A comparative study of near-surface velocity model building derived by 3D traveltime Tomography and Dispersion Curves Inversion techniques

摘要

In an attempt to reconstruct the near-surface velocity for the shallow subsurface in the area under study using two different approaches: 3D Traveltime Tomography and Dispersion Curves Inversion (DCI), a total of 26286 shot records were used from Matruh 3D onshore seismic survey; acquired in 1998. Building a near-surface velocity model based on 3D traveltime tomography required First-Breaks (FB) picking in the offset-time (x-t) domain. While, a near-surface model from DCI approach needed a data transformation into frequency-slowness (f-p) domain to allow for dispersion curves picking per shot gather. The inversion of each approach set of picks resulted in a gridded near-surface velocity model on 100 mx100 mx50 m grid for X, Y and Z, respectively. The results clearly demonstrated that both models showed similar velocity behavior in the first 100 m, however, DCI output model exhibited a high velocity layer (100 m thick) with velocities up to 5000 m/s between depth (-100 m and 0 m) compared to slower ones in the tomographic model within this depth interval. Meanwhile, the tomographic inversion results have revealed this high velocity layer quite deeper; starting at +100 m depth and were able to distinguish only the top of this high velocity layer. The validity of the two output near-surface models was examined through: (i) direct calculation of static corrections and (ii) merging the near surface velocity model with a deeper depth model then running pre-stack depth migration (PSDM) on two seismic test lines. Statics corrections were not able to distinguish a more valid model over the other, as they had very subtle impact on the stack quality. On the other hand, PSDM migration results clearly indicated that DCI model is constraining the shallow subsurface better compared with the tomographic model.