AUTHORS

Julia F. W. Gale � Bureau of EconomicGeology, John A. and Katherine G. JacksonSchool of Geosciences, University of Texasat Austin, J. J. Pickle Research Campus,10100 Burnet Road, Austin, Texas 78758;julia.gale@beg.utexas.edu

Julia Gale obtained a Ph.D. in structural geologyfrom Exeter University in 1987. She taughtstructural geology and tectonics for 12 yearsat the University of Derby. She moved to theUniversity of Texas at Austin in 1998, workingas a research associate first in the Departmentof Geological Sciences and then at the Bureau ofEconomic Geology. Her interests include frac-ture characterization in carbonate and shalehydrocarbon reservoirs.

Leonel A. Gomez � Department of Geologi-cal Sciences, John A. and Katherine G. JacksonSchool of Geosciences, University of Texasat Austin, Austin, Texas 78713; present address:ExxonMobil Upstream Research Company,3120 Buffalo Speedway, Houston, Texas 77098;leonel.gomez@exxonmobil.com

Leonel Gomez obtained a Ph.D. in structuralgeology from the University of Texas at Austinin 2007, working on the quantification of spa-tial arrangement of fractures. Before his grad-uate studies, he worked for 7 years at theColombia offices of Shell and Chevron. He hasa B.Sc. degree in geology from UniversidadIndustrial de Santander in his native Bucara-manga, Colombia. Leonel is currently working atthe upstream research center of ExxonMobilin Houston.

ACKNOWLEDGEMENTS

The fracture characterization project describedhere was undertaken as part of a UniversityLands Advanced Recovery Initiative betweenGoldrus Production Company and the Bureauof Economic Geology. Funding was providedby the University of Texas System and the U.S.Department of Energy (DOE award numberDE-PS26-02NT15378). Members of the Univer-sity of Texas Fracture Research and ApplicationConsortium supported our general fracture re-search. Rob Reed supervised the scanning elec-tron microscope-based cathodoluminescence

Late opening-mode fracturesin karst-brecciated dolostonesof the Lower OrdovicianEllenburger Group, west Texas:Recognition, characterization,and implications for fluid flowJulia F. W. Gale and Leonel A. Gomez

ABSTRACT

Two distinct groups of fractures in an Ellenburger Group reservoir

in Barnhart field, Reagan County, west Texas, were identified. The

oldest fractures (FBR) are the most numerous; have irregular shapes,

sediment, and baroque dolomite fill, and no preferred orientation;

and have been attributed by previous workers to brecciation asso-

ciated with the collapse of Lower Ordovician paleocave systems.

Younger, subvertical, opening-mode fractures (FY) that have consis-

tent east-southeast and south-southwest strikes postdate the baroque

dolomite cement. FY fractures therefore formed during the late

stages or after the Pennsylvanian Ouachita orogeny. We analyzed FYfracture orientation, intensity, and openness using well image logs,

oriented rotary-drilled sidewall cores, and a full-diameter core. FYfracture aperture sizes range from several micrometers to a fewmilli-

meters, and the fracture intensity is consistentwithin and between the

wells studied at 1.8–4.0� 10�2 fractures/mm2 for fractures�1 mm

(�0.04 in.) wide.

Dolomite cement that is synchronous with FY fracture opening

seals fractures in some locations, but is limited to fracture linings and

mineral bridges in other places. Calcite, which grew after FY fractures

stopped opening, is variably present and postdates dolomite cements.

Where present, calcite occludes most remaining FY fracture porosity.

Diagnosing the presence of postkinematic calcite is therefore an im-

portant step in being able to predict open fractures and was done for

part of Barnhart field using rotary-drilled sidewall cores.

AAPG Bulletin, v. 91, no. 7 (July 2007), pp. 1005–1023 1005

Copyright #2007. The American Association of Petroleum Geologists. All rights reserved.

Manuscript received June 6, 2006; provisional acceptance August 22, 2006; revised manuscript receivedDecember 14, 2006; final acceptance March 13, 2007.

DOI:10.1306/03130706066

INTRODUCTION

Variation in performance between wells in hydrocarbon reservoirs

can be caused by fracture systems. An understanding of the origin

and timing of the fractures is essential if reliable predictions of frac-

ture attributes and their effects on fluid flow are to bemade between

wells. Unfortunately, there are typically too few large, measurable

fractures from readily available data sources, such as vertical core and

well image logs, on which to base these predictions.

Opening-mode fracture sets generally comprise a range of fracture

aperture sizes from approximately 1 mm up to about 1 m (3.3 ft). We

define amacrofracture as a fracture that can be observed with the eye,

whereas a microfracture requires magnification greater than 10� to

be detected. Microfractures form at the same time as the larger,

fluid-transmitting fractures, as part of a power-law aperture-size dis-

tribution (Marrett et al., 1999; Gale, 2004; Gale et al., 2004). The

microfractures are generally subparallel tomacrofractures in the same

set (Laubach, 1997, 2003). Microfractures are sufficiently abundant

in core samples for representative measurements of fracture orien-

tation, intensity, and sealing to be obtained. We then use scaling rela-

tionships to predict intensity of macrofractures. Observations of

fracture cements on the thin-section scale provide information for

predicting whether large fractures are open, even if those large frac-

tures have not been observed directly (Laubach, 2003).

These techniques for macrofracture attribute prediction were

developed in sandstones (Laubach, 1997, 2003;Laubach et al., 2004a),

but are here applied to fractured dolostones of theLowerOrdovician

Ellenburger Group in Barnhart field, Reagan County, west Texas.

Fractures associated with cave-collapse breccias (here termed FBR)

are well known in the Ellenburger Group and were encountered in

this study. We described these briefly, but this study focuses on a

second group of opening-mode fracture sets (here termed FY) that

postdate both FBR fractures and late baroque dolomite cement. These

late fractures have not previously been characterized. Because they are

locally open, FY fractures can enhance permeability by linking porous

but low-permeability volumes of rock to the wellbore. Our intention

with this article is to demonstrate techniques for using core data,

including sidewall cores, to better characterize fracture systems.

We do not have sufficient data to provide a full fracture model for

Barnhart field; instead, we offer a methodology for obtaining reli-

able fracture attribute data and show how these methods helped us

characterize a previously undescribed fracture system.

Geological Setting and Barnhart Field Production History

The Ellenburger Group reservoir in Barnhart field (Figure 1) com-

prises shallow-water, Lower Ordovician dolostones and limestones

(Ross, 1976; Holtz and Kerans, 1992). They typically have low per-

meabilities and high initial water saturation despite relatively high

porosity. Reservoirs developed in these rocks were assigned to the

selectively dolomitizedEllenburger ramp carbonate subplay byDutton

and energy-dispersive spectroscopy data col-lection and assisted with the interpretation ofmicrostructures. Steve Laubach, Bob Loucks,and Steve Ruppel provided valuable insight anddiscussion and helpful reviews of early drafts.Steve Cumella, two anonymous reviewers, andFrances Whitehurst, AAPG geology consultant,suggested several improvements. West TexasGeological Society and the Bureau of Eco-nomic Geology gave permission to use figures.This article is published with permission of thedirector of the Bureau of Economic Geology,University of Texas at Austin.

