Testing Models of En Echelon Normal Fault Evolution Using 3D Computer Modeling
Hannah Mathy, Benjamin Surpless, and Samuel Simoneau; Department of Geosciences; Trinity University; San Antonio, TX, 78212 [contact e-mail: [email protected]]
Modeling: Propagating Normal Faults
Introduction
ult
Sev
lt
STRESS E-W (MPa)
0.0186
6.3
MEAN STRESS (MPa)
299.8
84.7
270.5
67.7
4.5
0.0135
2.6
55.0
241.1
0.8
42.2
219.1
0.0106
-1.1
89
29.5
-3.9
p fa
To
row
ea
I-15
VOLUME DILATION
9.1
UTAH
ARIZONA
197.1
0.0070
-6.7
COLORADO
PLATEAU
-9.5
16.7
175.1
145.7
-8.8
0.0005
116.4
-25.8
-16.0
BASIN &
RANGE
PROVINCE
Fig. 1
N
50 km
Ocean
Rocky
Mountain
s
Co
lor
Riv ado
er
94.4
-18.8
72.4
50.3
-0.0068
-59.8
28.3
-26.3
-0.0104
Figure 1. Physiographic context for the Sevier fault zone
study area within the Basin and Range-Colorado Plateau transi on zone (see inset). In combina on with the Grand Wash,
Hurricane, and Paunaugunt faults, the Sevier- Toroweap fault
helps accommodate extension across the transi on zone.
Ball is on the hanging wall of the west-dipping faults. Detailed geology of the Sevier fault study area (boxed) is displayed in Figure 2. Digital shaded relief modified from Thelin
and Pike (1991). Figure significantly modified from Reber et
al. (2001).
-76.8
6.3
-93.8
-30.0
-15.7
-0.0140
-33.7
-37.4
-110.8
-37.7
-0.0176
-127.8
-67.1
MODEL 2: EVEN SEGMENTS, 400 m DISPLACEMENT
DISPLACEMENT (m)
Background
-42.8
-0.0032
-22.5
COLO
PLAT RADO
EAU
ARIZONA
Vertical Displacement is the magnitude of displacement in the vertical direction, and as expected with
normal faults, negative displacement values are on the hanging wall of both fault segments. At low
displacement/overlap values, negative value regions are isolated from each other and decrease in
magnitude toward the west and toward the fault tips. The scales range from -37 to 9 m in Model 1 to
-378 to 111 m in Model 4.
As the faults propagate and maximum displacement increases, vertical displacement begins to affect
the overlap zone. Displacement is greatest (in both positive and negative directions) along the fault
zones farther from the fault tips, and a relay ramp forms between the two segments as displacement
increases and the faults propagate. Model 4 reveals a smooth transition in displacement from one fault
segment to the other.
4.0
0.0041
-13.2
NEVADA
RESEARCH QUESTIONS
1) How do stresses adjacent to an en echelon normal fault system change as segments
within that system overlap?
2) How do these changes in the local stress field impact subsequent propagation of fault
segments?
3) At what orientations do subsidiary structures (joints, fractures, minor faults) form as two
normal fault segments propagate and interact?
STUDY
AREA
(Fig. 2)
ugunt fau
ie r fa
u
Orderville
9
DISPLACEMENT (m)
Paunsa
NEVADA
UTAH
14
Toquerville
St. George
MODEL 1: UNDERLAPPED SEGMENTS, 100 m DISPLACEMENT
89
lt
I-15
Cedar City
fault
The Sevier fault zone of southern Utah near the town of Orderville has been well studied in the past (e.g.,
Reber et al., 2001; Schiefelbein, 2002; Schultz et al., 2010) and displays well-defined fault segments and
relay zones between those segments. It also has excellent cross-sectional and map-view field exposure (Fig.
1), so the relay zone near Orderville is an excellent natural laboratory to investigate the coupled evolution of
segmented normal faults and fracture systems. In this study, we use 3D computer numeric modeling to
investigate the evolution of a simple segmented fault system.
