Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

download Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

of 7

Transcript of Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    1/16

    AAPG Bulletin, v. 86, no. 12 (December 2002), pp. 2089–2104   2089

    Estimating fracture traceintensity, density, and meanlength using circular scan lines

    and windowsM. B. Rohrbaugh Jr., W. M. Dunne, and M. Mauldon

    A B S T R A C T

    Fracture characterization protocols that reduce sampling bias are

    likely to yield higher quality input for exploration and development 

    decisions when dealing with naturally fractured reservoirs. A new

    set of estimators for fracture density, intensity, and mean tracelength corrects for sampling biases and provides a useful integrated

    description for bulk aspects of a fracture network. These estimators

    are based on counts of intersections between fracture traces and

    circular scan lines and of trace terminations in circular windows.

    Application to synthetic fracture patterns with known parameters

    validates the use of the new estimators, which are then applied to

    natural fault trace maps from seismic volumes and joint trace maps

    from rock pavements. The new estimators are distribution inde-

    pendent and eliminate the effects of orientation, censoring, and

    length biases, which limit the effectiveness of other sampling tech-

    niques. Estimator accuracy improves as sample size increases, par-ticularly for larger circles that exceed a fracture-defined block size.

    Estimator accuracy for mean trace length improves when the sam-

    ple exceeds threshold count values for fracture terminations based

    on guidance from the analysis of similar synthetic patterns. These

    new estimators also provide both inputs and independent checks of 

    predictions for fracture-generator programs used to model fracture

    populations in a rock volume.

    I N T R O D U C T I O N

    Quantification of fracture parameters such as density, size, and in-

    tensity aids in the assessment of hydrocarbon flow and storage in

    fractured reservoirs (Reiss, 1982; Nelson, 1985; Dershowitz and

    LaPointe, 1994; Narr, 1996) but is complicated by the difficulty of 

    deciding on the best approach for incorporating fractures into a

    reservoir model. Problems arise from the lack of consensus as to

    Copyright2002. The American Association of Petroleum Geologists. All rights reserved.

    Manuscript received January 24, 2000; revised manuscript received January 10, 2001; final acceptance

     June 15, 2002.

    A U T H O R S

    M. B. Rohrbaugh Jr. Tennessee Department of Environment and Conservation, Division of Underground Storage Tanks, 540 McCallie Avenue, Suite

     550, Chattanooga, Tennessee, 37402–2013

    M. Bruce Rohrbaugh Jr. received his B.S.degree in geology from West VirginiaUniversity in 1997 and his M.S. degree instructural geology from the University of Tennessee, Knoxville in 2000. His researchinterests include hydrogeology, application of computers to solving geologic problems, andstructural geology. He is currently employedas a geologist with the Tennessee Departmentof Environment and Conservation.

     W. M. Dunne  Department of Geological Sciences, 306 G&G Building, University of Tennessee, Knoxville, Tennessee, 37996-1410; [email protected]

     William M. Dunne, although born in theUnited States, received his B.S. degree andPh.D. in geology from the University of Bristol,England. He joined the Department of Geological Sciences at the University of Tennessee in 1988 as an associate professorand is now is a professor and departmenthead. His research interests include fracture

    characterization particularly in younger rocks,deformation in thrust belts from large to smallscale, and deformation analysis of sedimentary rocks.

    M. Mauldon  Department of Civil and  Environmental Engineering, Virginia Tech, 200 Patton Hall, Mail Code 0105, Blacksburg,Virginia, 24061; [email protected]

    Matthew Mauldon, although born in England,has geology (B.A.) and civil engineering (M.S.)degrees and a Ph.D. in civil engineering from

     the University of California at Berkeley. Hespent eight years on the faculty at theUniversity of Tennessee, where hecollaborated with Dunne and Rohrbaugh.Mauldon is now an associate professor in the Via Department of Civil and EnvironmentalEngineering at Virginia Tech, where he teaches and conducts research in the areas of rock mechanics, engineering geology, andgeotechnical engineering.

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    2/16

    2090   Fracture Characterization Using Circular Scanlines and Windows

    which parameters to quantify, the difficulties in measuring the pa-

    rameters, and intrinsic sampling biases. To address these problems,

    we offer a set of estimators for fracture parameters based on the

    use of circular scan lines and windows where a scan line is the pe-

    rimeter and the window is the interior of a circle. These estimators

    reduce sampling biases and provide a useful integrated description

    of bulk aspects of a fracture network. Use of these estimators re-

    quires only counts of the number (n) of fracture traces intersectingthe circumference and/or counts of the number (m) of fracture

    traces terminating in the circle interior. Performance of the new

    circular scan line/window estimators is evaluated for both synthetic

    and natural fracture patterns to demonstrate that estimates con-

    verge on true values for a fracture trace population, to demonstrate

    that the new estimators outperform or match existing estimators,

    and to discuss issues of estimator performance.

    Fracture Parameters

    A redundant, commonly mutually inconsistent vocabulary exists todescribe the amount of fracturing in a rock. Rather than reviewing

    this terminology, we define three key parameters of a fracture pat-

    tern: density, size, and intensity (Table 1).

    Density

    Fracture density is commonly treated as the number of observed

    isolated fractures or fracture segments per unit length, area, or vol-

    ume (Dershowitz and Herda, 1992; Ghosh and Daemen, 1993).

    This is a scale-dependent quantity that we call “apparent density.”

    Fracture density is defined in this article as the number of fractures

    per unit length, area, or volume, enumerated in terms of uniquepoints, such as fracture centers (Mauldon, 1998; Mauldon and Der-

    showitz, 2000). Apparent density overestimates density (Kulatilake

    and Wu, 1984; Mauldon et al., 2001), and the magnitude of this

    overestimation increases as sample size decreases (Figure 1B). For

    example, the apparent density (number of visible traces divided by

    circle area) of fractures of Set 1 in Figure 1A is 0.0014 per m 2 for

    a circular window of radius 75 m and increases to 0.0020 per m2

    for a smaller window of radius 25 m. If trace centers (dots in Figure

    1A) are counted to estimate true density for Set 1, the estimates

    (count divided by circle area) are 0.0010 per m2 for both the circle

    of radius 25 m and the circle of radius 75 m. In practical applica-tions, one half the number of fracture trace terminations is used as

    an unbiased estimate of the number of trace centers, because cen-

    ters cannot be identified unless both ends of a trace are visible

    within the window (Mauldon, 1998; Mauldon et al., 2001).

    Size

    Fracture size is defined in one, two, or three dimensions, as frac-

    ture trace length, area, and volume, respectively. For composite

    connected fractures, an investigator should decide whether to

    characterize individual segments or the entire composite. Typically

    A C K N O W L E D G E M E N T S

     Acknowledgment is made to the donors of The Petroleum Research Fund, administeredby the American Chemical Society, for partialsupport of this research. Additionally, theGeological Society of America and the South-eastern Section of the Geological Society of  America are thanked for their partial supportof the field work. We would also like to thank M. F. Schaeffer for permission to use the frac- ture trace map from Rocky Creek, South Caro-lina, and Camilo Montes, Yen-Yit Chan, You Li,and Jim Calcagno for programming assistance.Steve Laubach, John Lorenz, and Bill Der-showitz are thanked for their insightful andconstructive reviews.

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    3/16

    Rohrbaugh et al.   2091

    Table 1.  Vocabulary for Fracture Parameters*

    Parameter Definition Estimator

    Density Linear Number of fractures per unit length    m /2p r 2 q̂

     Areal (q) Number of fractures per unit area

     Volumetric Number of fractures per unit volume (in all cases count is conducted using

    unique points such as fracture centers or half of the fracture terminations)

    Size Linear (l) Mean fracture trace length   (p r  /2)( n / m) l̂

     Areal Mean fracture area

     Volumetric Mean fracture volume

    Intensity Linear Number of fracture per unit length (L0 /L1 L1)   Î  n /4 r 

     Areal (I) Fracture length per unit area (L1 /L2 L1)

     Volumetric Fracture area per unit volume (L2 /L3 L1)

    *Where L is a dimension of length and  r  is radius. See Figure 3 for illustration of  m  and  n.