1006 Fractures in Ellenburger Dolostones

et al. (2005). Lee and Friedman (1987), Amthor and

Friedman (1991), andKupecz andLand (1991)described

the petrography and geochemistry of dolostones of the

Ellenburger Group in west Texas. All described cement

sequences in the rock mass and briefly mentioned the

presence of opening-mode fractures. Fracture size, shape,

and intensity were not described.

The Ellenburger Group in this region is commonly

brecciated. The brecciation is attributed to karst-related

processes that acted during two extensive periods of

subaerial exposure, one during the early Middle Ordo-

vician and one during the Pennsylvanian (Kerans, 1990;

Candelaria and Reed, 1992, and references therein).

Paleocave facies associated with karst modification, in-

cluding chaotic breccia, cave-sediment fill, disturbedhost

rock, and collapsed ceiling slabs (Loucks, 1999; Loucks

and Mescher, 2002; Loucks et al., 2004), cause marked

reservoir heterogeneity (Hammes et al., 1996). These

features and postbreccia baroque dolomite were de-

scribed in detail in the Barnhart field by Combs et al.

(2003) and Loucks (2003), who recognized some late

postbaroque dolomite fractures but described them no

further. Breccias in the Barnhart field contain many iso-

lated open interclast pores and partly open FBR fractures

that may contribute to the storage capacity of the reser-

voir, but which are unlikely to enhance permeability

(Combs et al., 2003). Combs et al. (2003) calculated a

mean porosity of 5.7% and a permeability of 5.12md for

breccias with sediment. These are slightly higher than

the total mean for all the collapsed paleocave facies

(4.8% and 3.4 md), which are typical of Ellenburger

reservoir rocks in west Texas (Holtz and Kerans, 1992).

The Barnhart field was discovered in 1941. Produc-

tion peaked at 4500 BOPD 4 yr after discovery, although

production rates and bottom-hole pressure declined

sharply (Gomez et al., 2001). In 1968, a pilot waterflood

programwas introduced to reverse pressure decline and

increaseproduction, butmost of thewells in the fieldwere

abandoned by the early 1970s. Production response and

well-pressure data suggested a fracture influence in the

field (J.Lomax,Goldrus,2001,personal communication).

DATA COLLECTION AND METHODOLOGY

Data were collected from three wells (Figure 1). We

first analyzed rotary-drilled sidewall cores and image

logs from two 50-yr-old wells that had been reentered

(the wells had been completed open-hole): the Supe-

riorOilCompany 48A-1University (A-1) and Superior

Oil Company 48A-6 University (A-6). After sidewall-

core analysis was completed, a new well, the Goldrus

unit 3 Barnhart (unit 3), was drilled. We next studied

the 4-in. (10-cm)-diameter core and image log from that

well. Theunit 3 corewas not oriented during drilling. An

attempt to orient the unit 3 core using the image log was

successful only for the top30 ft (9.1m)becausewe could

not identify common features on account of brecciation.

Figure 1. Regional map of the Permian Basin showing playtypes (from Dutton et al., 2005, reprinted with permission fromthe Bureau of Economic Geology) and the location of the Barnhartfield. Inset shows the contoured structure map of the top ofthe Ellenburger reservoir in the Barnhart field and distributionof University leases. From Cotton (1966), reprinted with per-mission from the West Texas Geological Society.

Gale and Gomez 1007

The core was broken into several sections that could not

be fit back together, and becausemost of the pieces were

not orientable, it was not possible to measure fracture

orientations directly from the core. We therefore used

the image log to obtain fracture orientation information

for this well. Little unbrecciated rock exists in the unit 3

well, and precise stratigraphic correlation with the A-1

and A-6 wells is not possible. In this article, we describe

the A-1 and A-6 sidewall cores first, followed by the

unit 3whole core data to reflect the chronological order

in which our study proceeded. Findings are then inte-

grated in the discussion.

Macrofracture Description in Sidewall Cores

The first step in fracture analysis is to identify types

of fractures present and their orientation. Initially, our

only source of fracture information was the two 50-yr-

old wells. No fracture information had been collected

in these wells when they were drilled. We sought frac-

ture information from a combination of image logs and

sidewall cores. The wellbores were pitted because of

many years of alteration of the open hole and damage

caused by acidizing treatments performed to enhance

oil production. Because thewellbores were badly pitted,

the pads of the image logging tool did not maintain good

contact with the wellbore surface, and the image logs

were poor quality, although the holes where the side-

wall coreswere collected are clearly seen in theA-1well

image log (Figure 2).

An account of the methods for fracture orientation

using sidewall cores was given in Gomez et al. (2001).

To keep track of the orientation of sidewall cores, we

mapped the shape, size, and location of fractures and

sedimentary features in each sidewall core (Figure 3).

A technique for orienting rotary-drilled sidewall

coreswas developed by Laubach andDoherty (1999) in

the absence of an orienting device on the sidewall-core

drill. This technique, along with evidence that small

fractures are valid guides to large fractures, means that

sidewall cores can be used to collectmeaningful fracture

data, although large fractures may not be sampled. A

quantitative measure of the reliability of sidewall-core

orientation was obtained using a numerical ranking

scheme devised by Laubach and Doherty (1999). For

successful orientation of rotary-drilled sidewall cores,

three conditions must be met. First, the azimuth of the

sidewall core must bemeasured in a wellbore image. Sec-

ond, the core-facing direction must be established: one

end of the core cylinder will face the wellbore, whereas

the other end will face the rock formation. Third and

most challenging, the up direction for the core cylinder

must be established. Themost reliable top indicator is a

sedimentary structure, although they are rare in such

small cores. In the absence of such structures, Laubach

and Doherty (1999) found that mechanical indicators,

resulting from the drilling and breakoff of rotary-drilled

sidewall cores, could be used (Figure 3). The certainty

of each observation is assigned a rank, and the ranks for

all observations are summed to indicate the most re-

liably oriented cores (Table 1). The strategy for decid-

ing which sidewall cores to use for fracture-orientation

analysis was to select intact core having reliable orien-

tation and visible fractures (Table 1). For this selection,

the core hole must be visible on the image log, and it

must have a top indicator rank of �3. Of the 45 side-

wall cores drilled in the A-1 well, 38 could be iden-

tified on the image log. Natural fractures were present

in 22 cores in the A-1 well. Relative fracture orien-

tations were measurable in 15 cores, but only 8 cores

satisfied all conditions. Although 23 cores contain nat-

ural fractures in the A-6 well, there was no azimuth

information because the sidewall cores were drilled

after the image log had been acquired. Fractures in

cores having low orientation certainty were examined

petrographically, but were not included in the orien-

tation analysis.

