A vast majority of marine geological research is based on academic
seismic data collected with single-channel systems or short-offset
multi-channel seismic cables, which often lack reflection moveout for
conventional velocity analysis. Consequently, our understanding of earth
processes often relies on seismic time sections, which hampers
quantitative analysis in terms of depth, formation thicknesses, or dip
angles of faults. In order to overcome these limitations, we present a
robust diffraction extraction scheme that models and adaptively
subtracts the reflected wavefield from the data. We use diffractions to
estimate insightful wavefront attributes and perform wavefront
tomography to obtain laterally resolved seismic velocity information in
depth. Using diffraction focusing as a quality control tool, we perform
an interpretation-driven refinement to derive a geologically plausible
depth-velocity-model. In a final step, we perform depth migration to
arrive at a spatial reconstruction of the shallow crust. Further, we
focus the diffracted wavefield to demonstrate how these diffraction
images can be used as physics-guided attribute maps to support the
identification of faults and unconformities. We demonstrate the
potential of this processing scheme by its application to a seismic line
from the Santorini-Amorgos Tectonic Zone, located on the Hellenic
Volcanic Arc, which is notorious for its catastrophic volcanic
eruptions, earthquakes, and tsunamis. The resulting depth image allows a
refined fault pattern delineation and, for the first time, a
quantitative analysis of the basin stratigraphy. We conclude that
diffraction-based data analysis is a decisive factor, especially when
the acquisition geometry of seismic data does not allow conventional
velocity analysis.