    Figure 1.  Problems due to censoring and length bias whensampling fracture traces. (A) Fracture pattern with two sets sam-pled by three progressively larger circles (dots  trace centers;

     r  radius). (B) Decreasing density overestimates for increasingsample (circle) size. (C) Decreasing mean trace-length under-estimates for increasing sample size. (D) Undersampling of joint trace lengths due to censoring as shown by the probability dis- tribution function (pdf). (E) Oversampling of longer traces due to length bias as shown by a pdf.

    for fracture studies on exposed surfaces, trace lengthor aperture is measured because fracture areas and

    volumes are commonly not directly measurable (Der-

    showitz and Herda, 1992; Marrett et al., 1999; Ortegaand Marrett, 2000). For the two-dimensional fracture

    patterns discussed in this article, we use mean trace

    length as our size parameter (Table 1; Figure 1).

    Intensity

    Fracture intensity is a pattern characteristic that in-

    corporates both density and size (Dershowitz and

    Herda, 1992; Mauldon and Dershowitz, 2000). Inten-

    sity is defined as number of fractures per unit sample

    length, fracture length per unit surface area, or frac-

    ture area per unit rock volume, in one, two, or threedimensions, respectively (Table 1). Consequently, in-

    tensity has the same dimensions whether calculated

    linearly, areally, or volumetrically. In this article, we

    examine two-dimensional samples and, therefore, use

    the areal intensity: fracture length per unit area.

    Current Measurement Methods

    Two common sampling methods are used for esti-

    mating fracture parameters: straight scan lines and

    areal sampling (Figure 2) (LaPointe and Hudson,1985; Priest, 1993; Wu and Pollard, 1995, Becker and

    Gross, 1996; Marrett et al., 1999; Ortega and Marrett,

    2000). Straight scan lines sample the fractures they

    intersect and are used to systematically record fracture

    characteristics, such as number, orientation, aperture,

    and so on. (Priest and Hudson, 1981). Areal sampling

    involves mapping the fracture trace pattern and re-

    cording desired fracture characteristics at locations in

    the map area (e.g., Priest, 1993; Wu and Pollard,

    1995).

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    4/16

    2092   Fracture Characterization Using Circular Scanlines and Windows

    Figure 2.  Circular scan line/window (dotted circle), areal (ir-regular white window), and straight scan line (dotted line) sam-pling of a fracture trace population. Solid lines represent visiblefracture traces, and dashed lines represent covered fracture traces.

    Sampling Biases

    Orientation bias occurs where scan lines or the long

    axes of inequant sampling areas are not perpendicular

    to a fracture set. In both cases, intensity is underesti-

    mated (Terzaghi, 1965; Priest, 1993; Mauldon and

    Mauldon, 1997). Scan line estimates are corrected by

    dividing by the cosine of the angle between the scan

    line and the normal to the fracture set (Terzaghi, 1965;Peacock et al., in press). However, as this angle ap-

    proaches 90, the cosine approaches zero, and cor-

    rected estimates approach infinity, significantly over-

    estimating intensity (Priest, 1993). No procedures are

    currently available for correcting orientation bias from

    inequant sampling areas.

    Censored fracture traces extend beyond the ex-

    posure or seismic coverage, so that one or both ends

    are not visible (Figure 1A) (e.g., Cruden, 1977;

    Baecher and Lanney, 1978; Einstein and Baecher,

    1983; Kulatilake and Wu, 1984; LaPointe and Hudson,1985; Pickering et al., 1995; Mauldon, 1998; Marrett 

    et al., 1999). Such traces are referred to as singly or

    doubly censored, respectively. Using censored traces to

    estimate density and size directly, rather than using es-

    timators such as those in Table 1, leads to overesti-

    mates of density and underestimates of size (Figure 1).

    For example, a count of all visible trace segments in a

    circle of radius 25 m (4 segments vs. 2 centers) of Set 

    1 (Figure 1A) overestimates density by a factor of two

    (0.002 vs. 0.001 per m2). Similarly, using the censored

    lengths, the estimate of mean trace length for Set 1

    (total visible trace length divided by number of traces)

    is 21.3 m, in contrast to the true mean trace length of 

    40 m. In both cases, increasing the size of the sampling

    circle relative to fracture size decreases censoring and,

    consequently, error. For example, increasing circle ra-

    dius from 25 to 75 m (Figure 1A) decreases apparent 

    density from 0.0020 to 0.0014 per m2 (true density0.001 per m2) and increases apparent mean trace

    length from 21.3 to 27.4 m (true value 40 m).

    Length bias occurs because longer fracture traces

    have a greater probability of being sampled than

    shorter traces (Baecher and Lanney, 1978; Einstein and

    Baecher, 1983; LaPointe and Hudson, 1985; Mauldon

    1998). Consequently, mean trace-length estimates and

    trace-length distributions are skewed toward longer

    fractures (Figure 1E). For example in Figure 1A, Sets

    1 and 2 have the same fracture density, but Set 1 has

    longer traces. As a result, the largest window (75 mradius) samples 24 of the longer Set 1 fractures vs. 17

    of Set 2. If these data are taken at face value,

    mean trace length for the whole pattern will be

    overestimated.

    Pattern heterogeneity refers to a change in fracture

    parameters with a change in position. This sampling

    bias can occur, for example, when sampling rock vol-

    umes with localized fracture development near faults.

    Heterogeneity is partly a function of scale and may be

    especially pronounced for borehole sampling, which

    provides limited samples of subsurface fracture net-works. Use of multiple subdomains with homogeneous

    parameters minimizes the effects of heterogeneity

    (Turner and Weiss, 1963; Whitten, 1966; LaPointe

    and Hudson, 1985; Kulatilake et al., 1997). Alterna-

    tively, geostatistical methods may be employed to de-

    termine the magnitude and rate of change of a param-

    eter as a function of distance and direction (LaPointe

    and Hudson, 1985; LaPointe, 1993; Priest, 1993; Jian

    et al., 1996).

    Advantages/Disadvantages of Current Estimation Methods

    Straight scan lines (Figure 2) provide rapid estimates

    of fracture intensity (Priest and Hudson, 1981; La-

    Pointe and Hudson, 1985; Becker and Gross, 1996).

    Unprocessed straight scan line data, however, are sub-

     ject to orientation bias, length bias, censoring, and pat-

    tern heterogeneity (Terzaghi, 1965; Baecher and Lan-

    ney, 1978; Priest and Hudson, 1981; LaPointe and

    Hudson, 1985; Priest, 1993; Mauldon and Mauldon,

    1997; Peacock et al., in press).

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    5/16

    Rohrbaugh et al.   2093

    Figure 3.  Fracture trace pattern with sampling circle. (A) Soliddots are intersection points ( n) between fractures and circle. (B)Triangles are fracture endpoints ( m) in the circular windows.

    Areal sampling (Figure 2) reduces censoring and

    length bias as compared to scan line sampling but (1)

    is subject to orientation bias in the plane for anything

    except a circular sampling area, (2) is more time con-

    suming than use of scan lines, (3) may disguise pattern

    heterogeneities, and (4) may introduce censoring and/

    or length biases for inequant or small sampling areas

    (LaPointe and Hudson, 1985; Kulatilake et al., 1993; Wu and Pollard, 1995; Kulatilake et al., 1997).

