Quantitative Curvature Analysis: A Case Study

Contributors: Evan Staples, Kurt Marfurt, Ze’ev Reches

Last month in this space we analyzed the relations of fracture patterns and layer curvature in clay models. This month we examine these relations in a central Oklahoma field developed by Pathfinder Exploration, Norman, Okla.

The dolomitized reservoir is 50-100 feet thick within the Hunton Group of Late Ordovician to Early Devonian age.

The data include a 3-D time-migrated seismic survey of about nine square miles, and 15,622 feet cumulative length of image-logs in seven horizontal wells.

The interpreted image logs (figure 1) of all wells revealed 3,971 fractures, as well as bedding surfaces and fault-zones. The majority of the fractures are sub-vertical to vertical, and their strike orientations are plotted with color-coding of our quality ranking (A to D) based on visual quality and continuity (figure 1, left portion).

Fracture density was binned in 55-foot bins, which are half the size of 110-foot 3-D seismic bins, for comparison between fracture density and seismic attributes. We assumed that lithological and thickness variations within the horizontal wellbores are minimal.

The fracture orientations are fairly scattered, yet three major trends can be recognized (figure 1):

  • A scattered ENE-WSW trend in wells 1, 2, 3 and 4.
  • A scattered ESE-WNW trend in wells 5 and 6.
  • A NNE-SSW trend in well 7.

We test the hypothesis that the fractures formed primarily as tensile fractures due to local curvature (see part 1 of this column in the July EXPLORER), and compare their density to the 3-D seismic curvature.

Figure 2 shows a top Hunton horizon slice through the most-positive curvature volume. The horizontal wells are displayed with color-coded fracture density (fracture number/55-foot length).

Figure 2 (left) displays a few areas of good correlation between high fracture density and high most-positive curvature values (yellow arrows). In the next step (figure 2), the strike direction for high curvature values are plotted with color denoting the direction as shown by the time slice.

The general E-W strike directions of the curvatures appear to correspond with the high fracture densities in wells 3–7 (figure 1). Wells 1 and 2 do not cross areas with curvature zones of E-W strike directions.

To further examine the correlations between fractures and the curvature, we used a workflow for azimuthally-limited weighted average of curvature features from the 2010, 80th Annunal International Meeting of the SEG entitled Seismic attribute illumination of Woodford Shale faults and fractures, Arkoma Basin, Oklahoma, by Guo and Marfurt.

Azimuthal intensity is:

[total strike length] / [total area of the search window]

This technique is similar to fracture intensity calculations in part one, but filters curvature strike direction by azimuth.

We calculated azimuthal intensity in 15-degree sections and correlated them to fracture densities in the image-logs; high correlation exists where “r” approaches unity (table 1). For example, the interpreted fractures in well 3 strike 30-90 degrees (figure 1), and we found high correlation between fracture density and the azimuthal intensity at 45 degrees and 75 degrees, which are within the range of the interpreted fractures.

One should note that these high correlations are localized only in areas of high curvature, and do not exist along the entire wellbore. The other wells, excluding #1, exhibited similar behavior of areas with high fracture density and high curvature, and also had high correlation with one or more trends of azimuthal intensity.

In last month’s (part one) compressional clay experiments we found that a critical magnitude of the curvature is needed to generate tensile fractures – and below this critical curvature there was no correlation between curvature and fractures.

We think that a similar situation occurs in the horizontal wells: areas of high correlation between azimuthal intensities and fracture density also show that most-positive curvature values highly correlate with fracture densities.

To identify the critical curvature magnitude in the present 3-D seismic area, we took the areas of high correlation between azimuthal intensity and fracture density in the wells and computed the curvature ranges for each high correlation interval. We then link them to the fracture density range (table 2).

It appears that the critical magnitude of curvature in the study area is between 8.71x10-3 mi-1 and 2.58x10-2 mi-1 as these ranges of curvature correspond to fracture densities > 0.5 (fracture/feet), suggesting that curvature induced most of these fractures.