Basin and Range Colorado Plateau
transiƟon zone
BASIN AND RANGE
PROVINCE
N
Grand Wa
sh
Normal faults often form as segments in an en echelon pattern (e.g., Ferrill et al., 1999; Peacock, 2002;
Schiefelbein, 2002), and as the segments propagate toward each other, many subsidiary structures form
between the segments, including joints, fractures, and other faults (e.g., Ferrill et al., 1999). We can use
these subsidiary structures to inform us about the evolution of these segmented boundaries and offer insight
about how the local stress field proximal to the fault segments may have evolved during fault propagation. In
addition, because faults, fractures, and joints increase permeability, there is greater potential for fluid flow at
segment boundaries, so the results from this study may inform our understanding of subsurface flow of
natural resources including water, oil, and natural gas.
STRESS E-W (MPa)
VOLUME DILATION
38.5
0.0610
Dilation is a proxy for opening-mode fracturing. Dilation is the volumetric strain each location on the
observation surface has been subjected to. Positive values (”warm” colors) represent volume gain and
negative values (”cold” colors) volume loss. In addition to the stress field, dilation values are affected
by Poisson's ratio (the less rigid the lithology, the more likely that the material will simply deform instead
of actually opening up space, or dilating). The scales range from -0.0176 to 0.0186 in Model 1 to -0.2043
to 0.1229 in Model 4.
In the underlapping model, the highest dilation values are in the hanging wall, adjacent to the fault, and
the lowest values are in the footwall, also adjacent to the fault. The dilation values approach zero as you
move toward the fault tips or move east or west, away from the fault. As the fault tips propagate toward
each other in Models 2 - 4, the dilation values between the northern fault and the southern fault begin to
link and interact. The high values on the hanging wall and the negative values on the foot wall link when
the fault tips overlap, curving toward each other and acting as one fault.
MEAN STRESS (MPa)
28.7
705.7
259.3
18.9
0.0419
West
Stewart
Canyon
relay ramp
East
Stewart
Canyon
relay ramp
Stewart Canyon
Central
Stewart
Canyon
relay ramp
Orderville
Orderville
relay ramp
N
Red Hollow Canyon
Elkheart
exposure
1 km
Figure 2. Structure map of the fault network within the
Orderville study area. Thick yellow lines indicate faults that
accommodate significant (>100 m) displacement or play an
important role in fault linkage and displacement transfer.
Relay ramps that are part of this proposed study are labeled
and dip to the NNE. Faults shown here are based primarily
on Schiefelbein (2002).
2.6
0.0310
195.5
535.4
131.8
-7.2
0.0201
84.0
450.3
-20.2
36.2
0.0065
365.2
-0.0071
280.0
-33.2
-11.7
-46.3
-59.5
194.9
-123.2
109.8
-187.0
-0.0207
-59.3
In a fault zone where en echelon fault segments overlap, subsidiary structures
such as relay ramps, minor linking faults, and fractures form. We focus on the
fault and fracture system developed near Orderville, Utah (Fig. 2), where several
relay ramps formed and evolved as the major fault segments propagated along
strike as slip magnitude increased (Schiefelbein, 2002). Relay ramps
accommodate differences in vertical displacement between overlapping normal
fault segments (e.g., Ferrill et al., 1999; Schiefelbein, 2002).
The displacement along a normal fault is related to the distance between fault
segment tips (Ferrill and Morris, 2001; Schiefelbein, 2002) as shown in Figure 3.
The displacement of a fault at a given point depends on its overlap with other
faults. More overlap leads to less displacement accommodated by each fault, as
the displacement is shared (Ferrill and Morris, 2001). Importantly, as fault
segments propagate past each other, they affect the surrounding stress field,
promoting the formation of subsidiary structures that may not be compatible with
the regional stress field (e.g., Crider and Pollard, 1998; Peacock, 2002).
Importantly, changes in strike-parallel vertical displacement along a fault varies
greatly according to what subsidiary structures surround the fault in three
dimensions.
620.6
9.2
The Sevier normal fault has accommodated extension across the transition zone from the Basin and Range Province
to the relatively stable Colorado Plateau since the Miocene (e.g., Reber et al., 2001). The fault is the northern portion
of the Toroweap-Sevier fault, which extends from northern Arizona to southern Utah (Fig. 1), and it is likely that many
segments of the Sevier fault reactivate older, high-angle contractional Laramide-age structures (e.g., Schiefelbein
and Taylor, 2000). Previous workers have demonstrated that the steeply west- dipping fault zone is segmented in
map view, with variations in the geometry of the linkages between normal fault segments (e.g., Davis, 1999; Schiefelbein, 2002; Doelling, 2008).