    N E W E S T I M A T O R S

    Because neither areal sampling nor straight scan lines

    are completely satisfactory in terms of efficiency, ac-

    curacy, and lack of bias, this article examines the use

    of circular scan lines and windows for characterizing

    intensity, density, and mean trace length (Figure 2).

    Data from circular scan lines and windows are appliedto estimators (Table 1) that do not require knowledge

    of fracture spacing, trace length, or orientation distri-

    butions and are, therefore, distribution independent 

    (Mauldon, 1998; Mauldon et al., 2001).

    E S T I M A T O R P E R F O R M A N C E F O R  S Y N T H E T I C F R A C T U R E P A T T E R N S

    The ability of the new estimators to correct for sam-

    pling biases was investigated by comparing estimatesof intensity, density, and mean trace length to known

    values for synthetic fracture trace patterns. A new

    computer program called JAWS (Joint Analysis using

     Windows and Scanlines) (Rohrbaugh, 2000), was used

    to generate and sample traces with uniformly distrib-

    uted centers in a square region. These traces were sam-

    pled with 100 circles of known radius placed randomly

    and independently in a smaller square analysis region

    centered on the generation region so as to avoid edge

    effects (Gilmour et al., 1986). Trace-circle intersec-

    tions and trace terminations inside circular windowswere counted (Figure 3), and counts were input into

    estimators (Table 1) to yield comparison values.

    For example, two synthetic fracture sets with ori-

    entations of 60 30 and 150 30 were deployed

    to produce orientation bias when sampled, with frac-

    ture lengths of 50 10 L (L is an arbitrary unit of 

    length) and 100 10 L that exceed circle radius of 20

    L by factors of 2.5 and 5, respectively, to create a cen-

    soring bias, and with a difference in length between

    sets of a factor of two to yield a length bias. Running

    means for the estimates (Figure 4) converge on the in-

    put values for intensity, density, and mean trace length

    and correspond closely to them after about 40 samples,whereas individual estimates fluctuate about the mean.

    These results show that the new estimators deal suc-

    cessfully with orientation, censoring, and length biases.

    Similar results were found for all other synthetic cases.

    Comparison to Other Estimators

    Having established that the new estimators overcome

    sampling biases, their accuracy was compared to that 

    from straight scan lines and/or areal sampling of a syn-

    thetic pattern. To achieve equivalent sampling com-parison, estimates were obtained using 100 circular

    scan lines/windows of radius 10 L , 100 straight scan

    lines of the same length (2p (10 L)    63 L), and

    areal samples consisting of 100 circle interiors.

    For example, a fracture set with orientation of 060

    30, mean trace length of 40 20 L, intensity of 

    1.95 L/L2, and density of 0.049 per L2 was generated

    (Figure 5A). Intensity estimates from both circular

    scan lines and areal samples accurately estimate inten-

    sity, but map construction for areal samples is likely to

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    6/16

    2094   Fracture Characterization Using Circular Scanlines and Windows

    Figure 4.  Estimates (jagged black lines) and running meansof estimates (small open circles) for intensity, density, and mean trace length of the synthetic pattern with dashed lines and num-bers for control values, showing that estimators successfully deal with sampling bias issues.

    be much more time intensive than is deployment of 

    circular scan lines. In contrast, running means for the

    straight scan lines underestimate intensity by a factor

    of 1.7 with the Terzaghi correction applied to the

    mean set orientation (Figure 5B).The running mean density estimate from the cir-

    cular window estimator corresponds closely to the con-

    trol value of 0.049 per L2, whereas the same size areal

    samples overestimate density by a factor of 3.5 if trace

    segments are counted (Figure 5C). Running means for

    circular window estimates of mean trace length also

    correspond to the control value, whereas the same size

    areal samples tend to yield estimates of a little less than

    4 L, which is an underestimate by a factor of 11 as

    compared to the input value of 40 L (Figure 5D).

    Based on these results, the use of the estimators based

    on circular scan lines and windows is recommended.

    E S T I M A T O R P E R F O R M A N C E W I T HN A T U R A L F R A C T U R E P A T T E R N S

    Given the success of the new estimators with syntheticfracture patterns, we investigated their applicability to

    natural patterns. This comparison used new and exist-

    ing trace maps of fault and/or joint patterns that range

    from single sets of parallel fractures to multiset pat-

    terns to near polygonal patterns (Table 2; Figure 6).

    This geometric variety was selected so as to provide a

    thorough test of the estimator equations. Also, fault 

    trace maps from seismic reflection data (Figure 6C, D;

    Table 2) were included to demonstrate that circular

    scan lines may be used effectively with this common

    industry data source. Trace maps were digitized andimported into JAWS for analysis. After specifying the

    number and size of the sample circles, JAWS randomly

    distributed them in the sample area and only retained

    intersection and termination counts for circles that 

    were completely contained within a map (Stoyan et al.,

    1995). The new circle-based estimates for the natural

    data sets were compared to estimates from areal sam-

    ples of the maps.

    Circle-based intensity estimates mostly match

    areal estimates for both fault and joint trace patterns

    (e.g., Figures 6; 7C, D), as would be expected from thesimilar correspondence that was found during the anal-

    ysis of synthetic fracture patterns (e.g., Figure 5B). The

    circle-based density and mean trace-length estimates

    mostly matched the apparent density and mean trace-

    length estimates where fracture/fault traces are small

    relative to window size, so that censoring is minimal

    (e.g., Figures 6C, D; 7E, F). In contrast, at Llantwit 

    Major, 86% of the master joint traces are censored, and

    direct density and mean trace-length estimates from

    areal samples differ by a factor of 4 from the circle-

    based estimates (density: 1.5 vs. 0.35 per m2

    ; meantrace length: 2.2 vs. 8.7 m)(Figures 6A; 7A, B). Having

    already tested and demonstrated the performance of 

    the circle-based estimators using synthetic traces, we

    interpret these circle-based estimates at Llantwit Ma-

     jor as being representative of true population charac-

    teristics. It follows that the direct areal estimates for

    the characteristics of these greatly censored traces are

    off by a factor of 4. Again, the circle-based estimators

    yield superior results to previous estimators where the

    potential for sampling biases exists.

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    7/16

    Rohrbaugh et al.   2095

    Figure 5.  Estimator performances for a synthetic fracture trace pattern. (A) Part of a synthetic fracture pattern used to test relativeperformance of new and old estimators. Pattern characteristics are described in text. (B) Intensity estimates from circular scan lines(thick black line), circular areas (light gray line), and straight scan lines (thin line). Control value is 1.95 L/L 2. (C) Density estimatesfrom circular windows (thick black line) and circular areal sampling of apparent density (light gray line). Control value is 0.049 perL2. (D) Mean trace length estimates from circular windows (thick black line) and circular areal sampling of apparent trace length(light gray line). Control value is 40.0 L. Open circles in (B), (C), and (D) are the running means for appropriate estimators.

    D I S C U S S I O N O F E S T I M A T O R  P E R F O R M A N C E

    Effects of Block Size on Estimates

    Block size in two dimensions is the unfractured area

    bounded by fracture traces from two or more sets

    (LaPointe, 1988). Block size depends on the fracture

    spacing or frequency and could be an issue for the new

    estimators for circles smaller than block size. Block size

    effects on the intensity estimator were investigated at 

    Llantwit Major and the light gray region at Amroth

    (Figure 6; Table 2). Block size at Llantwit Major is

    about 0.7–1 m by 0.3–0.5 m, whereas at Amroth the

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    8/16

    2096   Fracture Characterization Using Circular Scanlines and Windows

           T     a        b        l     e

           2  .     G    e    o     l    o    g    y    o     f     F    r    a    c     t    u    r    e     T    r    a    c    e     P    o    p    u     l    a     t     i    o    n    s

         N    a    m    e ,

         L    o    c    a     t     i    o    n ,

         F     i    g    u    r    e     N    o .