Figure 2. Representative section of image log from the A-1well showing vertical streaking caused by imperfect pad-wellborecontact of the tool, which produced a poor-quality image. Twosidewall-core holes are easily seen, however, making the im-age log useful for sidewall-core orientation (from Gomez et al.,2001, reprinted with permission from the West Texas Geolog-ical Society).

1008 Fractures in Ellenburger Dolostones

Microfractures

Statistically significant populations of microfractures

were obtained from horizontal, oriented thin sections of

the sidewall cores. In the Ellenburger Group, dolomite

fracture cements grow in crystallographic continuity

with dolomite in the fracture walls (Gomez et al., 2001)

andare indistinguishableusing conventional petrography,

but canbe seenusing scanning electronmicroscope-based

cathodoluminescence (SEM-CL) (Figure 4). Because of

small luminescence variations between different phases

of dolomite, small dolomite-filled fractures can be seen

and measured with SEM-CL, and textures marking ce-

ment growth are apparent. We used techniques devel-

oped byReed andMilliken (2003) to overcome imaging

challenges in carbonate rocks caused by phosphores-

cence. The SEM was also used to acquire secondary

electron imagery (SEI) and element mapping over spe-

cific areas of the polished thin section to show porosity

and compositional variation, respectively.

Because well trajectories and fracture orientation

are both subvertical, very few FY macrofractures in any

sample and similarly fewmicrofractures in SEM images

exist. To obtain a sufficiently large data set for orien-

tation determination, we collectedmicrofracture data

from an imagemosaic. Themosaic comprises many con-

tiguous images over an elongate rectangular area of the

thin section in a direction normal to the orientation of

the fracture set being quantified. Typically, image mo-

saics are approximately 300 mmwide (single individual

SEM-CL image using �250 magnification) and a few

millimeters long.

Rose diagrams of fracture strike were generated

for each horizontal thin section. Because fractures are

Figure 3. Schematic diagram and photographs of mechanical indicators in a sidewall core. Four elevations of sidewall-corephotographs showing pitting on the borehole end, which facilitates distinction between the borehole end and the formation end. A lipand smear may be present on one side of the formation end of the core and, together with the remnant curvature of the boreholeend, indicate the core top. The two remaining elevations display sketches of a sidewall core with the location of two orthogonal thinsections (TS) and orientation notches.

Gale and Gomez 1009

commonly curved, we measured the azimuth of two

segments from the midpoint of the widest part to the

tips and then length-weighted each segment in the rose

diagram. Such measurement gives a better reflection of

the orientation distribution of the fractures than would

a non–length-weighted diagram of singlemeasurements

(Gomez and Laubach, 2006; their figure 5). Both FYfracture sets are subvertical; therefore, a rose diagram

of fracture strike is representative of the variation in

orientation between different fracture sets.Where pos-

sible, rose diagrams are presented as true orientations; if

true orientation is not known, diagrams are presented

relative to a datum to demonstrate that relative strikes

of sets could be identified.

RESULTS

Fracture Orientation from Sidewall Cores

Macrofracture Orientation

Two sets of steeply dipping fractures, striking northeast-

southwest and northwest-southeast, are apparent from

13 oriented fractures measured in the sidewall cores

(Figure 5a). These data were used to inform drilling of

the unit 3 well. By contrast, four fractures in 116 ft

Table 1. Number of Sidewall Cores from Wells A-1 and A-6

Meeting One or All of the Selection Criteria*

Number of Sidewall Cores

Condition A-1 Well A-6 Well

Drilled 45 34

Recovered 44 (98%) 32 (94%)

Observed on borehole

image logs

38 (86%)

116 ft (35.3 m)

of image log

0

115 ft (35.0 m)

of image log

With top-core rank � 3 15 (34%) 14 (44%)

One or more

fracture sets

22 (50%) 23 (71%)

Meets all conditions 8 (18%) 0

*Complete recovery, azimuth located on borehole image, core-top rank �3,and presence of macroscopic fractures. The core-top rank is the sum of theorientation indicator ranks. For example, the certainty for an azimuth basedon an image log can take a range of 0–4, with 0 being zero certainty. Thehighest possible numerical scores for all core-top indicators total 24. A corerank of 3–5 gives acceptable core orientation reliability. Cores with totalscores below 3 were rejected for fracture orientation purposes. The per-centage of recovered cores with core-top ranks greater than 3 was 34% inA-1 and 44% in A-6; the greatest rank attained by an individual core was 6.

Figure 4. Micrographs of fractures from the A-6 well sidewall core comparing SEM-based imagery with conventional petrography.(a) Plane-polarized light photomicrograph; (b) SEM-CL image of pale-gray dolomite (L) forming the rock matrix and dark-graydolomite (D) partly filling the large fracture, sealing the small fracture (newly identified fracture), and filling some pores in the matrix.A low-luminescent dolomite may be seen in the large fracture, forming the outer zone of some dark-gray crystals. Fracture porosity(P) is present in the large fracture. Note that the small fracture is not visible in (a), and there is no distinction between the differentdolomite cements. Modified from Gomez et al. (2001). (c) Enlarged section of (b), showing rhombic dolomite bridge.

1010 Fractures in Ellenburger Dolostones

(35.4 m) of the A-1 image log have no consistent orien-

tation and are neither numerous nor consistent enough

to indicate preferred fracture orientation (Figure 5a).

One steeply dipping fracture at 9103 ft (2775m) trends

north-northeast–south-southwest, but the other three

fracture planes have much shallower dips. After the

unit 3 well was drilled, we were able to compare the

sidewall-core fracture orientations with fractures indi-

cated on the unit 3 image log through an equivalent in-

terval to determine whether the sidewall cores provided

accurate orientation data. A total of 111 fractures exist,

striking north-northeast–south-southwest and north-

west-southeast, in 180 ft (55m) of the unit 3 well image

log (Figure 5b, c). These match fracture orientations

from the sidewall cores and one fracture on the A-1

image log at 9103 ft (2775m). The other three fracture

orientations on the A-1 image log do not match frac-

tures on the unit 3 image log.

Microfracture Orientation

Rose diagrams for microfractures from two oriented

sidewall-core image mosaics from 9055 and 9087 ft

(2760 and 2770 m) show strong preferred orienta-

tion of fractures in the northeast-southwest orientation

(Figure 6). However, because cathodoluminescencemo-

saics are oriented perpendicular to the dominant frac-

ture set, they undersample fractures subparallel to the

mosaic length. Orthogonal fracture sets were observed

in two of the three image mosaics from the A-1 well at

9031 and 9055 ft (2753 and 2760 m) (Figure 6c). The

northeast-southwest set is dominant, but an orthogo-

nal west-northwest–east-southeast set is also present.

These orientations correspond closely to macrofracture

orientations.

Fracture Characterization in Full-Diameter 4-inch Core

The unit 3 core is extensively brecciated: some fractures

are partly filled with sediment and baroque dolomite

cement, and others are partly filled with dolomite or

calcite cement. Apparent macrofracture intensity varies

with depth. We use the term ‘‘apparent intensity’’ to

indicate that the local intensity of fractures observed

at any sampling point in core or an image log may not

reflect the fracture intensity away from the wellbore.

