Predicting Reservoir Presence from Seismic Velocity Mapping and Pore Pressure Prediction

Enhanced seismic velocity models (e.g. semblance or tomography) are key inputs for greenfield pore pressure prediction (PPP) in clastic basins, enabling velocity extraction along the planned well-path and also from a 3d earth-model surrounding the drill-site to test sensitivities and ranges. If velocity model resolution is sufficient, the resultant pressure gradients extracted along interpreted stratigraphic horizons from one of these three-dimensional (3D) earth models yields valuable insights into lateral pressure transfer efficacy, thus informing the net-to-gross (e.g. reservoir presence or absence) for the interval. The following inputs are required: 1) 3D seismic velocity model; 2) locally calibrated transforms for Compressional Velocity (Vp) to Vertical Effective Stress (VES) to generate a 3D pressure model from seismic velocities; 3) high-quality seismic interpretation horizons. The critical step is to extract from the 3D pressure model the pressure along each mapped horizon and plot the pressure with depth for each. For a shale horizon, the result reflects a linear shale gradient (e.g. ~0.83 psi/ft) consistent with restricted lateral pressure communication. Conversely, extraction along sand-interval yields a brine gradient (e.g., ~0.47 psi/ft) indicative of effective lateral pressure communication. Two Deepwater Gulf of Mexico (GOM) salt-withdrawal clastic mini-basins are depicted in seismic cross-section, each with representative suite of interpreted seismic horizons for pressure analysis within the depth-range of interest. It is not apparent to the uncalibrated-eye, in lieu of well data, which of the two is reservoir-prone and which is reservoir-absent. A detailed summary of methods and assumptions will follow, including empirical transforms, seismic earth cube generation, map-based extraction, and pressure gradient plots. The resultant gradient plots reveal one basin with uniform brine gradients and ubiquitous lateral pressure communication, compared with a suite of exclusively shale gradients in the other. Finally, well results, one test from each basin, confirm ubiquitous reservoir in the basin with the predicted brine gradients, contrasted with absence of reservoir in the basin with shale gradients. Following the presentation, the audience will require nothing more than graphical inspection of the event-extracted pressure gradient plots to discriminate a sand interval from a shale, and a reservoir-prone mini-basin from a reservoir-lean one.