-72.4
-0.0343
45.9
-250.7
-85.4
-17.9
-0.0479
-314.5
-98.5
-103.1
-378.2
-0.0616
-111.5
-0.0752
VOLUME DILATION
Vertical Displacement (defined in Modeling: Propagating Normal
Faults) gradients are highest where a single fault accommodates all displacement, with shallower gradients associated with the multi-fault sections of the system.
Figure 6. Map of a por on of the
Sevier fault zone near Orderville,
Utah (modified from Schiefelbein,
2002). See Fig. 2 for more detailed
fault map.
Coulomb Stress is the value that describes how much closer a fracture
is to the failure envelope than in its previous state. The Coulomb Stress
map reveals that Coulomb stress is highest where one fault is accomodating the strain, suggesting that fractures will be closer to failure at
these locations. Coulomb stress is near zero where multiple faults are
accomodating strain.
MEAN STRESS (MPa)
2356.4
652.7
2839.5
38.6
0.0852
1957.2
454.5
1729.1
341.3
1500.9
228.0
16.5
Figure 3. (a) Schema c block diagram illustra ng
an extensional relay ramp. (b) Distance versus
displacement diagram for the profile of line XY in
(a). (c) Distance versus fault-cutoff eleva on for
the profile XY in (a) (Ferrill and Morris, 2001).
169.2
1.7
0.0605
2271.6
-20.4
109.1
0.0358
-49.9
49.1
114.8
1272.8
0.0161
Methods: propagating normal faults
-0.0086
-138.5
-86.0
-168.3
-0.0333
1135.8
588.3
-161.1
-281.5
-0.0580
We used numerical modeling and outcrop-based observations to investigate the evolution of a simple en echelon normal fault system. We utilized 3D modeling using Move2016 software to test different magnitudes of displacement and fault propagation. We can compare our model results to detailed field observations of the structures in the study area (Simoneau et al., 2016), including structural measurements of fractures, detailed field
sketches, and photographs of the joints, fractures, and faults in the Sevier fault zone near Orderville, Utah.
360.2
-197.5
-236.1
-394.7
132.1
-0.0827
-227.0
567.9
-507.9
-39.1
-256.5
-0.1074
-267.2
-0.1321
-495.3
-762.7
-476.3
MODEL 4: OVERLAPPED SEGMENTS (PRESENT DAY), 1000 m DISPLACEMENT
DISPLACEMENT (m)
VOLUME DILATION
STRESS E-W (MPa)
110.9
0.1229
62.0
0.0901
VOLUME DILATION
MEAN STRESS (MPa)
3266.6
2782.5
We analyzed four models using the Fault Response Modeling module to record spatial distribution of vertical
displacement, dilation, stress in the east-west direction, and mean stress as the fault zone evolved. We use
these model results (Modeling: propagating normal faults) to interpret changes in fault transfer zone properties over time. Note that scales vary for each image displayed in the Modeling section. In the images, the fault
tips are emphasized by small white dots, and the zero value on the scale bar is indicated by a white arrow.
The models also show a significant change in the stress field adjacent to the fault segments as displacement and lateral propaga on increase. Importantly, both stress and strain are greatly amplified adjacent to and between the
fault segments and quickly diminish with distance from the fault system. These results support a model where interac ng, laterally propaga ng fault segments constantly change the stress field between them, impac ng the forma on and propaga on of minor features such as joints and shear fractures.
-339.6
In our preliminary modeling of the more complex Orderville Fault Network, we have difficulty interpre ng our results in the context of fault and fracture theory. Future efforts will be focused on: 1) including the linking faults that
help transfer slip; 2) adjus ng the interac on of faults at depth within the models; 3) carefully adjus ng the slip tapering associated with each fault to more closely represent the total offset; and 4) extending the boundary of our
model to remove poten al edge effects.
STRESS E-W (MPa)
0.0625
865.7
0.0416
711.4
991.1
References
754.4
3.3
0.0574
0.0247
2459.7
2136.9
619.2
0.0207
582.8
1814.2
-65.2
325.6
-143.5
1491.4
-0.0407
1168.6
213.4
-0.0734
Peacock, D. C. P., 2002, Propagation, interaction and linkage in normal fault systems: Earth-Science Reviews, v. 58, p. 121-142.
Reber, S., Taylor, W. J., Stewart, M., Schiefelbein, I. M., 2001, Linkage and reactivation along the northern hurricane and sevier faults, southwestern Utah: Utah Geological Association Publication, v. 30 – Pacific
-0.0210
Section American Association of Petroleum Geologists Publication GB78.