Fractures may be more or less intense than suggested

by the sampling.

A synthesis of petrography on samples from the

unit 3 well (Figure 7) shows a paragenetic sequence

common to most samples:

1. dolomitization (rockswere already dolomitized prior

to fracturing)

2. brecciation fracturing (FBR) and vug formation

(Figure 7a, b)

3. sediment fill between breccia clasts, inwedge-shaped

gaps between blocks and in FBR fractures

Figure 5. (a) Lower hemisphere equal-area projection of oriented fractures from the A-1 well, showing poles to fracture planesfrom oriented sidewall cores (n = 13, dots) and poles to fracture planes from borehole images (n = 4, crosses). (b) Lower hemisphereequal-area projection with 1% area contours for poles to fracture planes (n = 111, small crosses) from borehole images of unit 3.(c) Representative segment of electric image log from unit 3 well. Two approximately orthogonal fractures were interpreted from thissegment. A total of 111 fractures were interpreted in 180 ft (55 m) of log. From Gomez et al. (2001), reprinted with permission fromthe West Texas Geological Society.

Gale and Gomez 1011

4. baroque dolomite filling remaining FBR pore space

(Figure 7c)

5. pressure solution creating subvertical stylolites

(Figure 7d)

6. opening-mode fractures (FY) that crosscut breccia

clasts and baroque dolomite (Figure 7a, b)

Kupecz and Land (1991) determined that the El-

lenburger Group was mostly dolomitized before the

onset of pre-Middle Ordovician karstification, and we

agreewith this finding on the basis of fracture-dolomite

crosscutting relationships. Breccias in the unit 3 core

were interpreted by Loucks (2003) as cave-collapse brec-

cias of various facies, recognizing that sediment-filled

fractures, in addition to interclast gaps, must have been

open while the sediment was being transported through

the cave system. Remaining pore space in these frac-

tures and vugs between clasts may contain white ba-

roque dolomite, sometimes forming geopetal structures

(Figure 7c). FBR fracture orientations are highly variable,

and we did not measure them in this study. Moreover,

FBR fracture paths are commonly tortuous and branched

(Figure 7a, b), initiating or terminating at subtlemechan-

ical boundaries. Fractures radiating out from a point at

clast-clast contacts are also seen. Other small fractures

follow subtle mottling and sedimentary layering and

are, hence, very shallow dipping. The overall effect is a

highly tortuous fracture network on the centimeter

scale. Loucks (1999) described this type of network as

‘‘crackle breccia.’’

Truncated fragments of breccia and baroque dolo-

mite cement are preserved within the teeth of steeply

dipping tectonic stylolites (Figure 7d), indicating that

the stylolites postdate both the brecciation and the

low-aspect-ratio baroque dolomite-sealed fractures. The

trend of the stylolites is not known because they were

observed in fragments of unoriented core. Fractures

trending normal to the stylolites are present and could

be concurrentwith the stylolites, but it is unclear if they

are part of the FY set.

Breccia-related fractures are postdated by straight

persistent fractures (FY) that crosscut clasts, matrix,

and cement (Figure 7a, b). FY fractures have a range of

sizes (length, height, and aperture), and lengths and

sometimes heights are commonly truncated by the

core. FY fractures have greater length-to-aperture as-

pect ratios than FBR fractures. The FY fractures are typi-

cally clustered, although we have not assessed whether

the fractures are more clustered than expected for a

random distribution.

Figure 6. Length-weighted rose diagrams of microfracture orientations derived from thin-section analysis. (a, b) Fracture segmentsfrom horizontal thin sections from two oriented sidewall cores from the A-1 well. The azimuth of these fractures matches theorientation of one of the sets observed in the unit 3 well image log. (c) Scanning electron microscope-based cathodoluminescenceimage of microfractures in the A-1 well, showing two sets of fractures.

1012 Fractures in Ellenburger Dolostones

Fracture Orientation

Fracture orientations were difficult to determine in the

unit 3 core because the onlymeans of orienting the core

was to compare fractures in the image log with frac-

tures in the core, and this comparison had to be done

separately for each section of the core that could not be

fit together. We attempted to determine the dominant

orientation of FY fractures at selected depths where a

Figure 7. Core slabphotographs of sectionsfrom the unit 3 well:(a) 9036.5 ft (2754 m).Steeply dipping fracturesthat postdate white ba-roque dolomite, vug-filling cement. Thin frac-tures are confined withina single breccia clast,whereas others are wider(�1 mm; �0.04 in.) andhave an irregular pathlinking several vugs.(b) Sample at 9050.1 ft(2758 m). Fractures re-lated to brecciation cutby straight-sided, high-aspect-ratio, low-tortuosityfractures that cut throughboth clasts and cement.These have a dominantorientation, are steeplydipping (approximately70j), and are complete-ly sealed with cement.(c) Sample at 9071.05 ft(2764.85 m). Networkof steep and layer-parallelfractures of uncertainorigin. Fracture fill is whitebaroque dolomite andsediment. The large whitefracture at top right con-tains a geopetal fill struc-ture (GP). Fracture path-ways and terminations arecontrolled by sedimen-tary and diagenetic fabrics.This facies was identifiedas a crushed slab related tokarst collapse by Loucks(2003). (d) Sample at9121.2 ft (2780.1 m).Steeply dipping, tectonicstylolites (S) postdatebrecciation and baroquedolomite cement.

Gale and Gomez 1013

dominant orientation was also seen on the image log.

We have assumed that the fractures in the core are those

seen on the image log. Fracture orientations in the unit 3

image log (previously shown in Figure 5b) are displayed

in stereographic projections split out by depths with

common dominant orientation (Figure 8). These depth

sections do not correlate with any facies or log signature

and possibly reflect mechanical rock properties at the

time of fracturing that are not apparent now. Fracture

strikes for the whole image log are shown by depth

(Figure 9a). Between depths of 9069 and 9081 ft (2764

and 2768 m), a fairly consistent fracture orientation

was on the image log (Figure 9b, c), together with de-

monstrably late, strongly oriented FY fractures in the

core (Figure 9d, e). We therefore matched the west-

northwest–east-southeast fractures on the image log

with the observed fractures in the core at those depths.

We selected these for intensity measurements because

they are the most certain orientations we could obtain

for any FY fractures in the unit 3 core. Similar, demon-

strably late FY fractures were also observed at depths

of 9036.5, 9042, 9050.1, 9052, and 9090.4 ft (2754,

2756, 2758, 2759, and 2771 m) in the unit 3 core.

