Convvra: A Center of Excellence in Earth Sciences and Engineering
Convvra: A Center of Excellence in Earth Sciences and Engineering
Convvra: A Center of Excellence in Earth Sciences and Engineering
July 3, 2001
Contract No. NRC-02-97-009
The attached manuscript is an outgrowth of work done through the Southwest Research Institute's Internal
Research and Development Program. This work identifies a new way that normal faults control the flow of
groundwater and hydrocarbons. When the fault plane transitions from a weak rock to a stronger rock the fault
plane may be refracted, or bent slightly. This bend produces a zone of dilation in the stronger rock that may result
in a migration pathway in saturated rocks or an infiltration pathway in the unsaturated zone. The results of this
study, although conducted in Carbonate rocks in Texas, may be applicable to groundwater studies at Yucca
Mountain. This copy of the manuscript is sent to you for informational purposes only and does not require NRC
acceptance.
If you have any questions please contact Dr. David Ferrill at (210) 522-6082 or me at (210) 522-5252.
Sincerely,
,IBudhi Sag ar
Technical Director
rae
Attachment
D:\GLGPGroupVettersXmiscVRD-07-03-200lhlni.wpd
CNWRA, Southwest Research Institute, 6220 Culebra Road, San Antonio, TX 78238
2 Department of Earth and Environmental Science, University of Texas at San Antonio, San
Antonio, TX 78249
Abstract - At low differential effective stress and with minimum principal compressive effective
stress near or below zero, rocks fail in multiple modes and with variable failure angles. Under
small-displacement, normal fault dip. Less competent layers fail in shear mode along fractures
that approximate the failure angle predicted by a standard rock-mechanics analysis. Deformation
of more competent layers, which is driven in part by interaction with the more rapidly deforming
incompetent layers, produces hybrid mode failure in which failure angles are smaller than in
shear mode. Analyses of small normal faults cutting Cretaceous carbonate strata in west Texas
indicate that fault geometries resulting from this effect commonly display dilational jogs where
the fault traverses more competent beds. Similar dilational jogs along faults within carbonates of
the Cretaceous-age Edwards Group near San Antonio, Texas have been enlarged by groundwater
I
1. Introduction
Normal faults in stratified rocks commonly have dip changes that are visible in
dip-parallel profiles. Variation in normal fault dip can be caused by various mechanisms,
including; (i) differential compaction of sedimentary layers after fault formation (Xiao and
Suppe, 1989); (ii) active fault deformation (e.g., by slip or shear along layering or intersecting
faults; Ferrill et al., 1998; Ferrill et al., 2000); (iii) linkage of an originally vertically-segmented
fault (Childs et al., 1996; Mansfield and Cartwright, 1996); and (iv) fault initiation with failure
angles controlled by rock mechanical properties and effective stresses at the time of failure
where characteristic fault dips are associated with particular lithologic units. These faults lack
diagnostic traits of passive deformation (differential compaction) or active deformation (e.g., slip
interpret the dip changes to be primary features of the faults, developed during fault nucleation
and growth, and controlled by the failure mode and failure angle of the mechanical layers
experiencing the faulting. Finally, we use examples of faults with dip changes in Edwards
aquifer strata in central Texas to illustrate the importance of dilation on steep fault segments to
2
2. Background
The angle of shear failure (0, measured between the maximum principal compressive
stress and the failure plane) in rock can be predicted from the angle of internal friction (friction
Failure mode depends upon the effective differential stress and the effective minimum principal
compressive stress at the time of failure, and upon the strength characteristics of the rock
(1) Shear failure, in which displacement is parallel to the fracture surface, and the normal
stress acting on the fracture surface, ;3 < ao >_ 0. The failure angle for this mode is given
(2) Tensile failure ("mode I", e.g., Twiss and Moores, 1992), in which displacement is
perpendicular to the fracture surface. In this case G3 = an < 0 and Ia 3 1 ; tensile strength of
the rock. The failure angle for this mode is 0' (Fig. lb).
(3) Hybrid failure, in which displacement is oblique to the fracture surface ("dilatant faults,"
Mandl, 1988). In this case G3 < an < 0 and 1a3 1 < tensile strength of the rock. The failure
angle for this mode lies between 0' and the angle predicted by equation 1 (Fig. Ic).