Outside this curvature range the fracture densities are too low to be clearly correlated to the curvature.

We thus propose that the azimuthal intensity method can help to identify locations where curvature strike orientations in the subsurface appear to be locally related to fracture density

Our main conclusions are:

Curvature calculations in clay models and the subsurface appear to follow similar patterns.

In clay models, a critical value of curvature is needed to initiate fractures. Fracture density rapidly increases with increasing strain until the saturation point is reached and few new fractures are generated.

Indications for similar behavior were observed in our subsurface analysis of image-logs and seismic data.

Azimuthal-intensity by strike orientation is an effective filter to compare curvature orientation to fracture orientation.

In our study, correlations between curvature azimuthal intensity and fracture density indicated areas where curvature and fracture density are also highly correlated.

The curvature attribute can serve as a better proxy for fracture intensity when compared with horizontal image logs.

However, strain is only one component in fracture generation, with thickness and lithology (estimated by vertical logs, two-way travel time thickness, and seismic impedance inversion) also playing important roles.

Acknowledgements: Thanks to Pathfinder Exploration for providing the data used in this project, and to Schlumberger for providing software for this research at the University of Oklahoma.

Comments (0)


Geophysical Corner

Geophysical Corner - Satinder Chopra
Satinder Chopra, chief geophysicist (reservoir), at Arcis Seismic Solutions, Calgary, Canada, began serving as the editor of the Geophysical Corner column in 2012.

Geophysical Corner

The Geophysical Corner is a regular column in the EXPLORER that features geophysical case studies, techniques and application to the petroleum industry.


Image Gallery

Meet the Authors

Kurt Marfurt
Kurt Marfurt

Evan Staples is with ConocoPhillips in Houston, and Ze’ev Reches and Kurt Marfurt are with the University of Oklahoma in Norman, Okla. All are AAPG members.

See Also: Bulletin Article

Using diverse geologic and geophysical data from recent exploration and development, and experimental results of analysis of gas content, gas capacity, and gas composition, this article discusses how geologic, structural, and hydrological factors determine the heterogeneous distribution of gas in the Weibei coalbed methane (CBM) field.

The coal rank of the Pennsylvanian no. 5 coal seam is mainly low-volatile bituminous and semianthracite. The total gas content is 2.69 to 16.15 m3/t (95.00–570.33 scf/t), and gas saturation is 26.0% to 93.2%. Burial coalification followed by tectonically driven hydrothermal activity controls not only thermal maturity, but also the quality and quantity of thermogenic gas generated from the coal.

Gas composition indicates that the CBM is dry and of dominantly thermogenic origin. The thermogenic gases have been altered by fractionation that may be related to subsurface water movement in the southern part of the study area.

Three gas accumulation models are identified: (1) gas diffusion and long-distance migration of thermogenic gases to no-flow boundaries for sorption and minor conventional trapping, (2) hydrodynamic trapping of gas in structural lows, and (3) gas loss by hydrodynamic flushing. The first two models are applicable for the formation of two CBM enrichment areas in blocks B3 and B4, whereas the last model explains extremely low gas content and gas saturation in block B5. The variable gas content, saturation, and accumulation characteristics are mainly controlled by these gas accumulation models.

Desktop /Portals/0/PackFlashItemImages/WebReady/variable-gas-content-saturation-and-accumulation.jpg?width=50&h=50&mode=crop&anchor=middlecenter&quality=90amp;encoder=freeimage&progressive=true 5687 Bulletin Article

The Sierra Diablo Mountains of west Texas contain world-class exposures of Lower Permian (Leonardian) platform carbonates. As such, these outcrops offer key insights into the products of carbonate deposition in the transitional icehouse to greenhouse setting of the early to middle Permian that are available in few other places. They also afford an excellent basis for examining how styles of facies and sequence development vary between inner and outer platform settings.

