GNGTS 2019 - Atti del 38° Convegno Nazionale

GNGTS 2019 S essione 3.1 583 in order to improve the velocity analysis reliability. Once data pre-processing was satisfactory completed, attention turned to define an accurate velocity model for KPSDM. An iterative top-down workflow was adopted to progressively fix the upper layer velocities and horizon positions before moving to define the next series of deeper formations. The workflow borrowed from vintage pre stack time migrated velocities to build the initial velocity model. Velocity analysis methods based on tomography make use of the residual moveout of events (non-flatness) in depth migrated gathers, and their purpose is to determine how to modify the velocity model to reduce progressively the residual move-out. The general rule is: if the event appears as a frown in depth gathers, the overlying velocities in the existing model are too fast; if an event appears as a smile, the velocities are too slow. If the correct velocity is used, primary events on depth migrated gathers align horizontally along the offsets axis and stack coherently, with the result that target reflectors are well imaged and in the proper depth position. On the contrary, if they show a residual curvature, the velocity model needs to be updated. In principle, tomography may result in a very accurate and very high resolution velocity model, However, uniqueness of the solution is not guaranteed. In other words, several velocity models can be equally consistent with the data, and tomography alone cannot pick the right one. This problem can be mitigated by the incorporation of a (reliable) structural framework (e.g., geological horizons) in the velocity model building process. To this purpose, in this project a series of key horizons were selected to define boundaries between tomographic iterations in the shallower terrigenous succession: • Sea bed • Top Pliocene • Messinian Unconformity Another limitation, caused by the data, emerged after several iterations targeting from sea bottom to Messinian Unconformity. Indeed, the maximum source-receiver offset was limited of 3600 mt. The net effect is that residual curvature estimates from image gathers was relatively insensitive to velocity change at depths greater than the maximum available offset. Accordingly, the reliability of tomography had to be considered restricted to the shallower section of the subsurface. In light of such issue, an alternative strategy was used to build the velocity model in the deeper evaporitic domain below the Messinian Unconformity. Sonic logs from a sub-set of good quality check–shot wells were used to derive and propagate interval formation velocities between the major horizons of this deep section until a maximum depth of 12 km. This approach effectively avoided the short-offset limitation for tomography, providing a reliable well-derived interval velocity model. Fig. 1 - Comparison between the vintage PoSTM volume and the new Kirchhoff PSDM on a south-west/north-east section crossing the thrusts.