Structural Diagenesis

Timing of fracturing relative to diagenetic and tectonic

events is critical. For example, fractures growing very

near the surface in sandstones and dolostones may re-

main open because quartz and dolomite cement pre-

cipitation is slow at low temperatures. For fractures

growing in amore reactive environment, cement growth

is faster, and diagenesis and fracturing are intimately

linked, reflecting the sum of mechanical and chemical

processes (Gale et al., 2004; Laubach et al., 2004a). It

is particularly desirable to know whether cements pre-

cipitated before, during, or after fracture formation

(pre-, syn-, or postkinematic). Prekinematic cements

affect mechanical rock properties and, therefore, frac-

ture patterns (Olson et al., 2004). Synkinematic cements,

present as overgrowths in the host rock, seal narrow

fractures and bridge and line wider fractures, leaving

the widest fractures in a set partly open. The relative

rates and amount of fracture opening and synkinematic

cement growth control whether fractures seal during

fracturing (Laubach, 2003). Crack seal is a deformation

mechanism in which fractures open in multiple small

increments, and each small fracture is sealed by cement

precipitation before the next occurs (Ramsay, 1980;

Laubach, 2003). Crack-seal texture is themain evidence

of synkinematic cement. Postkinematic cements poten-

tially cause the most damage to fracture porosity be-

cause they can completely occlude larger fractures.

Only small amounts of postkinematic cement are pres-

ent in the host rock because most porosity has already

been occluded by earlier cement. To make accurate pre-

dictions of open fractures, however, recognizing post-

kinematic cement is important (Laubach, 1997; 2003),

and we show here that it can be accomplished using

sidewall cores.

Cements in FY Fractures in Sidewall Cores

Key observations of relative fracture and cement tim-

ing are summarized in Table 2. Crack-seal texture oc-

curs in both low luminescence (LLC) (Figure 10) and

moderate luminescence (MLC) (Figure 6c) dolomite ce-

ments in largemicrofractures (>100 mm). The smallest

fractures do not display crack-seal texture because they

become sealed after only one increment of opening

(Laubach, 2003). Low luminescence dolomite (black

in images) is present in both sets of FY fractures, com-

pletely filling the thinmicrofractures and formingbridges

in the wide ones (Table 2, observation A). Moderate

luminescence dolomite postdates the LLC, and calcite

cement is latest, sealing the remaining fracture poros-

ity. Intersections between fractures from FY sets are rare

(Table 2, observation B), but where seen, the cements

appear to seal the whole intersection area with no ob-

vious boundaries, suggesting that both fracture sets

formed simultaneously. Evidence of fracture propaga-

tion along internal dolomite crystal boundaries indicates

that matrix dolomitization predated fracture growth

(Table 2, observationC).Discontinuous fractures and

fractures with irregular pathways are common in both

A-1 (Figure 11f, g) and A-6 wells (Figure 4). In a few

cases, LLC is the latest cement (Table 2, observations D

and E). If the LLC precipitated broadly synchronously

across the field, these fractures must be older than

other FY fractures.

In both the A-1 and A-6 wells, parts of some

fractures are filledwith bitumen, and in other fractures,

thewalls are coatedwith, or are surroundedby, bitumen

(Gomez et al., 2001, their figure 11). Approximately

half the fractures in the A-1 are filled with dolomite

or calcite, and half are filled with bitumen. Most A-6

well fractures contain bitumen, but only a few have

calcite cement. Bitumen could have acted as a cement-

growth inhibitor, preventing postkinematic calcite from

precipitating.

We recorded the openness of FY fracture sets in the

A-1 and A-6 wells (Figure 12). The number of fracture

1014 Fractures in Ellenburger Dolostones

Figure 8. Lower-hemisphere stereograph-ic projections of polesto fractures in the unit 3well image log, separatedso that dominant orien-tations are identified bydepth. Poles to east-southeast– trending frac-tures are solid circles;poles to south-southwest–trending fractures areopen squares.

Gale and Gomez 1015

Figure 9. (a) Strike of fractures depicted in Figure 8. Strike value is the one viewed along a dipping plane when the dip is down tothe right (right-hand azimuth). Gray-white shading denotes depth intervals corresponding to the individual stereograms in Figure 8.(b) Stereographic projection and (c) rose diagram of fracture orientations in the 9069- to 9081-ft (2764- to 2767-m) depth range. Inthis interval, there are demonstrably late (FY) fractures. (d, e) Rose diagrams of fracture trends and photographs of core disks usedfor FY fracture intensity analysis.

1016 Fractures in Ellenburger Dolostones

sets identified is plotted for all sample depths for each

well, together with the ratio of open to present frac-

ture sets. If all fractures are open, the ratio is 1; if all

fractures are sealed, the ratio is 0. Nearly all fractures are

sealed in the A-1 well, whereas there are many open

fractures in the A-6 well. FY fractures in the unit 3 well

are mostly too narrow to retain porosity after dolo-

mite precipitation, so we do not know whether larger

fractures would retain porosity or be sealed with cal-

cite. The wider distribution of postkinematic calcite in

the Barnhart field between the sampled wells remains

to be constrained.

Element Mapping

Wemapped elementswithin selected areas using energy-

dispersive spectroscopy to determine variation in com-

position of the different dolomite crystals present. No

compositional variationwas detected between the LLC

crystals of the rock matrix and the MLC present in

some intergranular pores and in fractures (Figure 13).

However, element mapping indicates that the LLC

present in some fractures contains slightly more iron

and less magnesium than other dolomite.

Fracture Aperture Size Distributions

Fracture aperture sizes were measured from four SEM

image mosaics (Figure 14). Two samples from the A-1

well (Figure 14a, b) yielded microfractures of the same

fracture set (SW; Figure 6) sufficient to generate a re-

liable cumulative frequency distribution. Both distribu-

tions follow power-law regressions over approximately

two orders of magnitude, and coefficients and expo-

nents are similar. Extrapolating to an aperture of 1 mm

Table 2. Summary of Observations of Fracture Cement in the Sidewall Cores

Well Depth, ft (m) Obs.* Figure Observation and Interpretation

A-1 9031 ft (2753 m) A 11b, c Dolomite is synkinematic. Low-luminescence dolomite is earliest.

Medium-luminescence dolomite follows. Calcite is postkinematic.

9055 ft (2760 m) B 6c Intersection of two sets of FY fractures; fractures are synchronous.

9055 ft (2760 m) C 11d, e Host rock dolomitized prior to FY growth.

9031 ft (2762 m) D 11a Low-luminescence dolomite is latest cement in a few fractures.

A-6 9116 ft (2779 m) E 4b, c Low-luminescence dolomite is latest cement in a few fractures.

9116 ft (2779 m) F 4 Open fractures in A-6 well. Locally no calcite.

*Obs. = observation.

Figure 10. FY fractures in the unit 3 well sealed with low-luminescence dolomite and showing crack-seal structure (CS).

Gale and Gomez 1017

Figure 11. Examples of LLC dolomitein FY fractures in the A-1 well. (a) LLC(or low-luminescence dolomite) rimming adolomite rhomb fracture bridge, (b) SEM-SEI, and (c) SEM-CL image of a bridgeof synkinematic LLC in a fracture at 9031 ft(2753 m). Calcite cement precipitated inthe remaining fracture porosity. (d) SEM-SEIand (e) SEM-CL image of a fracture at9055 ft (2760 m). Two dolomite crystals(image centers) were cut during initial frac-ture opening and sealed with LLC. Sub-sequent fracture opening allowed a secondphase of midgray luminescence dolomitecement to precipitate in the fracture. Thefracture propagated along internal crystalboundaries, indicating that matrix dolo-mitization predated fracture growth.(f ) SEM-CL image of fractures sealed withLLC at 9031 ft (2752 m), and (g) SEM-CLimage of fracture at 9087 ft (2770 m),with an irregular pathway resulting frompropagation along rhombohedraldolomite-grain boundaries.