3
3. Analysis of natural examples from west Texas, U.S.A.
Horizontal to gently dipping layers of the Cretaceous Buda Limestone, exposed along
Interstate Highway 10 (1-10), in west Texas are cut by small-displacement normal faults. At a
roadcut exposure along eastbound I-10 at mile marker 195 (approximately 16 km west of
Balmorhea, Texas), several faults have consistent dip changes through the various limestone
layers (Fig. 2a). Layers in the outcrop dip gently, between horizontal and 100, and dip changes
Faults are undulatory, cutting layering at 80-90' in the more competent layers (K and H;
Figs. 2 b through e and Table 1), and generally 65-75' in less competent layers. Fault dip
changes tend to be gradual rather than abrupt (Figs. 2 a through e). Slip on the faults produced
normal offset, was approximately parallel to more shallowly dipping fault segments, and
produced dilation at steep fault segments within the more competent layers. In the context of
normal slip parallel to the more gently dipping fault segments, the steep segments are dilational
jogs, equivalent to releasing bends. In one case, a small dissolution cavity has formed,
suggesting localized groundwater flow along the dilational jog (also see Sibson, 1996).
Fault zone materials consist of a combination of crystalline calcite veins, rock fragments,
and clay sized material (Fig. 3a). Along the releasing bend walls, veins of milky calcite (up to
5 cm thick) are locally present (Fig. 3b). Calcite crystals in these veins range from 5-15mm
4
across, and in some cases have euhedral terminations toward the vein centers. Along slip
surfaces, coarse calcite veins that are well bonded with the wall rock are common. These veins
are typically 2-5 mm thick, and consist of equant to elongate crystals 2-5 mm across. The core
of the fault zones generally consists of angular to rounded carbonate rock fragments (1-30 mm
diameter) and clayey (some iron rich) material (Fig. 3b). The filling tends to be amorphous,
especially in fault releasing bends, but in some cases the fill is layered parallel to the walls along
fault segments that parallel the slip direction. Tan colored, fine grained calcite veneer is common
on wall rock interfaces and on open fracture surfaces within the fault zone. This finely crystalline
material has the appearance of flowstone and suggests precipitation along open cavity walls
3.3 Interpretation
Failure angles for competent rocks undergoing shear tend to be on the order of 10-20'
for competent rocks, and 15-45' for less competent rocks (Mandl, 1988). These angles decrease
to 0' for hybrid failure modes (Fig. 1). If we assume that faulting initiated when layering was
horizontal, then the angles between the faults and bedding represent the initial fault dips and the
failure angle for each fault segment is 90°-fault segment dip (Table 1). The faults investigated
here failed at a range of angles from 0-40', and individual mechanical layers consistently failed
at similar angles where cut by the four faults analyzed (Table 1). The lowest failure angles are
indicative of competent rock failing in the hybrid mode, and is manifest by the dilational sense of
displacement (see layers K and H in Figs. 2b through e and 3a and b; Mandl, 1988). Net slip on
the faults from west Texas presented here was parallel to the sections with shallower dips.
5
Consequently, the subvertical fault sections dilated an amount equal to the fault-perpendicular
Variations of fault dip are the result of variations in the mechanical strength of rock
layers coupled with low values of differential effective stress and a minimum principal
compressive stress (a3) that is weakly compressive to tensile (Fig. 4). These conditions cause
different layers to fail in different modes, hence generating different fracture orientations from
bed to bed.
Initially, the stress state within a weaker layer generates shear failure (stress state A,
Fig. 4). When the propagating tip of the shear fracture reaches the stronger layer, traction
between the deforming weaker layer and the as yet undeformed stronger layer causes stress
magnification in stronger layer. Because the stronger layer will not fail at the same differential
stress level as the weaker layer, the differential stress increases as a 3 decreases (becomes
overlying rock. Ultimately, deformation of the weaker layer generates a large enough differential
stress for the stronger layer to fail in hybrid mode (stress state B, Fig. 4). A consequence of
failure occurring under different conditions in the two layers is that the failure angle in each
layer is different, specifically, the failure angle in the stronger layer is smaller than that in the
weaker layer (Fig. 4). This can be thought of as refraction of the fault trajectory as the fracture
propagates through the rock mass. In general, the optimal fracture trajectory will be shorter in
6
The stress conditions illustrated in Fig. 4 are those for unsaturated carbonate rocks at
depths of about 250-300 meters. A similar situation could arise in saturated conditions where the
pore fluid pressure is sufficiently elevated to generate a tensile ( 3, and differential effective
Slip tendency is the ratio of resolved shear stress to resolved normal stress on a surface
(Morris et al., 1996). At the time of sliding, slip tendency equals the frictional resistance to
sliding on the surface. Slip tendency analysis of these fault yields the results illustrated in Fig. 5.
