As oil and gas exploration and production occur in deeper basins and more complex geologic settings, accurate characterization and modeling of reservoirs to improve estimated ultimate recovery (EUR) prediction, optimize well placement and maximize recovery become paramount. Existing technologies for reservoir characterization and modeling have proven inadequate for delivering detailed 3D predictions of reservoir architecture, connectivity and rock quality at scales that impact subsurface flow patterns and reservoir performance. Because of the gap between the geophysical and geologic data available (seismic, well logs, cores) and the data needed to model rock heterogeneities at the reservoir scale, constraints from external analog systems are needed. Existing stratigraphic concepts and deposition models are mostly empirical and seldom provide quantitative constraints on fine-scale reservoir heterogeneity. Current reservoir modeling tools are challenged to accurately replicate complex, nonstationary, rock heterogeneity patterns that control connectivity, such as shale layers that serve as flow baffles and barriers.

Sedimentary deposits are archives of the physical processes that operated on the Earth surface over multiple timescales. These processes can be highly non-linear and fluctuate in time and space to create complex sedimentary and stratigraphic patterns. Enabled by the rapid advancement in digital and computational technology, computational stratigraphy is a physics-based forward modeling system that simulates the Earth surface processes through computation of fluid flow and sediment transport to generate high resolution, geologically realistic models. Using computational stratigraphy models, we can link flow and sediment transport processes to the emergence of earth surface geomorphic structures and predictable stratigraphic patterns at multiple scales. Consequently, computational stratigraphy models can overcome the data sparsity and resolution gap and quantitatively predict reservoir heterogeneity in 3D, across all scales and for any depositional environment (fluvial, shallow marine and deep water).

By integrating computational stratigraphy with existing reservoir characterization and modeling workflows, we can significantly improve reservoir performance forecasts and development uncertainty assessments. Integrated computational stratigraphy workflow combines characterization and interpretation of field data with sedimentologic and stratigraphic concepts to develop geologic scenarios for forward computational stratigraphy modelling. A range of models is constructed to capture potential variations in depositional environment, stratigraphic patterns and basin setting. The developed ensemble of digital analogs defines specific analog predictions that can be validated with well logs and seismic data. Once the ensemble of digital analogs is in place, quantitative prediction of reservoir properties and their 3D distribution can be obtained, and the range of reservoir development uncertainties can also be quantitatively forecasted.

Comparing with traditional reservoir models, from which the predictions, e.g. EUR, are often too narrow and the average predictions are centered on cases that are too homogeneous, computational stratigraphy models capture the full range of reservoir heterogeneities. In addition, computational stratigraphy reservoir forecast variations emerge from a consideration of the fundamental geologic uncertainties, rather than being products of ad hoc rock property ranges and spatial correlation structures. Therefore, computational stratigraphy models with the greatest differences can be compared more directly with the field data, and the range of predictions can more confidently be related to physical reality.