1018 Fractures in Ellenburger Dolostones

(0.04 in.), we calculated a single fracture intensity from

the power-law equation for each sample to be able to

compare samples. The fracture intensity for fractures

�1 mm (�0.04 in.) wide lies between 1.8 � 10�2 and

2.0 � 10�2 fractures/mm2 for the A-1 well. Fracture

intensities of samples from the unit 3 well (Figure 14c,

d) are consistent with each other (3.4� 10�2 and 4.0�10�2 fractures/mm2), but are slightly higher than those

in the A-1 well.

DISCUSSION

Fracture Characterization

It is not always possible to assign individual fractures to

the FBR or FY groups if crosscutting relationships are ab-

sent. For example, in the unit 3 core, there are several

sections of relatively unbrecciated rock with a few steep-

ly dipping fractures. Loucks (2003) interpreted these sec-

tions of core as collapsed ceiling slabs, in which fractures

are related to roof stresses, but they could also have re-

sulted from a separate structural event, unrelated to cave

collapse. We infer that fractures in the sidewall cores

belong to the FY groupbecause their orientationsmatch

fractures in the unit 3 well that are demonstrably FY.

The common orientations of microfractures measured

in thin section and macrofractures observed in the im-

age log and sidewall cores support the contention that

FY microfractures and macrofractures are part of the

same population, and fracture intensity predictions are

made on this basis. Correlating fracture sets on the basis

of orientation is not the most robust approach, but we

appeal to the otherwise chaotic nature of FBR fracture

orientations to justify doing so in this study.

North-northeast–south-southwest–trending frac-

tures in the unit 3 image log and northeast-southwest–

trending fractures in theA-1 sidewall cores are interpreted

to be part of the same set. The consistent orientation of

late opening-mode fractures indicates that a common

stress field prevailed during fracture growth. The sedi-

ment fill of FBR fractures does not appear to have acted

as a barrier to fracture propagation, suggesting that the

fill had lithified prior to FY fracturing. The local paleo-

stress field, however, was likely to have been affected by

the presence of mineral-filled vugs or interclast pores,

Figure 12. Graph of the number of fracture sets present(lower X-axis, filled triangles) and the ratio of open fracturesto fracture sets present (upper X-axis, open circles) againstdepth in the (a) A-1 and (b) A-6 wells. Up to three fracture setsexist, and none, one, two, or all of the sets might be filled. Inthe A-1 well, none of the fractures are open except at onedepth. In the A-6 well, there are several depths where all setspresent are open.

Figure 13. Imagery and element mapping of a sample from the A-1 well at 9031 ft (2752 m). All images are of the same part of thesample, and brightness in element maps is proportional to concentration, brighter areas being more concentrated. (a) Secondary electronimage, where porosity is black, dolomite is midgray, and calcite is pale gray; (b) SEM-CL image, which, used together with the secondaryelectron image, shows more than one dolomite. LLC (or low-luminescence dolomite) appears black in this image (c) magnesium map,showing the LLC to be less concentrated in magnesium than the other dolomite and (d) calcium map showing the distribution of calcite.

Gale and Gomez 1019

acting as stress-concentration points for the initiation

of steep opening-mode fractures (Figure 7a).

Present-day maximum horizontal stress (SHmax)

orientation in thePermian Basin is variable (Tingay et al.,

2006). The openness of FY fractures is not likely to be

affected by present-day SHmax, however, because of

substantial syn- and postkinematic cementation in FYfractures and the host rock (cf. Laubach et al., 2004b).

Synthesis of Fracture Timing in the Barnhart FieldEllenburger Group

Combs et al. (2003) published a burial history incor-

porating Loucks’ (2003) paleocave-collapse model for

origin of the breccias, which we use here to frame the

discussion of fracture timing and constrain conditions for

FY fractures (Figure 15). Paleocave collapse and brec-

ciation can occur at a range of depths and times after

karstification, leading to a coalesced collapsed-paleocave

system with several brecciation events (Loucks, 1999).

Loucks (1999) indicated that open caves are present

down to at least 19,000 ft (�5,800m) and Combs et al.

(2003) reported a solution-enlarged opening approxi-

mately 3 ft (1 m) in diameter in the Barnhart field.

Karst development during theOrdovician andPenn-

sylvanian provides an indication of when the Ellen-

burger Group was near the surface (Figure 15, points 1

and 5). Combs et al. (2003) suggested that brecciation

occurred over a range of depths and times on the basis

of the variability of fracture fill in fractures associated

with breccia. Some earlier fractures contain geopetal

structures, with sediment being deposited in the lower

parts of fractures and later baroque dolomite cement

occluding the upper part.We concur that fractures with

sediment fill are likely to have formed during the early

part of cave collapse, when sediment transport was

possible. Other irregular, randomly oriented fractures

Figure 14. Aperture-size distributions for data from SEM imagery (a, b) samples from the A-1 well (9055 and 9087 ft; 2759 and2769 m). Sufficient microfractures of the same fracture set (NE) are present to generate power-law distributions with comparablevalues of fracture intensity. (c, d) Fracture intensity in two samples from the unit 3 well (9070.9 and 9079.6 ft; 2764.8 and 2767.4 m).Intensities are similar. Fracture intensities in the unit 3 well are slightly higher than those in the A-1 well.

1020 Fractures in Ellenburger Dolostones

contain no sediment, only baroque dolomite. These are

also probably FBR fractures and could be associated

with continued mechanical compaction of the system

(Combs et al., 2003). The dominant orientations of FYfractures, however, are consistent in two wells more

than 500m (1640 ft) apart, and on this basis, we argue

that these fractures are related to a regional structural

event. A consistently oriented stress field and, therefore,

fracture orientation would be unlikely during mechan-

ical compaction of caves. Moreover, FY fractures con-

tain synkinematic, nonbaroque dolomite, indicating

that ambient conditions during FY were different from

those prevailing during FBR fracturing.

Kupecz and Land (1991) identified early (pre-

MiddleOrdovician) and late (post-MiddleOrdovician)

phases of dolomitization. Early dolomitization pre-

dates all karsting, whereas late dolomitization predates

only karsting associated with the Pennsylvanian surface-

exposure event. Oxygen stable-isotope data from late

dolomite, which accounts for 10% of total dolomite,

gave temperatures of 60–110jC (Kupecz and Land,

1991). Late-stage dolomite was related partly to karst

breccias by Kupecz and Land (1991). These researchers

also recognized baroque dolomite cements that post-

dated the late matrix dolomite. They linked the late-

stage dolomite and baroque dolomite cements to evolv-

ing reactive Pennsylvanian pore fluids derived from

basinal siliciclastics, and fluid migration was associated

with the Ouachita orogeny.

