f An efficient approach to the development of 3D velocity models

The selection of appropriate velocities for stacking and migration of seismic data plays a key role in determining the quality of the subsurface image obtained. For 3D surveys this selection is complicated by the large data volumes involved and the practical realities of time constraints. The requirements for quantifiable accuracy and interpretive input makes the construction of the stacking velocity model a lengthy proces. Typically stacking velocity analysis may occupy up to one quater of the total 3D survey processing time with around 1000 local analyses of stacking velocity being involved. Techniques for optimising the efficiency with which these analyses can be made are thus of significant importance. The construction of a reliable velocity model for the migration process is further complicated by the difficulty associated with estimating subsurface interval velocities from the seismic measurements. Routine techniques, utilising the Dix interval velocity calculated from the measured stacking velocities, may leave unquantified uncertainties in the accuracy of the derived velocity model .