197.0
112.0
10.5
-182.6
Ferrill, D. A., and Morris, A. P., 2001, Displacement gradient and deformation in normal fault systems: Journal of Structural Geology, v. 23, p. 619-638.
Ferrill, D. A., Stamatakos, J. A., and Sims, D., 1999, Normal fault corrugation: implications for growth and seismicity of active normal faults: Journal of Structural Geology, v. 21, p. 1027-1038.
348.7
-0.0080
-104.3
Crider, J., and Pollard, D., 1998, Fault linkage – three-dimensional mechanical interaction between echelon normal faults: Journal of Geophysical Research, v. 103, p. 24,373 – 24,391.
Davis, G. H., 1999, Structural Geology of the Colorado Plateau Region of Southern Utah, With Special Emphasis on Deformation Bands: Geological Society of America Special Paper 342.
454.2
483.9
-0.0002
Figure 5. Oblique view of Model 2 from below
the observation surface. This model shows the
faults extending to a depth of 8000 m and
shows fault slip tapering on the fault plane
(purple = highest slip magnitude, blue = lowest
slip magnitude).
Results from modeling the two-fault propaga ng normal fault system are consistent with what we observe in the
field (Simoneau et al., 2016). The displacement-propaga on models show the forma on of a relay ramp between
overlapped fault segments, a feature documented at several loca ons near Orderville, Utah. The fractures in the
models at different stages of fault overlap and displacement are systema c and strongly affected by the posi on rela ve to the transfer zone. Fractures are less systema c near the fault ps in the footwall of the fault system. This
is also seen in the field, where joint orienta ons in the hanging wall and within the transfer zone display more systema c joint orienta ons that can be related to the models presented here, and fractures in the footwall of the fault
zone become significantly less systema c both in our models and in the field (Simoneau et al., 2016).
32.6
-26.1
The Models
We varied overlap and fault displacement magnitudes for the four models presented here. Model 1 is an underlapped normal fault pair with a displacement of 100 m. Model 2 shows the same faults, which have propagated
along strike until the fault tips are even with each other, and each fault displays 400 m maximum displacement.
Model 3 tests a moderate fault overlap and 700 m displacement. Model 4 shows the present-day fault pair (significant overlap) and 1000 m displacement.
Stress in the East- West Direction (defined in Modeling: Propagating Normal Faults)
shows high compressional stresses within the overlap zone and near the southern
boundary of the model, with tension at other locations.
-311.2
-621.2
-386.2
-293.4
Opening-mode fractures (defined in Modeling: Propagating Normal Faults) are most
complex within the segment overlap zone near Stewart Canyon and become more
systematic and predictable with increasing distance to the east and west of the fault
network.
-55.0
816.5
-168.0
In Move2016, we created models of a segment pair at different stages of overlap (Fig. 4) similar to the scale of
those observed along the Sevier fault zone near Orderville. We developed models to test four different stages
of fault overlap and normal fault displacement. For each model, we held the following attributes constant: our
observation surface sits at a depth of 1000 m below the atmosphere-rock interface; the synthetic normal fault
segment pair dips at 72°; and each fault reaches a maximum fault depth of 8000 m. For all models, the material
properties include a Poisson’s Ratio of 0.15 and a Young’s Modulus of 10000 MPa to simulate the characteristics of the Navajo sandstone (Rogers and Engelder, 2004). Figure 5 reveals how fault displacement varies on
each fault plane, with maximum displacement at 1500 m depth and displacement decreasing toward the fault
tips.
1703.7
-10.9
29.9
1044.6
-108.9
Volume Dilation (defined in Modeling: Propagating Normal Faults) displays positive
values in most areas to the west of the network, with the highest values associated
with the fault to the northeast of the overlapping fault system.
Interpretations
COULOMB STRESS(MPA)
DISPLACEMENT (m)
274.3
-79.4
Figure 4. Oblique view of Model 4 displaying
mean stress on the observation surface. This
model shows normal faults with significant
overlap. Model dimensions are 10000 m
(E-W), 18000 m (N-S), and 9000 m (vertical).