         R    o    c

         k     A    g    e    a    n     d

         F    o    r    m

        a     t     i    o    n     N    a    m    e

         S    o    u    r    c    e ,

         T    r    u    n    c    a     t     i    o    n

         L     i    m     i     t ,

         M    a    p    p     i    n    g     S    c    a     l    e ,

         S    u    r     f    a    c    e     A    r    e    a

         S    e     t    s

         P    a     t     t    e    r    n     G    e    o    m    e     t    r    y

         O    r     i    g     i    n

         L     l    a    n     t    w     i     t     M    a     j     o    r ,     W    a     l    e    s ,

         U    n     i     t    e     d     K     i    n    g     d    o    m ,

         S     S     9     5     7     5     6     7     5     4     *

         (     F     i    g    u    r    e     6     A     )

         E    a    r     l    y     J    u    r    a

        s    s     i    c ,

         P    o    r     t     h     k    e    r    r    y

         F    o    r    m    a     t     i    o    n     (     l     i    m    e    s     t    o    n    e     )

         T     h     i    s    s     t    u     d    y ,

         2     0    c    m ,

         1    :     2     5 ,

         4     9 .     5    m

            2

         J    o     i    n     t    s    c    r    o    s    s   —     0     7     5     

        m    a    s     t    e    r   —     1     6     5     

         L    a

         d     d    e    r    p    a     t     t    e    r    n    w     i     t     h    a

         b     i    m    o     d    a     l     t    r    a    c    e     l    e    n    g     t     h

         d     i    s     t    r     i     b    u     t     i    o    n

         1     6     5         s    e     t     f    o    r    m    e

         d     d    u    r     i    n    g     L    a     t    e

         C    r    e     t    a    c    e    o    u    s   –     E

        a    r     l    y     M     i    o    c    e    n    e     A     l    p     i    n    e

        c    o    m    p    r    e    s    s     i    o    n

         (     N    e    m    c    o    c     k    e     t    a     l . ,

         1     9     9     5     )    ;

         0     7     5         s    e     t     f    o    r    m

        e     d     d    u    e     t    o    r    e     l    a    x    a     t     i    o    n    o    r

        c    o    n     t    r    a    c     t     i    o    n    a

         t     l    a     t    e    r    s     t    a    g    e     (     R    a    w    n    s     l    e    y

        e     t    a     l . ,

         1     9     9     8     )

         A    m    r    o     t     h ,

         W    a     l    e    s ,     U    n     i     t    e     d

         K     i    n    g     d    o    m ,

         S     N     1     7     5     6     0     7     2     2     *

         (     F     i    g    u    r    e     6     B     )

         E    a    r     l    y     W    e    s

         t    p     h    a     l     i    a    n

         (     C    a    r     b    o    n

         i     f    e    r    o    u    s     ) ,

         E    a    r     l    y

         W    e    s     t    p     h    a     l     i    a    n    c    o    a     l

        m    e    a    s    u    r    e

        s     (    s    a    n     d    s     t    o    n    e     )

         T     h     i    s    s     t    u     d    y ,

         4     0    c    m ,

         1    :     2     0 ,

         1     6     2 .     0    m

            2

         (    e    x    p    o    s    e

         d

        a    r    e    a     ) ,     1     0     4 .     7    m

            2

         (    s     t     i    p    p     l    e     d    a    r    e    a     )

         J    o     i    n     t    s     2     0     0      ,     2     9     0      ,

         3     1     6     

         O    r     t     h    o    g    o    n    a     l    p    a     t     t    e    r    n

         (     2     0     0          &     2     9     0          )    a    n     d

        o    n    e    y    o    u    n    g    e    r     j     o     i    n     t

        s    e     t

         2     0     0         a    n     d     2     9     0     

        s    e     t     f    o    r    m    e     d     d    u    r     i    n    g

        a     l     t    e    r    n    a     t     i    n    g     r

            2

        a    n     d     r        3

        s     t    r    e    s    s

         d     i    r    e    c     t     i    o    n    s     d    u

        r     i    n    g     V    a    r     i    s    c    a    n     t     h    r    u    s     t     i    n    g

         (     D    u    n    n    e    a    n     d     N    o    r     t     h ,

         1     9     9     0     )

         S     l    e     i    p    n    e    r     V    e    s     t     fi    e     l     d ,

         N    o    r     t     h     S    e    a ,

         (     F     i    g    u    r    e     6     C     )

         J    u    r    a    s    s     i    c ,     H

        u    g     i    n

         F    o    r    m    a     t     i    o    n ,    r    e    s    e    r    v    o     i    r

        s    a    n     d    s     t    o    n    e

         O     t     t    e    s    e    n     E     l     l    e    v    s    e     t    e     t

        a     l .

         (     1     9     9     8     ) ,        2     0     0    m ,

        u    n     k    n    o    w    n ,

         1     1     0 .     7

         k    m

            2

         F    a    u     l     t    s    v    a    r     i    e     t    y    o     f

        o    r     i    e    n     t    a     t     i    o    n    s

         S    o

        m    e    w     h    a     t    p    o     l    y    g    o    n    a     l

         L    a     t    e     J    u    r    a    s    s     i    c   –     E

        a    r     l    y     C    r    e     t    a    c    e    o    u    s

        e    x     t    e    n    s     i    o    n ,    p    o

        s    s     i     b     l    y    r    e     l    a     t    e     d     t    o    s    a     l     t

        p     i     l     l    o    w     f    o    r    m    a     t     i    o    n     (     O     t     t    e    s    e    n     E     l     l    e    v    s    e     t    e     t

        a     l . ,

         1     9     9     8     )

         C    a    r     t     i    e    r     T    r    o    u    g     h ,

         T     i    m    o    r

         S    e    a     (    u    n    s    p    e    c     i     fi    e     d     )

         (     F     i    g    u    r    e     6     D     )

         U    p    p    e    r     C    r    e     t    a    c    e    o    u    s     t    o

         H    o     l    o    c    e    n

        e    s    e     d     i    m    e    n     t    s

         W    a     l    s     h    e     t    a     l .     (     1     9     9     6     ) ,

              4     0    m

         d     i    s    p     l    a    c    e    m

        e    n     t ,

        u    n     k    n    o    w    n ,

         5     5     5 .     6

         k    m

            2

         F    a    u     l     t    s     W     S     W

         S     i    n    g     l    e    s    e     t     i    n     t    e    r    m    s    o     f

        o    r     i    e    n     t    a     t     i    o    n

         P     l     i    o    c    e    n    e   –     P     l    e     i    s     t    o    c    e    n    e    n    o    r    m    a     l     f    a    u     l     t    s

         (     W    a     l    s     h    e     t    a     l . ,

         1     9     9     6     )

         T    e     l    p    y    n     P    o     i    n     t ,     W    a     l    e    s ,

         U    n     i     t    e     d     K     i    n    g     d    o    m ,

         S     N     1     8     3     3     0     7     3     1     *

         (     F     i    g    u    r    e     6     E     )

         L    a     t    e     N    a    m

        u    r     i    a    n

         (     C    a    r     b    o    n

         i     f    e    r    o    u    s     ) ,

         U    p    p    e    r

         S    a    n     d    s     t    o    n    e     G    r    o    u    p

         T     h     i    s    s     t    u     d    y ,

         4     0    c    m ,

         1    :     2     5 ,

         2     4     7 .     6    m

            2

         J    o     i    n     t    s     2     0     0      ,     2     6     7      ,

         2     9     0      ,     3     1     8     

         O    r     t     h    o    g    o    n    a     l    p    a     t     t    e    r    n

