In the last few years, shale resource plays have brought a welcome renaissance to the US onshore E&P business. Companies have competed for the best land positions using the best information available, often drilling ahead of seismic data out of necessity. Drilling has consistently produced results in plays like the Marcellus, Eagle Ford, Bakken, and Niobrara, and many shale players are taking the opportunity to optimize production from resource plays using technology fit for the areas' unique subsurface challenges.

Profit margins are thin in unconventional resources, especially in areas associated with dry gas. Relatively small increases in average initial production rates (IPs) and estimated ultimate recoveries (EURs) across a portfolio make a material impact on the long-term economic viability of a play and bring significant value to shareholders. The search is on for technologies that truly optimize these critical production metrics.

Subsurface models

Subsurface understanding is a key component for production optimization in resource plays because total organic content (TOC), rock brittleness, porosity, closure stress, natural fractures, and pore pressure drive production. By maximizing the insight gained from each available data type, subsurface models identify and predict reservoir rock, meeting the necessary criteria for highly successful wells. Success is determined by knowledge of rock physics properties and correlation to production information plus local knowledge of the field. Wells planned and fraced using these "drill here" models minimize the risk of poor performance due to poor reservoir quality and significantly increase average IP rates and EURs.

Fracture directions were derived using a cumulative oil model. (Images courtesy of SIGMA3)

The workflow

The SIGMA3 workflow uses all available data and input from geophysicists, geologists, and reservoir engineers to accurately represent the subsurface at different scales. A patented workflow organizes and assembles those contributions to create a predictive model that ties all of the data together and tests then constantly against blind wells to validate their predictive capabilities.

Recent projects in North American resource plays have allowed the company to perform high-resolution inversions on prestack seismic data for physical rock properties. In the process, factors such as TOC, frac closure stress, rock brittleness, natural fracture density, and porosity have been analyzed for their weighted predictive capabilities for reservoir properties and production. This approach bridges the gap between geophysics and reservoir engineering, bringing new value to seismic data in shale plays. Finally, engineers have quantitative predictive models they can use to plan well locations and frac jobs.

Geologists provide vital input from core, outcrop, and well log data to create a fracture indicator log that ties seismic information to the physical properties of the earth. Pressure and completion data from reservoir engineers bring knowledge of historical reservoir production into the model. The geological model rigorously applies the geophysical attributes to empirically test the validity of results against well data throughout the process. Well logs are analyzed and used as ground truth to calibrate other interpretations, and quantitative and qualitative production data are incorporated to ensure the model takes into account the oil and gas that has already been produced.

Many recent resource play projects have revealed pore pressure as a key contributing factor affecting frac closure pressure and thermal maturation. This information is incorporated in the models, giving an extra dimension of insight into the factors driving production in some fields.

The map on the left shows cumulative Niobrara production, while the map on the right shows the probability of 25,000 bbl of cumulative oil.

Ahead of the curve in Niobrara

The workflow was recently applied in the Teapot Dome area on the Niobrara shale interval with the goal of understanding production drivers and optimizing future well placement. The data in this area are fairly limited – only 15 wells with sonic and density logs and only one core with some porosity and permeability measurements were available. Fortunately, 3-D post-stack seismic data were available over the entire area, and a full analysis of the available data could be performed.

During the course of this study, several seismic inversions were applied to the 3-D post-stack seismic data, including several deterministic and stochastic inversions, spectral imaging, and volumetric curvature. Each extracted a different aspect of the information contained in the post-stack seismic that was then used in the modeling process. Key seismic horizons and faults were interpreted in time and used to build a 3-D geocellular grid. The grid was depth-converted along with the seismic attributes.

Because naturally occurring fractures have been documented as potential drivers of production in the Teapot Dome area, the SIGMA3 continuous fracture modeling (CFM) technology was incorporated into the workflow. A fracture indicator log was created based on well log and core data and subsequently was correlated to an optimized combination of seismic attributes using a patented workflow involving supervised neural networks. The CFM model accurately predicted fracture location, orientation, and density. Modeling showed selected spectral imaging attributes, curvature, and impedance all played a meaningful role in predicting natural fractures.

When using prestack data and elastic inversion, rock mechanical properties such as Lambda-Rho, Mu-Rho, and the derived rock brittleness allow highly advanced and accurate modeling of the key rock properties that control the performance of the frac jobs and ultimately the performance of the well. The workflow uses extended elastic inversion, providing very high-resolution Vp, Vs, and density properties and additional properties that could be inverted directly from the prestack data. By using stochastic inversion on the mid, near, and far stacks, Vp, Vs , density, and these additional properties can be derived at a resolution of 0.5 ms to 1 ms.

The workflow delivered 3-D models of permeability, porosity, fracture orientation, and fracture density for Teapot Dome. By testing the model against blind wells, high confidence was established in the predictive capability of the work. A strong correlation was shown to exist between known historical production in the field and the 3-D permeability model. Because permeability is strongly tied to natural fracture density in the model, natural fractures were determined to be an important driver for production in this area.

Further modeling of well productivity used cumulative oil production as a proxy for fracture density around the wells. Using only 2-D maps of key rock properties and seismic attributes, it was possible to derive a map showing cumulative oil in place for the area of interest. This map clearly shows hydrocarbon sweet spots, some of which have yet to be drilled.

The way forward

There is no cookie-cutter solution for shale resource plays. Every shale is different, and the factors driving production are different as well. Drawing on extensive experience modeling these reservoirs in the US and abroad, the SIGMA3 solution has proven to be a scientifically sound pragmatic workflow for finding the key drivers and modeling them using the best information available from all the data.

References available.