Combs et al. (2003) also related baroque dolomite

cementation to the Pennsylvanian Ouachita orogeny

(Figure 15, point 4). If baroque dolomite was indeed

related to hydrothermal fluids permeating the system

during the Ouachita uplift, then the possibilities for

later events that could have caused FY opening-mode

fractures are partly constrained. Combs et al. (2003)

suggested that a second period of karstification occurred

during the Pennsylvanian. In the Barnhart field, how-

ever, there is no unequivocal evidence of karst or karst

collapse postdating baroque dolomite.

Dolomite cement in FY fractures postdates all ce-

ments described by Kupecz and Land (1991) and com-

monly shows crack-seal texture, which is evidence that

the cement was precipitating at the time of fracturing,

and that precipitation was keeping pace with fracture

opening (Figure 10). This is an important observation

for constraining conditions of FY fracture growth. Do-

lomite precipitation is thought to occur only at elevated

temperatures (>50jC, Machel, 2004) or under the

influence of sulfate-reducing bacteria (Wright and

Wacey, 2004). We have no evidence that dolomite pre-

cipitation was microbe assisted. FY fracturing, there-

fore, is likely to have occurred while the Ellenburger

Group was buried at a depth corresponding to a tem-

perature of at least 50jC, orwhile fluids at or above this

temperature were passing through the fractures. In the

former case, a depth of 2 km (6562 ft), assuming a geo-

thermal gradient of 25jC/km, would be sufficient for

Figure 15. Burial history andparagenesis synthesis plot fromCombs et al. (2003, reprintedwith permission from the WestTexas Geological Society). Pos-sible FY fracture timing is in-dicated by points A and B. Point Arepresents onset of fracturingon the downward part of theburial curve, and point B repre-sents fracturing associated withhydrothermal fluids near theend of the Ouachita orogeny.

Gale and Gomez 1021

dolomite precipitation. Late fractures would then have

formed on the downward burial curve after a period

during the late Pennsylvanian, when these rocks were

near the surface (Figure 15, FY fractures initiating at

point A). Alternatively, fracturing could be driven by

late pulses of hydrothermal fluid, out of thermal equi-

librium with the host rock on the uplift path, during

late Pennsylvanian exhumation as overburden stress

was reduced. These fluids would have been too cool to

allow baroque dolomite to precipitate, but hot enough

to allow synkinematic ordered dolomite precipitation.

In this case, FY opening-mode fractures would form

during the later stages of theOuachita uplift (Figure 15,

point B).

A study of the stable isotope signatures of the syn-

kinematic cements may distinguish between these pos-

sibilities. If synkinematic cements share common fea-

tures with the baroque dolomite, they are more likely

to be derived from late hydrothermal fluids. If the ce-

ment signatures aremore like dolomite in the host rock,

they aremore likely to be derived from local dissolution

during burial, although in this case, local sources would

also include baroque dolomite and could be difficult to

distinguish.

The orientation of FY fractures could be linked to

tectonic events, but care is needed. The east-southeast

set might be compatible with Ouachita orogenic com-

pression if FY fractures formed during the later phases

of the orogeny (point B, Figure 15). An association for

the synchronous south-southwest set is less clear, but

the orthogonal sets could reflect low stress anisotropy

at the time of fracturing. This might be more likely

during the burial phase (after point A, Figure 15).

CONCLUSIONS

Two groups of fractures are present in the Ellenburger

Group in the Barnhart field. The oldest fractures (FBR)

have irregular shapes, sediment, and baroque dolomite

fill, no preferred orientation, and result from one or

more brecciation events associated with karsting. Frac-

turing associated with brecciation is highly heteroge-

neous and probably occurred over a considerable time

span, and the origin of individual fractures remains

speculative. The second group (FY) consists of two sets

of younger, subvertical, opening-mode fractures having

consistent east-southeast and south-southwest strikes.

These fractures postdate brecciation and baroque dolo-

mite cementation. Fracture (FY) apertures range from

several micrometers to a few millimeters. FY fracture

intensity is relatively low and consistent within each

well, and there is only a small variation between wells.

The intensity for fractures �1 mm (�0.04 in.) wide

lies between 1.8� 10�2 and 2� 10�2 fractures/mm2

(A-1 well) and 3.4 � 10�2 and 4.0 � 10�2 fractures/

mm2 (unit 3 well).

Low-luminescent, dolomite cement seals narrow

FY fractures and form linings and mineral bridges in

large FY fractures. Other, more luminescent dolomite

cements overgrow the low-luminescent dolomite. The

dolomite cements are synkinematic with FY fractures.

Calcite postdates the dolomite cements and FY fracture

opening and is present in two wells (unit 3 and A-1),

where it occludes remaining FY fracture porosity. Frac-

tures are locally open in a third well (A-6) where cal-

cite is absent.

Diagnosing the presence of postkinematic calcite

is an important step in being able to predict open frac-

tures and, thereby, reservoir permeability and can be

done without sampling large fractures using sidewall

cores. Analysis of sidewall cores also provided a con-

sistent FY fracture orientation in the early stages of the

study, which was later verified by image log analysis.

REFERENCES CITED

Amthor, J. E., and G. M. Friedman, 1991, Dolomite-rock texturesand secondary porosity development in Ellenburger Group car-bonates (Lower Ordovician) west Texas and southern NewMexico: Sedimentology, v. 38, p. 343–362.

Candelaria, M. P., and C. L. Reed, eds., 1992, Paleokarst, karstrelated diagenesis and reservoir development: Examples fromOrdovician–Devonian age strata of west Texas and the mid-continent: Permian Basin SEPM Publication 92-33, 202 p.

Combs, D. M., R. G. Loucks, and S. C. Ruppel, 2003, Lower Ordo-vician Ellenburger Group collapsed paleocave facies and asso-ciated pore network in the Barnhart field, Texas, in T. J. Huntand P. H. Luftholm, eds., The Permian Basin: Back to basics:West Texas Geological Society Symposium: Proceedings, WestTexas Geological Society Publication 03-112, p. 397–418.

Cotton, C. B., 1966, Barnhart field, in Oil and gas fields in westTexas: Symposium: West Texas Geological Society Publica-tion 66-52, p. 38–41.

Dutton, S. P., E.M. Kim, R. F. Broadhead, C. L. Breton,W.D. Raatz,S. C. Ruppel, and C. Kerans, 2005, Play analysis and digital port-folio of major oil reservoirs in the Permian Basin: University ofTexas at Austin, Bureau of Economic Geology Report of Inves-tigations no. 271, 287 p., CD-ROM.

Gale, J. F. W., 2004, Self-organization of natural mode-I fractureapertures into power-law distributions: GulfRocks 2004. 6thNorth American RockMechanics Symposium (NARMS): RockMechanics Acron Borders and Disciplines, held in Houston,Texas, June 5–9, 2004, AmericanRockMechanics Association/NARMS 04-563, 13 p.