experienced by west Texas as the southern Basin and Range and Rio Grande Rift developed
during the late Tertiary. In this analysis, we assume a simple Andersonian normal faulting
regime, with the magnitude of the intermediate principal compressive stress centered between
the maximum and minimum compressive stress magnitudes. The steeper faults shown in profile
as blue lines have low resolved normal stress and low resolved shear stress - therefore they are in
low slip tendency orientations. These steep fault segments, approximately perpendicular to a 3
have a very high dilation tendency (Ferrill et al., 1999). High dilation tendency segments tend to
dilate in response to the same stress field that produces slip on the less steep fault segments.
4. Discussion
Net slip parallel to fault segments with lower dips causes steep fault segments to be
releasing bends or dilational jogs. These dilational jogs locally enhance fault permeability.
7
Small-displacement normal faults in Edwards aquifer strata from the Balcones Fault system near
San Antonio Texas exhibit dip changes at lithologic boundaries, similar to the examples from
west Texas (Fig. 6). These faults also slipped parallel to lower dip segments, producing dilation
of steep segments.
For vertical dilational fault segments, fracture porosity increase equals horizontal
extension magnitude (Fig. 7). Dilational jogs on small-displacement normal faults within the
Edwards aquifer are common sites of dissolution features (Figs. 6a and b). Because faults can
propagate through mechanically heterogeneous sequences, they tend to produce both vertically
and laterally connected flow pathways. Where these faults intersect the ground surface, they may
be important pathways for shallow infiltration of groundwater (Fig. 7). In the case of the
Edwards aquifer, the enhancement of flow pathways by dissolution is increased by the fact that
the weaker, lower-fault-dip layers are clay rich compared with the stronger, steeper-fault-dip
layers. Both the bed and the fault segment in the clay-rich layers have low permeability which
leads to ponding of water within the dilational jog, thus increasing its exposure to dissolution.
5. Conclusions
Rock failure in a mechanically layered sequence at shallow depths or high fluid pressures
results in variable failure angle. These variations in failure angle are caused by differences in
friction angle or failure mode from layer to layer. Faults that cut several mechanical layers will
have dip changes or will appear to "refract." The resulting refracted fault trajectories can
stratigraphy and stress conditions at the time of failure can be used to make predictions of fault
8
trajectories through a rock sequence. Refracted fault trajectories favor the development of
dilational normal faults, which can develop in mechanically heterogeneous rock sequences by
slip on faults cutting layers with different failure angles or different failure modes. Dilational
normal faulting significantly enhances vertical infiltration at the surface and along strike
Acknowledgments
This work was funded by Southwest Research Institute through the SwRI Internal
Research and Development Program (project #20.R9223.01.001), and supported by the Center
for Water Research at the University of Texas at San Antonio. We thank Larry McKague and
English Pearcy for helpful reviews, and Rebecca Emmot and Cheryl Patton for preparation of the
manuscript.
References
Childs, C., Nicol, A., Walsh, J.J., Watterson, J., 1996. Growth of vertically segmented normal
Ferrill, D.A., Morris, A.P., Jones, S.M., Stamatakos, J.A., 1998. Extensional layer parallel shear
Ferrill, D.A., Winterle, J., Wittmeyer, G., Sims, D.W., Colton, S. Armstrong, A., Morris. A.P.,
1999. Stressed rock strains groundwater at Yucca Mountain, Nevada. GSA Today 9(5):
1-8.
9
Ferrill, D.A., Morris, A.P., Sims, D.W., Stamatakos, J.A. 2000. Crossing conjugate normal
Goodman, R.E., 1980. Introduction to Rock Mechanics. John Wiley and Sons. 478 pp.
Hoek, E., Brown, E.T., 1988. The Hoek-Brown Failure Criterion-A 1988 Update. 15th
Mansfield, C.S., Cartwright, J.A., 1996. High resolution fault displacement mapping from
three-dimensional seismic data: evidence for dip linkage during fault growth. Journal of
Morris, A.P., Ferrill, D.A., Henderson, D.B., 1996. Slip tendency analysis and fault reactivation.
Peacock, D.C.P., Sanderson, D.J., 1992. Effects of layering and anisotropy on fault geometry.
Twiss, R.J., Moores, E.M. 1992. Structural Geology. W.H. Freeman and Company, New York.