         (     2     0     0          &     2     9     0          )    ;    o     t     h    e    r

        y    o    u    n    g    e    r    s    e     t    s

         S    a    m    e    a    s     A    m    r    o

         t     h

         W    a    r     d     L    a     k    e ,

         C    a     l     i     f    o    r    n     i    a ,

         U    n     i     t    e     d     S     t    a     t    e    s ,    n    o     t

        s    p    e    c     i     fi    e     d     (     F     i    g    u    r    e     6     F     )

         C    r    e     t    a    c    e    o    u

        s ,     M     t .     G     i    v    e    n    s

         G    r    a    n    o     d     i    o    r     i     t    e

         S    e    g    a     l     l    a    n     d     P    o     l     l    a    r     d

         (     1     9     8     3     ) ,        1     0     0    c    m

     ,

        u    n     k    n    o    w    n ,

         1     8     3     5 .     6

        m        2

         (    e    x    p    o    s    e     d    a    r    e    a     ) ,     1

         1     6     9 .     1

        m        2

         (    s     t     i    p    p     l    e     d    a    r    e    a

         )

         J    o     i    n     t    s     0     1     0   –     0     2     0     

         S     i    n    g     l    e    s    e     t

         A    g    e    o     f     j     o     i    n     t    s     h

        a    s     b    e    e    n    r    e    p    o    r     t    e     d    a    s

        e     i     t     h    e    r    p    r    e   -     E    o

        c    e    n    e    o    r    p    o    s     t   -     P     l     i    o    c    e    n    e

         (     S    e    g    a     l     l    a    n     d     P

        o     l     l    a    r     d ,

         1     9     8     3     )    ;     j     o     i    n     t    s    e     t    s

        a    r    e    o     f    r    e    g     i    o    n

        a     l    e    x     t    e    n     t

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    9/16

    Rohrbaugh et al.   2097

         P     1     0     0 ,

         Y    u    c    c    a     M     t . ,

         N    e    v    a     d    a ,

         U    n     i     t    e     d

         S     t    a     t    e    s ,     5     6     2     0     5     1 .     8

         f     t

         E         †

         7     6     3     4     0     7 .     7

         f     t     N         †

         (     F     i    g    u    r    e     6     G     )

         M     i    o    c    e    n    e ,

         T     i    v    a     C    a    n    y    o    n

         T    u     f     f     (    u    p    p    e    r     l     i     t     h    o    p     h    y    s    a     l

        z    o    n    e     )

         B    a    r     t    o    n    e     t    a     l .     (     1     9     9     3     ) ,     2     0

        c    m ,

         1    :     5     0 ,

         2     3     3 .     2    m

            2

         (    e    x    p    o    s    e     d    a    r    e    a     ) ,     1

         2     4 .     1

        m        2

         (    s     t     i    p    p     l    e     d    a    r    e    a

         )

         J    o     i    n     t    s    c    o    o     l     i    n    g   —     0     5     0      ,

         3     2     0         ;     t    e    c     t    o    n     i    c   —     0     1     0      ,

         0     4     0      ,     3     2     5      ,     3     5     9     

         O    r     t     h    o    g    o    n    a     l    s    e     t    o     f

        c    o    o     l     i    n    g     j     o     i    n     t    s    w     i     t     h

         t     h    e     0     5     0          b    e     i    n    g

         d    o    m     i    n    a    n     t    ;    p    o    o    r     l    y

         d    e    v    e     l    o    p    e     d     t    e    c     t    o    n     i    c

         j     o     i    n     t    s

         C    o    o     l     i    n    g     j     o     i    n     t    s     f    o    r    m    e     d     fi    r    s     t     d    u    r     i    n    g

         t     h    e    r    m    o    e     l    a    s     t     i    c

        r    e     l    a    x    a     t     i    o    n     f    r    o    m

        c    o    o     l     i    n    g    ;     t     h    e     t    e    c     t    o    n     i    c     j     o     i    n     t    s    p    o    s     t     d    a     t    e

        c    o    o     l     i    n    g     j     o     i    n     t    s

        ;     b    a    s    e     d    o    n     t    e    r    m     i    n    a     t     i    o    n

        r    e     l    a     t     i    o    n    s     h     i    p    s ,

         t     h    e     3     2     5          t    e    c     t    o    n     i    c    s    e     t     i    s

        o     l     d    e    s     t ,

         f    o     l     l    o    w

        e     d     b    y     3     5     9         s    e     t ,    w     i     t     h

         0     4     0         s    e     t    y    o    u    n    g    e    s     t .

         (     B    a    r     t    o    n    e     t    a     l . ,

         1     9     9     3     )

         R    o    c     k    y     C    r    e    e     k ,

         S    o    u     t     h

         C    a    r    o     l     i    n    a ,

         U    n     i     t    e     d

         S     t    a     t    e    s ,     5     1     1     2     1     5 .     3    m

         E     *     *     3     8     2     1     8     4     3 .     5    m

         N     *     *     (     F     i    g    u    r    e     6     H     )

         C    a    m     b    r     i    a    n

     ,     G    r    e    a     t     F    a     l     l    s

         M    e     t    a    g    r    a

        n     i     t    e

         M .

         F .

         S    c     h    a    e     f     f    e    r ,     1     9     9     8 ,

        u    n    p    u     b     l     i    s     h    e     d     d    a     t    a ,

         2     5

        c    m ,

         1    :     1     2     0 ,

         1     4     4     6 .     6

        m        2

         (    e    x    p    o    s    e     d    a    r    e    a     ) ,     9

         2     2 .     2

        m        2

         (    e    x    p    o    s    e     d ,    s     t     i    p

        p     l    e     d

        a    r    e    a     )

         J    o     i    n     t    s     0     6     8      ,     2     7     4      ,

         2     9     2      ,     3     5     6     

         M

        u     l     t     i    p     l    e     f    r    a    c     t    u    r    e    s    e     t    s    ;

         l    e    n    g     t     h     d     i    s     t    r     i     b    u     t     i    o    n     i    s

         i    n     d    e    p    e    n     d    e    n     t    o     f

        o    r     i    e    n     t    a     t     i    o    n

         J    o     i    n     t    s    a    r    e    a     t     h     i    g     h    a    n    g     l    e    s    a    n     d    o     f

         t    e    c     t    o    n     i    c    o    r     i    g     i    n    ;    g    r    e    e    n    s    c     h     i    s     t     f    a    c     i    e    s

        m     i    n    e    r    a     l    s    a     l    o    n

        g     t     h    e     j     o     i    n     t    s     3     0     0     M    a    ;

         t     h    e     l    a    u    m    o    n     i     t     i     t    e    m     i    n    e    r    a     l     i    z    a     t     i    o    n

        r    e     l    a     t    e    s     t    o     M    e

        s    o    z    o     i    c    r     i     f     t     i    n    g     1     5     0   –     2     0     0

         M    a     (     M .

         F .     S    c

         h    a    e     f     f    e    r ,     1     9     9     9 ,    p    e    r    s    o    n    a     l

        c    o    m    m    u    n     i    c    a     t     i    o    n     ) .

         *     B    r     i     t     i    s     h     N    a     t     i    o    n    a     l     G    r     i     d     R    e     f    e    r    e    n    c    e     S    y    s     t    e    m

         (     U    n     i    v    e    r    s    a     l     T    r    a    n    s    v    e    r    s    e     M    e    r    c    a     t    o    r    s    y    s     t    e    m     ) .

         *     *     U    n     i    v    e    r    s    a     l     T    r    a    n    s    v    e    r    s    e     M    e    r    c    a     t    o    r    s    y    s     t    e    m

         f    o    r     U    n     i     t    e     d     S     t    a     t    e    s .

