Reflection-Image-Spectroscopy of a subduction zone

S. Buske(1), S. Shapiro, M. Yoon, (1)Institute of Geological Sciences, Free University of Berlin, Germany
Prague, Oct 29, 2004

We propose a new method for extracting additional information from deep seismic reflection images. In complicated environments it is not always possible to deliver one uniquely valid seismic image. Instead, the image will differ significantly when focusing over different frequency ranges. For example enhanced concentrations of scatterers may be masked in one frequency band and become visible in another frequency band. We therefore apply 3D prestack Kirchhoff depth migration to data containing the full frequency range on one hand and different narrow frequency ranges on the other hand. We call this approach "Reflection Image Spectroscopy" (RIS). It provides frequency dependent images which enable for instance the characterization of the medium in terms of scatterer concentration and length scale. Furthermore, the analysis of these images allows to differentiate between small-scale structures in the high-frequency band and large-scale structures in the low-frequency band.

We applied this technique to an onshore deep seismic reflection line (ANCORP96) across the Chilean subduction zone. The dominant features in the full-frequency image are the mid-crustal Quebrada Blanca Bright Spot (QBBS) and the subducted oceanic crust (Nazca reflector). The latter appears as a thick reflector with no further interpretable internal structure. However, the narrow-frequency images yield more information (see figure below). At a depth of 70 km the Nazca reflector changes its appearance. The clearly separated double reflection zone continues downward but a wedge-shaped body attached to its top can be resolved. We interpret this body as a possible fluid trap above the slab. Serpentinization within this part of the mantle wedge may lead to the increased reflectivity. Further up towards the QBBS a strongly heterogeneous zone can be observed. This zone coincides with a tomographic low-velocity anomaly and can be explained as the location of possible fluid ascend paths, eventually along pre-existing faults. The QBBS therefore seems to be directly linked to the downgoing plate.

Another seismic reflection line (PRECORP95) in that area shows similar features and supports our interpretations.

Last edited Nov 2, 2004