We collected detailed data on the facies composition and architecture of lower Leonardian high-frequency cycles and sequences from outcrops that provide more than 2 mi (3 km) of continuous exposure. We used these data to define facies stacking patterns along depositional dip across the platform in both low- and high-accommodation settings and to document how these patterns vary systematically among and within sequences.

Like icehouse and waning icehouse successions elsewhere, Leonardian platform deposits are highly cyclic; cycles dominantly comprise aggradational upward-shallowing facies successions that vary according to accommodation setting. Cycles stack into longer duration high-frequency sequences (HFSs) that exhibit systematic variations in facies and cycle architectures. Unlike cycles, HFSs can comprise symmetrical upward-shallowing or upward-deepening facies stacks. High-frequency sequences are not readily definable from one-dimensional stratigraphic sections but require dip-parallel two-dimensional sections and, in most cases, HFS boundaries are best defined in middle platform settings where facies contrast and offset are greatest. These studies demonstrate that HFSs are the dominant architectural element in many platform systems. As such, the lessons learned from these remarkable outcrops provide a sound basis for understanding and modeling carbonate facies architecture in other carbonate-platform successions, especially those of the middle to upper Permian.

Desktop /Portals/0/PackFlashItemImages/WebReady/outcrop-based-characterization-leonardian.jpg?width=50&h=50&mode=crop&anchor=middlecenter&quality=90amp;encoder=freeimage&progressive=true 3661 Bulletin Article
We use samples from undeformed and deformed sandstones (single deformation band, deformation band cluster, slip-surface cataclasite, and fault core slip zone) to characterize their petrophysical properties (porosity, permeability, and capillary pressure). Relationships between permeability and porosity are described by power-law regressions where the power-law exponent (D) decreases with the increasing degree of deformation (strain) experienced by the sample from host rock (D, sim9) to fault core (D, sim5). The approaches introduced in this work will allow geologists to use permeability and/or porosity measurements to estimate the capillary pressures and sealing capacity of different fault-related rocks without requiring direct laboratory measurements of capillary pressure. Results show that fault core slip zones have the highest theoretical sealing capacity (gt140-m [459-ft] oil column in extreme cases), although our calculations suggest that deformation bands can locally act as efficiently as fault core slip zones in sealing nonwetting fluids (in this study, oil and CO2). Higher interfacial tension between brine and CO2 (because of the sensitivity of CO2 to temperature and pressure) results in higher capillary pressure and sealing capacity in a brine and CO2 system than a brine and oil system for the same samples.
Desktop /Portals/0/PackFlashItemImages/WebReady/insight-into-petrophysical-properties.jpg?width=50&h=50&mode=crop&anchor=middlecenter&quality=90amp;encoder=freeimage&progressive=true 3716 Bulletin Article

See Also: DL Abstract

Hydrocarbon exploration beneath the shallow allochthonous salt canopy of the ultra-deepwater central Gulf of Mexico has encountered three thick, sand-rich, submarine fan successions that punctuate an otherwise relatively condensed and fine-grained basin center stratigraphy. These sand-rich fans are Late Paleocene, Early Miocene, and Middle Miocene in age and each coincide with periods of very high sediment flux and basin margin instability. They are the primary exploration targets in most ultra-deepwater fields, recent discoveries, and failed exploration tests.

Desktop /Portals/0/images/_site/AAPG-newlogo-vertical-morepadding.jpg?width=50&h=50&mode=crop&anchor=middlecenter&quality=90amp;encoder=freeimage&progressive=true 3079 DL Abstract

See Also: Learn! Blog

The Granite Wash of the Anadarko Basin of Oklahoma and Texas has been a remarkable challenge with prolific results for operators who have solved the mystery of a highly complex play. With up to 18 producing zones, extreme compartmentalization, and pressures that can be challenging, all phases of exploration, drilling, and production require special knowledge.

Desktop /Portals/0/PackFlashItemImages/WebReady/AAPG-learn-blog-200x200.jpg?width=50&h=50&mode=crop&anchor=middlecenter&quality=90amp;encoder=freeimage&progressive=true 12055 Learn! Blog