             †     5     0     0     0     f     t    g    r     i     d     b    a    s    e     d    o    n     N    e    v    a     d    a     S     t    a     t    e     P     l    a    n    e    c    o    o    r     d     i    n    a     t    e    s ,     N    e    v    a     d    a     S     t    a     t    e     P     l    a    n    e    p    r    o     j     e    c     t     i    o    n .

    block size for the two dominant sets is about 2.5 m by

    1 m. At each pavement, 100 circles with diameters

    smaller than the block size and 100 circles with di-

    ameters larger than the block size (0.2 and 1.5 m at 

    Llantwit; 0.5 and 3.0 m at Amroth) were used to es-

    timate intensity. The smaller circles yield estimates

    with much greater variability than the larger circles

    (Figure 8). Therefore, to reduce the variability of es-timates, circles larger than mean block size are rec-

    ommended for estimating fracture parameters. Similar

    conclusions are applicable to the other estimators and

    to fault networks interpreted from seismic cubes. For

    example, the greatest estimator error occurs at Sleip-

    ner Vest and Cartier Trough for the smallest sampling

    circles (Figure 7E, F).

    Although these estimators are effective with sam-

    ples from two-dimensional surfaces such as rock pave-

    ments and maps, they cannot be used effectively with

    borehole imagery data because terminations inside theborehole cannot be uniquely constrained and, hence,

    m  counts are subject to interpretation. Thus, none of 

    the existing estimation approaches are particularly ef-

    fective with borehole data sets at present.

    Furthermore, inferring subsurface joint patterns

    from boreholes is, in general, problematic, because

    boreholes are likely to be smaller than block size in

    naturally fractured rocks. Perhaps this issue can be re-

    solved by considering microfractures in core specimens

    (Laubach, 1997; Ortega and Marrett, 2000), but im-

    plementation of the circle-based estimators on micro-fractures inside a solid core is beyond the scope of this

    article.

    Effects of Pattern Heterogeneity

    To consider the effect of pattern heterogeneities and

    circle size on estimator variance, intensity estimates for

    fractures from the rock pavements were analyzed as a

    function of circle size (Tables 2, 3; Figure 6). Fault 

    trace maps could not be included in the analysis be-

    cause their greater size as compared to the joint pat-terns precluded the use of sampling circles with the

    same size.

    The fracture patterns that were selected for this

    analysis encapsulate a wide range of likely fracture ge-

    ometries and settings (Table 2; Figure 6). They include

    a ladder geometry where cross-joint spacing is con-

    trolled by the spacing of master joints (Figure 6A); an

    orthogonal pattern of large fractures overprinted by

    younger, less persistent joints (Figure 6B); patterns

    with clustered fractures due to fault development or

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    10/16

    2098   Fracture Characterization Using Circular Scanlines and Windows

    Figure 6.  Fracture trace maps. (A) Llantwit Major. (B) Amroth. (C) Sleipner Vest (modified from Ottesen Ellevset et al., 1998). (D)Cartier Trough (modified from Walsh et al., 1996). Continued.

    igneous processes (Figure 6E, G); and patterns in crys-

    talline rocks that vary from a single fracture set in a

    granite (Figure 6F) to a complex array of fractures that 

    accumulated during a long uplift and erosion history

    (Figure 6H). Characteristics of these patterns include

    spatial heterogeneity due to different ages of joint sets,

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    11/16

    Figure 6.  Continued. (E) Telpyn Point. (F) Ward Lake (modified from Segall and Pollard, 1983). (G) P100 (modified from Barton etal., 1993). (H) Rocky Creek (modified from M. F. Schaeffer, 1998, unpublished data). Light-gray regions in (B), (F), (G), and (H)

    represent visually identified homogeneous subdomains, and black regions in (F) and (H) are unexposed. See Table 2 for geologicinformation for trace maps.

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    12/16

    2100   Fracture Characterization Using Circular Scanlines and Windows

    Figure 7.   Estimates of natural fracture characteristics using 10circular scan lines/windows (diamonds) and areal samples(dashed lines). (A) Density at Llantwit Major. (B) Mean trace

    length at Llantwit Major (MJ

    master joints; CJ

    cross joints).(C) Intensity at Amroth. (D) Intensity at Sleipner Vest. (E) Mean trace length at Sleipner Vest. (F) Density at Cartier Trough. Es- timates are in meters for Llantwit Major and Amroth and inkilometers for Sleipner Vest and Cartier Trough.

    varying degrees of persistence, clustering, and the local

    heterogeneities arising from the mechanics of joint 

    origin.

    Spatially homogeneous fracture domains were es-

    tablished initially by visual inspection. However, afirst-pass visual inspection to eliminate obvious hetero-

    geneities was not sufficient to identify homogeneous

    domains in all pavements. This insufficiency means

    that the circle-based estimators could be tested both

    for their performance in homogeneous regions and for

    their ability to detect pattern heterogeneity in regions

    of proposed homogeneity.

    Intensity was sampled for the entire pavements at 

    Telpyn Point and Llantwit Major and for subdomains

    at Amroth, Ward Lake, Yucca Mountain, and Rocky

    Creek (Figure 6). To simplify comparison among the

    different fracture geometries and sampling schemes,

    circle-based intensity estimates within 15% of the areal

    value were considered accurate and are marked with Y

    (for “yes”) in Table 3. Ranges of low (1 fracture per

    m), moderate (1–1.99 per m), and high (2 per m)

    were adopted for intensity magnitude to facilitate com-

    parison (Table 3). Circle radii of 1–2 m were used toachieve a high likelihood of intersections between frac-

    ture traces and circular scan lines as a function of block 

    size or fracture spacing, while not exceeding pavement 

    size. Ten circles were used to balance the need for a

    large sample with the need to be able to do work in a

    timely manner.

    The intensity of seven single fracture sets and two

    entire fracture patterns was evaluated (Table 3). Four

    fracture sets and two fracture patterns yielded accurate

    circle-based intensity estimates as compared to the es-

    timate for the entire pavement. We interpret theagreement between the two methods as indicating that 

    these patterns or fracture sets are spatially homoge-

    neous. Three fracture sets, however, yielded inaccurate

    results. Inspection of these pavements indicated that 

    the inaccuracy results from spatial heterogeneity. For

    example, the 316 set at Amroth is more abundant at 

    the eastern end of the pavement subdomain (Figure 6),

    so, unlike the 200  set, it is heterogeneously distrib-

    uted. The cross joints at Llantwit Major are less intense

    at the western end of the pavement, which is where

    Figure 8.  Variability of estimates as a function of circle size with respect to block size (small circles open diamonds; largecircles black diamonds; see text for size details). Variabilityis represented by the coefficient of variation (standard deviationdivided by the mean) for 100 circle counts for each sample.

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    13/16

    Rohrbaugh et al.   2101

    Table 3.  Accuracy of Intensity Estimates Using Ten Circular Scanlines for a Variety of Natural Fracture Patterns

    Estimate Accuracy, Scanline Radius

    Natural Fracture Patterns 1.0 m 1.5 m 2.0 m Intensity** (m/m2)

    Llantwit Major, cross joint set Y     Y    N   High

    14.2 8.8 18.6 2.00

    Llantwit Major, master joint set Y  

      Y 

      Y 

      High5.4 12.6 7.7 3.40

     Amroth, 316 set N   N   N   Low

    49.5   63.1   35.9 0.55

     Amroth, 200 set Y     Y    Y    Mod

    9.9   10.6   0.1 1.80

    Telpyn Point, 290 set Y     Y    Y    Low

    15 13.3 2.5 0.50

    Telpyn Point, 200 set Y     Y    Y    High

    9.0 14.0 3.4 3.60

     Ward Lake, 015 set N   Y    N   Mod

    50.7 2.9 23.4 1.00

     Yucca Mountain, all fracture sets Y    Y    Y    High

    7.6   9.7 8.2 2.80

    Rocky Creek, all fracture sets Y     Y    Y    High

    6.0   4.8 13.7 3.30

    *Y(yes)/N(no) indicates whether estimate is/is not within 15% of the areal value;  / denotes an over/underestimate.

