the empirical mode decomposition are conceptually different, they share a common link to the primitive features of the seismic waveforms. In this paper, we introduce a new decomposition technique, empirical mode skeletonization, that combines the features of both seismic skeletonization and empirical mode decomposition. By this process, it is now possible to subject the reflection seismic data to an analysis that could include a parameterization of the reflected wave field. We apply the new technique to a segment of seismic line 1 of the Lithoprobe Slave Northern Cordillera Lithosphere Evolution (SNORCLE) transect to extract event-oriented instantaneous frequency attributes and also to analyze the decomposition products for any scaling behavior of the reflected wave field.