Seismic interpretation is a complex process, in which many data types are considered and integrated to create a structural and/or stratigraphic model. In many cases, interpreters rely only on a seismic image or its attributes obtained after data processing and migration. Traditionally, interpretation was started after seismic imaging with little feedback to the seismic imaging process. However, modern depth imaging and tomography require integrating geologic concepts and constraints into the imaging process. Interaction between an interpreter, earth-model builder, and depth imager becomes necessary to improve the accuracy of seismic-derived structures and stratigraphic models. We have developed an example of closing the loop between geophysics and geology via a 3D geophysical basin modeling workflow that uses geologic concepts and rock-physics modeling to estimate a prior anisotropic seismic velocity model. In addition, the use of a geomechanical model to understand stress regimes in the area of interest and to map stress perturbations into a velocity perturbation ensures a stress-consistent earth model. Because the predictive capability of a numerical simulator was strictly related to the quantification of uncertainties related to its inputs (e.g., geometry, boundary conditions, and constitutive laws), uncertainty should be quantified at basin-length scales and geologic time scales. We used the Logan prospect in the deepwater Gulf of Mexico as an example to demonstrate the workflow.