    **Actual percent error for the intensity estimate are shown in the second row for each fracture system. Low/moderate/high intensity indicates values in the ranges

    (0:1), (1:2), or (2), respectively.

    larger sampling circles are restricted because of pave-

    ment width (Figure 6). The subdomain at Ward Lakeeliminates a low intensity region for the fracture set 

    but contains heterogeneities due to fracture clusters

    and covered areas that prevent random circle place-

    ment. Thus, the geometry of older joints, joint persis-

    tence, and joint clustering, as expected, generate spa-

    tial heterogeneity. More important, based on these

    results, we believe that a sampling strategy of 10 circles

    with a size that exceeds the block size or fracture spac-

    ing but is substantially smaller than the minimum di-

    mension of a sample area will yield an intensity esti-

    mate within 15% of the actual intensity for ahomogeneous fracture pattern.

    Minimum Count of  m for Reliable Tracelength Estimates

    Of the three estimators discussed here, the mean trace-

    length estimator exhibits the greatest variability. Low

    trace densities, large fracture trace lengths relative to

    circle size, or small total sample area can lead to small

    counts of fracture endpoints (m). Such small counts

    may be an issue for the petroleum industry be-

    cause boreholes are commonly small with respect to

    fractures. These small counts of  m  can lead to signifi-cant errors in the mean trace-length estimator (Table

    1), because as m  tends to zero in the denominator, the

    estimate tends to infinity. Thus, it is important to en-

    sure that a small  m  count is an adequate sample of a

    fracture pattern with large trace lengths and not a sam-

    pling artifact due to insufficient sample size, thereby

    producing an overestimate of mean trace length.

    To evaluate this issue, five simple synthetic trace

    patterns with a single fracture set of intensity 2 m/m2

    were examined. The five cases were for trace patterns

    with individual traces of length 10, 16, 20, 25, or 50m, respectively. For each case, a suite of 35 sampling

    strategies (combinations of circle size and number) was

    used, with circle radii ranging from 1 to 20 m and num-

    ber of circles ranging from 2 to 10. This combination

    of cases and strategies yielded a total of 175 samples in

    most of which trace lengths exceeded the circle di-

    ameters, consistent with the petroleum industry situ-

    ation of fractures larger than borehole diameter.

    The accuracy of circle-based estimates of mean

    trace length, as determined by comparison with the

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    14/16

    2102   Fracture Characterization Using Circular Scanlines and Windows

    Figure 9.   Graph illustrating the performance of the mean trace-length estimator as a function of the total number of  mcounts in terms of the percentage of accurate estimates for agiven total m  count.

    input mean trace-length values, is graphed as a func-

    tion of m counts in Figure 9. For this graph , an accurate

    result was defined as within 15% of the input value.

    Using this criterion, virtually all sampling strategies

    that yielded   m  counts of 30 or greater produced ac-

    curate estimates. These results (Figure 9) suggest that 

    for a specified fracture intensity, a threshold value of 

    m counts exists for which accurate estimates are a nearcertainty.

    This analysis could not be extended directly to nat-

    ural fracture patterns, because, due to problems of cen-

    soring and length bias, no direct technique exists to

    accurately estimate mean trace length for comparison

    to the results of the circle-based estimator. Although

    threshold values of   m   counts for accurate estimates

    could be determined for our synthetic patterns, our

    models had Poissonian trace center distributions,

    which is unlikely in nature. Still, results from the syn-

    thetic patterns should provide guidance, such as in thecase considered here, where a fracture pattern with

    large fractures and an intensity of 2 m/m2 that is being

    sampled by small circles (or small-diameter boreholes)

    needs an m  count greater than 30 to yield an accurate

    estimate at the 15% level.

    U S I N G N E W E S T I M A T O R S W I T HF R A C T U R E G E N E R A T O R S

    One way to evaluate fractures as factors in hydrocar-bon transport and storage in reservoirs is to use com-

    puter programs to generate representative fracture net-

    works in synthetic rock volumes (Dershowitz and

    LaPointe, 1994; Swaby and Rawnsley, 1996; Renshaw,

    2000). These synthetic networks are then modeled for

    fluid flow and storage. The primary contribution that 

    the circle-based estimators make to such an approach

    is providing more accurate estimates for program in-

    puts and validating outputs. Intensity estimates can, for

    example, substitute for dimensionally equivalent den-

    sity or frequency input values in programs. In addition,these estimates of density, size, and intensity provide

    tests for determining whether synthetic networks are

    truly representative.

    C O N C L U S I O N S

    1. New circle-based estimators for fracture intensity,

    density, and mean trace length virtually eliminate

    orientation, censoring, and length biases, which se-

    verely limit the effectiveness of the straight scan line

    and area methods.

    2. Estimator accuracy is improved and variability re-

    duced by using circles larger than mean fracture

    block size or fracture spacing for a single set.3. A sampling strategy of 10 circles with a diameter

    that exceeds block size or fracture spacing but is

    significantly less than the minimum dimension of a

    sample region yields an intensity estimate within

    15% of the actual intensity for a homogeneous frac-

    ture pattern.

    4. When using small circles to estimate mean trace

    length of large fractures, care should be taken to use

    sampling strategies that gather a sufficient  m  count 

    to eliminate spurious overestimates of fracture size

    from small samples. Guidance about the necessarynumber of  m  counts can be gained from analyzing

    synthetic patterns with similar characteristics.

    5. The new estimators provide both inputs and inde-

    pendent checks of predictions for fracture-genera-

    tor programs that model fracture populations in a

    rock volume.

    6. Fracture characterization protocols that deal with

    sampling biases, such as the ones presented in this

    article, are likely to yield improved input for explo-

    ration and development decisions.

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    15/16

    Rohrbaugh et al.   2103

    R E F E R E N C E S C I T E D

    Baecher, G. B., and N. A. Lanney, 1978, Trace length biases in joint surveys: Proceedings of the 19th U.S. Symposium on Rock Me-chanics, v. 1, p. 56–65.

    Barton, C. C., E. Larsen, W. R. Page, and T. M. Howard, 1993,Characterizing fractured rock for fluid-flow, geomechanical,and paleostress modeling: methods and preliminary resultsfrom

    Yucca Mountain, Nevada: United States Geological SocietyOpen-File Report 93-269, 62 p.Becker, A., and M. R. Gross, 1996, Mechanism for joint saturation

    in mechanically layered rocks: an example from southern Israel:Tectonophysics, v. 257, p. 223–237.

    Cruden, D. M., 1977, Describing the size of discontinuities: Inter-national Journal of Rock Mechanics and Mining Sciences andGeomechanical Abstracts, v. 14, p. 133–137.

    Dershowitz, W. S., and H. H. Herda, 1992, Interpretationof fracturespacing and intensity,  in  J. R. Tillerson and W. R. Wawersik,eds., Proceedings of the 33rd U.S. Symposium on Rock Me-chanics: Rotterdam, Balkema, p. 757–766.

    Dershowitz, W. S., and P. R. LaPointe, 1994, Discrete fracture ap-proaches for oil and gas applications,  in  R. A. Nelson and S.Laubach, eds., Rock mechanics: Rotterdam, Balkema, p. 19–

    30.Dunne, W. M., and C. P. North, 1990, Orthogonal fracture systems

    at the limits of thrusting: an examplefrom southwesternWales:Journal of Structural Geology, v. 12, p. 207–215.

