A new interpretation of seismic tomography in the southern Dead Sea basin using neural network clustering techniques

Abstract

The Dead Sea is a prime location to study the structure and development of pull-apart basins. We analyzed tomographic models of Vp, Vs, and Vp/Vs using self-organizing map clustering techniques. The method allows us to identify major lithologies by their petrophysical signatures. Remapping the clusters into the subsurface reveals the distribution of basin sediments, prebasin sedimentary rocks, and crystalline basement. The Dead Sea basin shows an asymmetric structure with thickness variation from 5 km in the west to 13 km in the east. Most importantly, we identified a distinct, well-defined body under the eastern part of the basin down to 18 km depth. Considering its geometry and petrophysical signature, this unit is interpreted as a buried counterpart of the shallow prebasin sediments encountered outside of the basin and not as crystalline basement. The seismicity distribution supports our results, where events are concentrated along boundaries of the basin and the deep prebasin sedimentary body. Our results suggest that the Dead Sea basin is about 4 km deeper than assumed from previous studies. Key Points Self-organizing maps are used for a joint interpretation of tomography models Deep prebasin sediments were revealed below asymmetric basin structure Seismicity distribution reflects geometry of derived lithological units.