532 pp.
Xiao, H., Suppe, J. 1989. Role of compaction in the listric shape of growth normal faults.
10
Figure Captions
Fig. 1. Mohr-space diagrams with Hoek-Brown failure envelope illustrating different failure
modes: (a) shear failure, (b) tensile failure, (c) hybrid failure.
Fig. 2. Field photographs showing normal faults cutting limestone layers in the Buda
Limestone, along 1-10 east near Balmorhea, Texas. Field Notebook, 12 cm x 19 cm for scale. (a)
Horst between northeast-dipping fault (4) and southwest-dipping fault (5). (b) Northeast-dipping
horst-bounding fault (4) with 9 cm displacement. (c) Southwest-dipping horst-bounding fault (5)
has 77 cm displacement. Note major and minor releasing bends and that horizontal dilation
across releasing bends equals heave component of fault slip. (d) Northeast-dipping fault (8) has 7
cm displacement near base of exposure, and loses displacement upward to a tip in layer J.
(e) Southwest-dipping fault (10) has 15 cm displacement near base of exposure, and loses
Fig. 3. Details of fault zone and releasing-bend-filling material along southwest dipping fault 5,
(see Fig. 2c). Note coarse calcite vein fill along walls of releasing bend visible in (b).
different failure mode and failure angles in different layers. A stronger rock (e.g., Tavernalle
Limestone) is more likely to fail in hybrid mode than a weaker rock (e.g., Bedford Limestone)
11
under the same normal faulting stress regime. See text for details. Equation and rock-quality
parameters for generating failure envelopes from Hoek and Brown, 1988; rock strength data
Fig. 5. Slip tendency analysis for faults illustrated in Figs. 2 and 3. Slip tendency plot for faults
G2 horizontal, azimuth 160; G3 horizontal, azimuth 070. (a) Slip tendency plot illustrates the
pattern of slip tendency, and triangles represent poles (dip corrected) to fault planes illustrated in
Fig. 2 and tabulated in Table 1. (b) Cross sectional profile of fault geometry color coded
according to slip tendency in interpreted stress field at time of faulting as shown in (a).
Fig. 6(a) and (b). Small-displacement normal faults in the Kainer Formation of the Edwards
Group Limestone exposed in the Balcones Fault Zone, along State Route 211, approximately
1 km south of San Geronimo, Texas. Note that dissolution cavities are localized at steep,
Fig. 7. Block diagram illustrating the role of dilational fault segments in shallow infiltration and
Table 1. Fault segment and bedding orientation data for faults in the Balmorhea exposure.
Failure angle is calculated assuming that faulting initiated when strata were horizontal. Unit
12
(a) v Failure angle = 0
"r Shear failure (
,2e
Tensile
strength
of rock
(Y3
G3 = an (01
(C)
e\ e o Y Failure angle < 0
f' Hybrid failure .e
(2
G an
Ferrill &Morris
Fig. I
(C, i Fig. 2b,c,d,e
Ferrill & Morris
I-"
C
LL LT
Hoek-Brown failure envelope Propagating
"t, MPa of stronger rock tip produces
(e.g., intact Tavernalle Limestone) drop in a3
5
0 5
Ferrill& Morris
Fig. 4
(a) (b)
(7G3 K
i .D, FG,IJ
Slip Tendency G2
* 0.64
* G
* F
* E
C
B
Fig. 5
Ferrill& Morris
I1
c14 I.
[4.
kQ\
Fig. 7
Ferrill& Morris
Table 1.
Corrected Fault Corrected Fault
Bed Fault strike Fault Dip Bedding Strike Bedding DIp Strike Dip Failure Angle
Fault 4
Fault 5
Fault 8
Fault 10
Lithologies
K (competent) bedded to moderately massive fossiliferous packstone/wackstone 86.5 3.5
J fossiliferous wackstone, wavy to sutured stylolites common, shaley base 72.25 17.75
calcareous sandstone 72.25 17.75
H (competent) massive, highly fossiliferous, poorly sorted, spar cemented grainstone 86.75 3.25
G buff wackstone/packstone 70.75 19.25
F gray packstone/wackstone 71.5 18.5
E buff colored wackstone/packstone 64.25 25.75
D gray packstone/grainstone 70 20
C buff colored wackstone/packstoDne 68.75 21.25
B gray packstonelgrainslone 76 14
A buff colored wackstonelpacastone 72.75 17.25