    Einstein, H. H., and G. B. Baecher, 1983, Probablistic and statisticalmethods in engineering geology, specific methods and exam-ples, part 1: exploration: Rock Mechanics and Rock Engineer-ing, v. 16, p. 39–72.

    Ghosh, A., and J. J. K. Daemen, 1993, Fractal characteristics of rock discontinuities: Engineering Geology, v. 34, p. 1–9.

    Gilmour, H. M. P., D. Billaux, and J. C. S. Long, 1986, Models forcalculating fluid flow in randomly generated three-dimensionalnetworks of disk-shaped fractures: theory and design of FMG3D, DISCEL and DIMES: Berkeley, California, Lawrence

    Berkeley Laboratory, Earth Sciences Division, LBL-19515,144 p.

    Jian, X., R. Olea, and Y. Yu, 1996, Semivariogram modeling byweighted least square: Computers and Geosciences, v. 22,p. 387–397.

    Kulatilake, P. H. S. W., and T. H. Wu, 1984, The density of discon-tinuity traces in sampling windows (technical note): Interna-tional Journal of Rock Mechanics and Mining Sciences andGeomechanical Abstracts, v. 21, p. 345–347.

    Kulatilake, P. H. S. W, D. N. Wathugala, and O. Stephansson, 1993,Joint network modeling with a validation exercise in StripaMine, Sweden: International Journal of Rock Mechanics andMining Sciences and Geomechanical Abstracts, v. 30, p. 503–526.

    Kulatilake, P. H. S. W., R. Fiedler, and B. B. Panda, 1997, Box fractaldimension as a measure of statistical homogeneity of jointedrock masses: Engineering Geology, v. 48, p. 217–229.

    LaPointe, P. R., 1988, A method to characterize fracture density andconnectivity through fractal geometry: International Journal of Rock Mechanics and Mining Sciences and Geomechanical Ab-stracts, v. 24, p. 421–429.

    LaPointe, P. R., 1993, Pattern analysis and simulation of joints forrock engineering,   in   J. A. Hudson, ed., Comprehensive rock engineering, volume 3—rock testing and site characterization:New York, Pergamon Press, p. 215–239.

    LaPointe, P. R., and J. A. Hudson, 1985, Characterization and in-terpretation of rock mass joint patterns: Geological Society of America Special Paper 199, 37 p.

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

    Marrett, R., O. J. Ortega, and C. M. Kelsey, 1999, Extent of power-law scaling fornatural fractures in rocks: Geology, v. 27,p. 799–802.

    Mauldon, M., 1998, Estimating mean fracture trace length and den-sity from observations in convex windows: Rock Mechanics andRock Engineering, v. 31, p. 201–216.

    Mauldon, M., and W. Dershowitz, 2000, A multi-dimensional sys-

    tem of fracture abundance measures: Geological Society of America Abstracts with Programs, v. 32, no. 7, p. A474.

    Mauldon, M., and J. G. Mauldon, 1997, Fracture sampling on a cyl-inder: from scan lines to boreholes andtunnels: Rock Mechanicsand Rock Engineering, v. 30, p. 129–144.

    Mauldon, M., W. M. Dunne, and M. B. Rohrbaugh Jr., 2001, Cir-cular scan lines and circular windows: new tools for character-izing the geometry of fracture traces: Journal of Structural Ge-ology, v. 23, p. 247–258.

    Narr, W., 1996, Estimating average fracture spacing in subsurfacerock: AAPG Bulletin, v. 80, p. 1565–1586.

    Nelson, R. A., 1985, Geologic analysis of naturally fractured reser-voirs: Houston, Texas, Gulf Publishing, 320 p.

    Nemcock, M., R. Bayer, and M. Miliorizos, 1995, Structural analysisof the inverted Bristol Channel Basin: implications for the ge-ometry and timing of fracture porosity,  in  J. G. Buchanan andP. G. Buchanan, eds., Basin inversion: Geological Society Spe-cial Publication 88, p. 355–392.

    Ortega, O. J., and R. Marrett, 2000, Prediction of macrofractureproperties using microfracture information, Mesaverde Groupsandstones, San Juan basin, New Mexico: Journal of StructuralGeology, v. 22, p. 571–587.

    Ottesen Ellevset, S., R. Knipe, T. Svava Olsen, Q. J. Fisher, and G.Jones, 1998, Fault controlled communication in the SleipnerVest field, Norwegian continental shelf: detailed, quantitativeinput for reservoir simulation and well planning,  in  G. Jones,Q. J. Fisher, and R. J. Knipe, eds., Faulting, fault sealing andfluid flow in hydrocarbon reservoirs: Geological Society SpecialPublication 147, p. 283–297.

    Peacock, D. C. P., S. J., Harris, and M. Mauldon, in press, Use of curved scan lines and boreholes to predict fracture frequencies:Journal of Structural Geology, v. 25, p. 109–119.

    Pickering, G., J. M. Bull, and D. J. Sanderson, 1995, Samplingpower-law distributions: Tectonophysics, v. 248, p. 1–20.

    Priest, S. D., 1993, Discontinuity analysis for rock engineering: NewYork, Chapman and Hall, p. 50–54.

    Priest, S. D., and J. A. Hudson, 1981, Estimation of discontinuityspacing and trace length using scan line surveys: InternationalJournal of Rock Mechanics and Mining Sciences and Geome-chanics Abstracts, v. 18, p. 183–197.

    Rawnsley, K. D., D. C. P. Peacock, T. Rives, and J.-P. Petit, 1998,Joints in the Mesozoic sediments around the Bristol ChannelBasin: Journal of Structural Geology, v. 20, p. 1641–

    1661.Reiss, L. H., 1982, The reservoir engineering aspects of fracturedformations: Houston, Texas, Gulf Publishing, 186 p.

    Renshaw, C. E., 2000, An example of the use of geological and me-chanical constraints to develop a conceptual model of flow infractured, granitic rock: Geological Society of America Ab-stracts with Program, v. 32, no. 7, p. A64.

    Rohrbaugh Jr., M. B., 2000, Estimating joint intensity, density, andmean trace length using circular scan lines and circular win-dows: M.S. thesis, University of Tennessee, Knoxville, 96 p., 1CD-ROM.

    Segall, P., and D. D. Pollard, 1983, Joint formation in granitic rock of the Sierra Nevada: Geological Society of America Bulletin,v. 94, p. 563–575.

  • 8/19/2019 Estimating Fracture Trace Intensity Density and Mean Lenght Using Circular Scan Line and Windows

    16/16

    Stoyan, D., W. S. Kendall, and J. Mecke, 1995, Stochastic geometryand its applications, 2d ed.: New York, John Wiley and Sons,p. 286–296.

    Swaby, P. A., and K. D. Rawnsley, 1996, An interactive 3D fracturemodeling environment: Society of Petroleum Engineers, SPE36004, p. 177–187.

    Terzaghi, R. D., 1965, Sources of error in joint surveys: Geotech-nique, v. 15, p. 287–304.

    Turner, F. J., and L. E. Weiss, 1963, Structural analysis of meta-

    morphic tectonites: New York, McGraw-Hill, 545 p.

     Walsh, J. J., J. Watterson, C. Childs, and A. Nicol, 1996, Ductilestrain effects in the analysis of seismic interpretations of normalfault systems, in P. G. Buchanan and D. Nieuwland, eds., Mod-ern developments in structural interpretation, validation, andmodelling: Geological Society Special Publication 99, p. 27–40.

     Whitten, E. H. T., 1966, Structural geology of folded rocks: Chicago,Rand McNally, 678 p.

     Wu, H., and D. D. Pollard, 1995, An experimental study of therelationship between joint spacing and layer thickness: Journal

    of Structural Geology, v. 17, p. 887–905.