An integrated workflow for selecting optimal drilling locations

When determining where to place horizontal wells in gas shale reservoirs, a number of criteria have to be considered, including petrophysical, geological, and geophysical factors. Water saturations, kerogen content, rock properties, stress fields, karsting, faulting, and fracturing of the Barnett have been evaluated and integrated into a single model for use in the decision-making process. Both wireline log data and rich-azimuthal 3-D seismic data are utilized in the interpretation workflow.

Petrophysical interpretation
Unconventional reservoirs require specialized interpretation techniques. The first step in the evaluation is to determine the volume of kerogen present in the Barnett shale interval. A modified Passey approach is used to calculate the volume of kerogen. Passey proposed that the volume of kerogen could be determined by overlying the sonic curve on the deep resistivity curve — the separation between the two curves is proportional to the volume and maturity of the kerogen. The modified Passey approach we are using models a sonic curve from the deep resistivity via the Smith transform, and the separation between the modeled sonic curve and the actual sonic log is directly proportional to the kerogen volume. Both the Passey method and the modified Passey method require some core data for calibration purposes in any particular basin.

The determination of the mineralogy and fluid content is the next step in the petrophysical interpretation process. The principle of optimizing petrophysics is used to calculate the volumes of minerals and fluids within the Barnett shale, with the modified Passey kerogen volume constraining the computation. The model consists of calcite, quartz, illite, pyrite, kerogen, bound water, free water, and gas. In addition, a seismic trace was extracted along the well bore and displayed with the curve data. This seismic-to-well display validates the velocity model used for time-to-depth conversion as well as allowing the geophysicist to correlate specific seismic responses to log responses. This petrophysical interpretation is displayed in Figure 1.

Figure 1. Petrophysical analysis of Marble Falls, Barnett shale, and upper Ellenberger formations. a) Gamma ray; b) lithology; c) depth; d) resistivity; e) closure Stress; f) water saturation, porosity, and gas volume; g) Poisson’s ratio, Young’s modulus, brittleness; and h) seismic trace. (Images courtesy of Paradigm)

Notice a closure stress curve displayed in track e of Figure 1 and additional rock properties in track g, including Poisson’s ratio in green, Young’s modulus in red, and a brittleness curve in purple. Both the closure stress and brittleness curves were calculated using equations published by Rickman, Mullen, et al in 2008. Desirable characteristics for a Barnett shale reservoir include high brittleness and low closure stress. Using these criteria, it can be noted that the Barnett intervals with higher quartz content are generally more amenable to fracture stimulation than zones with more limestone and clay.

Another characteristic of shale gas reservoirs that needs to be considered is the presence or absence of open natural fractures. These fractures can have a major impact on the completion and productivity of the Barnett shale interval. Fracture communication between the water-bearing Ellenberger and the gas-bearing Barnett will render the Barnett well non-commercial. The fracture stimulation propagation can be heavily influenced by open or healed fractures. A review of the formation microscanner images (FMI, Figure 2) in the Idaho #1 well indicates minimal, if any, open fractures, and the fractures that are imaged are the result of borehole breakout and present no hazard to the well completion. The borehole breakout is oriented from northwest to southeast and provides valuable insight into the anisotropic horizontal stress fields present in the Barnett shale.

Figure 2. FMI images over a Barnett shale interval.

Seismic processing and well ties
The seismic data volume was generated using continuous full-azimuth and angle-domain image processing. The process decomposes the recorded seismic wave-field into full-azimuth reflection angle gathers that can be used to generate high-quality seismic volumes using preferential azimuth and angle data. The decomposition process also generates full-azimuth directional angle gathers that can be used to separate specular (continuous) energy from scattered (discontinuous) energy for preferential enhancement of data.

The high-amplitude reflections at the top of both the Marble Falls and Ellenberger formations (Figure 1, track h) serve to highlight the accurate seismic-to-well tie. The top of the Barnett is a low-amplitude event that is not clear in many areas and is difficult to map.

Figure 3. Perspective view of top Ellenberger depth structure map.

Surface mapping and interpretation
The Ellenberger and Marble Falls horizons were mapped using a combination of automatic and manual methods. Maps of the Barnett and upper Barnett were generated using a technique incorporating both the Marble Falls and Ellenberger horizons. Fifteen proportional slices were generated between the Marble Falls and Ellenberger, the idea being that the top of the Barnett is better estimated using a combination of the bounding surfaces versus attempting to pick the Barnett directly. Specific slices that best fit the top Barnett and upper Barnett were converted into horizons.

Both Marble Falls and Ellenberger maps show the major structural features, including north/south trending major faults, fault ramps, and secondary en echelon fault segments. The details of complex Ellenberger karst features are enhanced using angled light sources (Figure 3). All karst features imaged on the Marble Falls show a characteristic drape structure. Coherence, curvature, and dip attributes on the Ellenburger and Marble Falls surfaces clearly show karsts, major faults, and minor lineations in the relatively unstructured or quiet areas. A quiet area is identified in the north-central portion of the 3-D survey, where risk of existing communication between the Barnett and Ellenberger is expected to be lower. This is the primary focus area of this study.

Figure 4. Barnett seismic trace variations (red and green).

Barnett trace shape and volume classifications
Examination of seismic traces within the Barnett show lateral wave shape variations (Figure 4). Neural network technology is used to classify the traces into 11 categories. The results show an east/west trending pattern likely representing changes in thickness across the survey. An upper Barnett “sweet” zone identified from petrophysical interpretation is characterized by an increase in quartz content. Lateral amplitude variations are detected in the Barnett interval throughout study area 1. Amplitude increases are interpreted as a slightly thicker zone of higher quartz content. A multiple attribute classification volume was generated focusing on this zone using seismic envelope and frequency. An upper Barnett seismic event within the attribute volume was detected and converted into a 3-D geobody to map the extent and thickness of this zone.

Figure 5. Volume blending of amplitude and coherence (blue), showing karsts.

Imaging permeability hazards
Potential high-risk permeability zones connecting the Barnett and water-charged Ellenberger need to be identified and avoided to increase potential for a commercial well. Major karsts and faults are imaged on surfaces and depth slices; however, detailed mapping of these features requires additional attribute analysis. Surface dip attribute and curvature volumes are used to identify geomorphologic features. Integrating this with enhanced vertical sections improves the understanding of structural elements. Three-D coherence and amplitude data are blended to enhance the visible appearance of the stratigraphic section to aid in defining vertical fault extents, karst zones, and subtle features in the relatively unstructured areas (Figure 5). A vector azimuth volume is generated from coherence data, and intervals within the Barnett are extracted and show patterns of lineaments that reflect the local stress fields and the density of stress-related events. The style and density variations appear to be related to faulting, karsting, and multiple regional stress regimes. An area of low-vector azimuth intensity corresponds to relatively unstructured “quiet” areas, which are potential drilling locations.

Figure 6. Perspective view of structure map on Ellenberger: seismic (grey), Barnett geobody (yellow-green), vector azimuth lineaments, and karst chimneys.

A fault-enhanced volume is generated from coherence data to help define karst chimneys for geobody extraction. Extraction of karst systems as 3-D multi-Z surfaces shows details of the collapsed chimney, often with intersecting fault planes that propagate beyond the karst geobody limit and extend the area of permeability risk. Integrating all of these interpretation results generates an optimal well location (Figure 6).

Data integration
Petrophysical analysis identifies important criteria for placement of a well path within Barnett shale. Integrating seismic, interpretation, and fault-related attributes with karst topology provide the operator with additional data necessary to design an optimal horizontal well.