Toward an accurate reservoir model of heterolithic, tidally-influenced deposits : an ongoing case study in the Sego Sandstone member of the Mancos Shale through second-generation, outcrop-to-subsurface 3D modeling
MetadataShow full item record
Heterolithic tidal reservoirs are common in hydrocarbon settings around the world, but can be underestimated as a contributor to total resource volume due to their complex lithologic and architectural nature. A preliminary, static geologic model was built for the Sego Sandstone, a tidally influenced unit that outcrops in the Book Cliffs of eastern Utah, to investigate variables influencing model construction and subsequent fluid flow. The initial model, built with DecisionSpace software, utilized 39 outcrop logs and 36 adjacent subsurface logs. Six bounding surfaces provided the 3D framework for a spatial and temporal grid containing six facies, each with unique lithology, permeability and porosity. In addition, bed thickness, bioturbation intensity, gamma response, and net sand percentage were noted in sections and utilized as modifiers in the model. Based on limitations of the previous DecisionSpace model results, a second-generation model was undertaken using Petrel software to investigate: 1. the differences between model design and construction within these two software packages, 2. petrel’s ability to incorporate fine-scale heterogeneities and smaller elements such as tidal channels, swatchways and zones of intense bioturbation, and 3. the impact of bar type morphology (confined versus unconfined) and associated shale distributions on fluid flow. Resulting models were used to investigate standard modeling techniques versus multiple-point geostatistical simulation and the construction of training images. Uncertainty analyses are performed on data parameters using successive property realizations to examine fluid flow connectivity between tidal bar-intrabar architectural elements. Model results show simulated architectural elements that match observed outcrop parameters, honor stacking patterns, and provide enhanced facies distribution predictions. Finally, software gridding algorithms are compared to identify which of these algorithms more accurately characterize subsurface geometries, and ensure best prediction of lithofacies distributions. Outcomes illustrate the importance of integrating subseismic scale data into reservoir models and will inform future work in characterizing tidally-influenced reservoir stratigraphy and modeling approaches utilizing outcrop and subsurface data.