Observations
Fault used in modeling
N
ORDERVILLE FAULT NETWORK (PRESENT DAY)
1000 m DISPLACEMENT
STRESS E-W (MPa)
0.1149
75.5
Opening-mode fractures (shown as short blue lines on all models) form perpendicular to 3 (the
minimum stress) and are strongly affected by fault segment propagation and displacement. The fracture
orientation maps shown for these models are constant for each overlap- displacement value. Importantly,
near and within the transfer zone between fault segments, fracture orientation changes with increasing
displacement and fault segment overlap, revealing obvious perturbations in the local stress field.
-457.9
-273.3
MODEL 3: OVERLAPPED SEGMENTS, 700 m DISPLACEMENT
DISPLACEMENT (m)
Methods
We developed a simplified three-fault
model to test the effects of fault slip within
the complexly-deformed Orderville fault
network. The map-view fault pa ern was
based on Figures 2 and 6, which show the
key fault segments within the network. We
used similar constraints as the simplified
two-fault model. We included 1000 meters
of dip-slip displacement and similar slip tapering. Our model includes the segments
shown in red in Fig. 6 but does not include
any linking faults.
-188.2
-124.6
Mean Stress is [(max + min)/2] and is measured in MPa. Negative values indicate a tensional stress field,
and positive values indicate a compressive stress field. The scales range from -128 to 85 MPa in Model 1
to -700 to 990 MPa in Model 4. Mean stress is greatest on the east side of the fault segments (hanging
wall), with negative stress values on the west side of each segment (footwall). The negative stresses
associated with each segment connect in Model 2, but the positive stress fields associated with each
segment do not yet interact. By Model 4, where displacement and overlap are greatest, the stresses from
the northern and southern faults interact, and the surrounding stresses are close to zero.
Preliminary modeling of the Orderville fault network
339.0
790.9
Stress in the East-West Direction displays the magnitude of east-west stress created by fault motion,
with positive values indicating compression and negative values tension. The scales range from -67 to 300
MPa in Model 1 to -768 to 3267 MPa in Model 4. Stress in the east-west direction is greatest along and
adjacent to the faults, with values approaching zero as you move away from the faults or toward the fault
segment tips. The stress field is greatly affected by the relative overlaps of the faults; as the faults
propagate toward one another and displacements increase, the stresses increase and spread along more
of each fault segment. The most compressive stress is on the fault, favoring the footwall, with the most
tensional stresses on the hanging wall. The stress fields interact when the segments overlap and join in
Model 4.
Rogers, C. and Engelder, T., 2004, The feedback between joint-zone development and downward erosion of regularly spaced canyons in the Navajo Sandstone, Zion National Park, Utah, In Cosgrove, J. W., and
-0.0419
845.9
94.2
Engelder, T. (Eds.), The initiation and propagation of joints and other fractures: Geological Society Special Publication 231, London, UK, p. 49-72.
Simoneau, S., Surpless, B., and Mathy, H., 2016, The evolution of subsidiary fracture networks in segmented normal fault systems: Geological Society of America, abstracts with program, Denver, Colorado.
-124.7
-221.7
-0.1062
200.3
-260.9
-260.0
-0.0663
-122.5
-395.2
-339.1
-378.3
-530.5
-265.9
Acknowledgments
-445.2
-0.2043
-768.0
-699.6
Schultz, R. A., Okubo, C. H., and Fossen, H., 2010, Porosity and grain size controls on compaction band formation in Jurassic Navajo Sandstone: Geophysical Research Letters, v. 37, L22306.
Thelin, G.P., and Pike, R.J., 1991, Landforms of the Conterminous United States - A Digital Shaded-Relief Portrayal: U.S.G.S. Geologic Investigations Series I – 2720.
-0.0906
-0.1716
Schiefelbein, I. M., 2002, Fault segmentation, fault linkage, and hazards along the sevier fault, southwestern Utah: MS Thesis, University of Nevada, Las Vegas.
Schiefelbein, I. M. and Taylor, W. J., 2000, Fault Development in the Utah Transition Zone and High Plateaus Sub-province: Geological Society of America Abstracts with Programs, v. 32, no. 7, p. A431.
-137.3
-0.1389
-300.0
-8.7
523.1
-0.1115
-420.2
This research has been supported by the Trinity University Department of Geosciences' Edward C. Roy fund for student activities and research and the Department of Geosciences at Trinity University. Special thanks to Bauer’s Canyon Ranch RV Park in Glendale,Utah, for
their hospitality during our fieldwork.