Gale, J. F. W., S. E. Laubach, R. A. Marrett, J. E. Olson, J. Holder,and R. M. Reed, 2004, Predicting and characterizing fracturesin dolostone reservoirs: Using the link between diagenesis and

1022 Fractures in Ellenburger Dolostones

fracturing, in C. J. R. Braithwaite, G. Rizzi, and G. Darke, eds.,The geometry and petrogenesis of dolomite hydrocarbon res-ervoirs: Geological Society (London) Special Publication 235,p. 177–192.

Gomez, L. A., and S. E. Laubach, 2006, Rapid digital quantificationof microfracture populations: Journal of Structural Geology,v. 28, p. 408–420.

Gomez, L. A., J. F. W. Stowell, S. C. Ruppel, and S. E. Laubach,2001, Fracture characterization using rotary-drilled sidewallcores: An example from the Ellenburger Formation, west Texas,in Proceedings, West Texas Geological Society fall meeting:West Texas Geological Society Publication 01-110, p. 81–89.

Hammes, U., F. J. Lucia, and C. Kerans, 1996, Reservoir heteroge-neity in karst-related reservoirs: Lower Ordovician EllenburgerGroup,west Texas, inE. L. Stoudt, ed., Precambrian–Devoniangeology of the Franklin Mountains, west Texas— Analogs forexploration and production in Ordovician and Silurian karstedreservoirs in the Permian Basin: West Texas Geological Society1996 annual field trip guidebook, West Texas Geological So-ciety Publication 96-100, p. 99–115.

Holtz, M. H., and C. Kerans, 1992, Characterization and classi-fication of west Texas Ellenburger reservoirs, in M. P. Cande-laria andC. L. Reed, eds., Paleokarst, karst related diagenesis andreservoir development: Examples from Ordovician–Devonianage strata of west Texas and the midcontinent: Permian BasinSEPM Publication 92-33, p. 45–54.

Kerans, C., 1990, Depositional systems and karst geology of theEllenburger Group (Lower Ordovician), subsurface, west Texas:University of Texas at Austin, Bureau of Economic GeologyReport of Investigations no. 193, 63 p.

Kupecz, J. A., and L. S. Land, 1991, Late-stage dolomitization ofthe Lower Ordovician Ellenburger Group, west Texas: Journalof Sedimentary Petrology, v. 61, p. 551–574.

Laubach, S. E., 1997, A method to detect natural fracture strike insandstones: AAPG Bulletin, v. 81, p. 604–623.

Laubach, S. E., 2003, Practical approaches to identifying sealed andopen fractures: AAPG Bulletin, v. 87, no. 4, p. 561–579.

Laubach, S. E., and E. Doherty, 1999, Oriented drilled sidewallcores for natural fracture evaluation: Society of Petroleum En-gineers, SPE Paper 56801, Formation Evaluation and ReservoirGeology, v. omega, p. 793–800.

Laubach, S. E., R. M. Reed, J. E. Olson, R. H. Lander, and L. M.Bonnell, 2004a, Coevolution of crack-seal texture and fractureporosity in sedimentary rocks: Cathodoluminescence observa-tions of regional fractures: Journal of Structural Geology, v. 26,no. 5, p. 967–982.

Laubach, S. E., J. E. Olson, and J. F. W. Gale, 2004b, Are openfractures necessarily aligned with maximum horizontal stress?:Earth and Planetary Science Letters, v. 222, no. 1, p. 191–195.

Lee, Y. E., and G. M. Friedman, 1987, Deep-burial dolomitizationin the Ordovician Ellenburger Group carbonates, west Texasand southeastern New Mexico: Journal of Sedimentary Pe-trology, v. 57, no. 3, p. 544–557.

Loucks, R. G., 1999, Paleocave carbonate reservoirs: Origins, burial-depth modifications, spatial complexity, and reservoir implica-tions: AAPG Bulletin, v. 83, no. 11, p. 1795–1834.

Loucks, R. G., 2003, Understanding the development of brecciasand fractures in Ordovician carbonate reservoirs, in T. J. Huntand P. H. Luftholm, eds., The Permian Basin: Back to basics:West Texas Geological Society Symposium: Proceedings, WestTexas Geological Society Publication 03-112, p. 231–252.

Loucks, R. G., and P. Mescher, 2002, Coalesced-collapsed palaeo-cave systems: Origins, spatial complexity and reservoir impli-cations: AAPG/SEPM Annual Meeting, Houston, Texas, So-ciety for Sedimentary Geology field trip guidebook, 81 p.

Loucks, R. G., P. K. Mescher, and G. A. McMechan, 2004, Three-dimensional architecture of a coalesced, collapsed-paleocavesystem in the Lower Ordovician Ellenburger Group, centralTexas: AAPG Bulletin, v. 88, no. 5, p. 545–564.

Machel, H. G., 2004, Concepts and models of dolomitization: Acritical reappraisal, in C. J. R. Braithwaite, G. Rizzi, and G.Darke, eds., The geometry and petrogenesis of dolomite hy-drocarbon reservoirs: Geological Society (London) Special Pub-lication 235, p. 7–64.

Marrett, R., O. Ortega, and C. Kelsey, 1999, Extent of power-lawscaling for natural fractures in rock: Geology, v. 27, no. 9,p. 799–802.

Olson, J. E., S. E. Laubach, and R. H. Lander, 2004, Improvingfracture permeability prediction by combining geomechanicsand diagenesis: GulfRocks 2004. 6th North American RockMechanics Symposium (NARMS): Rock Mechanics AcronBorders and Disciplines, held in Houston, Texas, June 5–9,2004, American RockMechanics Association/NARMS04-563,9 p.

Ramsay, J. G., 1980, The crack-seal mechanism of rock deforma-tion: Nature, v. 284, p. 135–139.

Reed, R. M., and K. L. Milliken, 2003, How to overcome imagingproblems associated with carbonate minerals on SEM-basedcathodoluminescence systems: Journal of Sedimentary Research,v. 73, no. 2, p. 328–332.

Ross, R. J., 1976, Ordovician sedimentation in the western UnitedStates, in M. G. Bassett, ed., The Ordovician system: Proceed-ings of a Paleontological Association Symposium, Birmingham,p. 73–105.

Tingay, M. R. P., B. Muller, J. Reinecker, and O. Heidbach, 2006,State and origin of the present-day stress field in sedimen-tary basins: New results from the world stress map project,in Proceedings, 41st U.S. Symposium on Rock Mechanics:50 Years of Rock Mechanics— Landmarks and Future Chal-lenges: Golden, Colorado, American Rock Mechanics Associ-ation/USRMS 06-1049, 14 p., CD-ROM.

Wright, D. T., and D. Wacey, 2004, Sedimentary dolomite: Areality check, in C. J. R. Braithwaite, G. Rizzi, and G. Darke,eds., The geometry and petrogenesis of dolomite hydrocar-bon reservoirs: Geological Society (London) Special Publica-tions 235, p. 65–74.

Gale and Gomez 1023