Professional Documents
Culture Documents
PII: S0191-8141(15)00033-4
DOI: 10.1016/j.jsg.2015.02.002
Reference: SG 3184
Please cite this article as: Stead, D., Wolter, A., A critical review of rock slope failure mechanisms: The
importance of structural geology, Journal of Structural Geology (2015), doi: 10.1016/j.jsg.2015.02.002.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
PT
7 Abstract
RI
8 Geological structures such as folds, faults, and discontinuities play a critical role in
9 the stability and behaviour of both natural and engineered rock slopes. Although
SC
10 engineering geologists have long recognised the importance of structural geology in
U
12 and theory into all stages of engineering projects. We emphasise the importance of
13
AN
structural geology to slope stability assessments, reviewing how structures control
18 control
EP
19 1. Introduction
C
23 the exploitation of large open pit mines. Design in such large-scale rock masses
25 tectonic scale. In this paper, we critically review rock slope failure mechanisms with
27 all stages in the slope characterisation or design process. We discuss the impacts of
28 broad structural feature types on rock slopes, highlighting important aspects using
PT
31 incorporated into numerical modelling of slopes, a useful technique in slope design.
32 Our objective is to demonstrate the critical role of structural features and processes
RI
33 in controlling rock slope stability and failure type, style, and mechanism.
SC
34 2. Effects of structural geology on rock slope stability
35 Structural features, such as folds, faults, and discontinuities, control rock mass
U
36 (i.e., intact rock dissected by discontinuities) behaviour and contribute to either the
37
AN
stabilisation or destabilisation of rock slopes, depending on their orientations and the
41 mechanism failure (Figure 1 and Figure 2). Stead et al. (2006) referred to the
43 section, we review the importance of tectonic environment and damage and common
44 brittle and ductile structures related broadly to lithology, with examples drawn from
C
45 the published rock slope literature. Table 1 summarises the structural and lithological
AC
46 features of the cited case studies and additional examples. Specific structural
47 features are discussed in relation to each lithological rock type; however, this does
51 determine if and how it fails; in situ stress conditions are important to consider when
52 assessing rock mass behaviour. Hoek et al. (2009) discussed the role of in situ
53 stresses in influencing the design of large open pits. They demonstrated numerically
PT
54 that in situ stresses are typically not significant as compared to uncertainties related
RI
56 pressures within the rock masses of a given pit slope. Thus, gravitational stress
SC
57 usually dominates slope stability in open pits. The exception is in areas of high
58 horizontal stress (compressional regime), where stresses can influence open pit
U
59 performance and should be considered. Kinakin and Stead (2005) demonstrated the
60
61
AN
importance of in situ stress conditions in the formation of sackung-type slope failures
63 geomorphological processes and in situ stress conditions and slope geometry. They
D
64 demonstrated that these factors influence slope failure type and behaviour. The
TE
65 World Stress Map (Zoback, 1992) is a useful tool in assessing general in situ stress
66 conditions.
EP
67 Past and present seismicity is also important to slope stability. Not only does each
68 earthquake have the capability of triggering slope failures, but the cumulative effect
C
69 of regional seismicity may also damage and weaken slopes. Moore et al. (2011), for
AC
70 example, illustrated the production of slope amplification effects at the 1991 Randa
72 simulated stability of large open pits in relation to seismic activity and showed that
73 open pits are less susceptible to seismically induced failure than natural slopes.
74 Brian et al. (2014) examined the role of microseismicity in producing fatigue and
ACCEPTED MANUSCRIPT
76 limited and may be episodic; it typically also only occurs in areas of pre-existing
PT
79 deformation and strength of rock masses in compression and extension. The most
RI
80 common criterion in rock slope analysis remains the simple, linear Mohr-Coulomb
SC
82 (1980), is also commonly applied to represent the strength of rock masses. Recently,
U
84 brittle fracture through tensile failure and spalling at low confinement. Leith et al.
85 AN
(2014a and b) applied this criterion to two Alpine valleys to investigate the effects of
88 and Eberhardt (2013) highlighted the relevance of a trilinear failure envelope to rock
D
89 slope damage.
TE
91 damage associated with geological structures (e.g., Shipton and Cowie, 2003; 2008).
92 A rock slope may be subjected to many forms of damage including tectonic damage,
C
94 and damage due to alteration or weathering. Only recently has the importance of
95 rock damage been fully realised in the rock slope stability community; much of this
96 has been due to development of both rock slope mapping technologies and
98 Brideau (2010) discussed the importance of damage in rock slope instability with
99 reference to several major landslides including the 1965 Hope, 2002 McAuley Creek,
ACCEPTED MANUSCRIPT
100 and 2007 Chehalis Lake landslides in British Columbia, and the South Peak, Turtle
101 Mountain instability in Alberta, Canada. Pedrazzini et al. (2011) demonstrated the
102 varied rock mass damage associated with folding at Turtle Mountain and its effect on
103 instability. More recently, Agliardi et al. (2013) and Stead and Eberhardt (2013)
PT
104 provided more in-depth discussions on the importance of rock mass damage to
105 slope stability. We emphasise the importance of tectonic damage on rock masses.
RI
106 This damage reduces the strength of rock masses and also provides the kinematics
SC
107 required for slope displacements; we suggest that tectonically induced damage has
U
109 other forms of damage.
110 AN
2.2. Importance of structures in sedimentary rock slopes
111 Sedimentary rock masses typically behave anisotropically due to bedding planes,
M
112 which act as planes of weakness along which slip may occur. Closely spaced
D
113 bedding planes dipping into the slope may lead to toppling slope failures. Such
TE
114 failures are described in surface coal mines, where repeated strata of arenaceous
115 and argillaceous rock masses dip into the slope (Benko, 1997). Stability of these
EP
116 slopes is controlled by the relative spacing of bedding and cross joints, the presence
117 of lateral release joints, and the interbed shear strength. The influence of tectonic
C
118 activity may lead to a reduction in the interbed shear strength with an important
AC
120 Originally horizontal beds are frequently uplifted, tilted, and folded. The folding of
121 sedimentary rock sequences has numerous and varied effects on both natural and
122 open pit slope stability (Figure 3). The orientation of the topographic surface or
123 excavation face with respect to bedding is of major significance. Knowledge of the
ACCEPTED MANUSCRIPT
124 fold type and orientation, and its structural complexity, are important in realistic rock
125 slope modelling. We suggest that the features of a conventional fold classification,
126 such as Fleuty (1964) or Ramsay (1967), should be considered with respect to how
127 they influence rock slope stability (Figure 4). Horizontal folds may be associated with
PT
128 simple translational failure (sliding of rock blocks in the true dip direction) on one limb
129 and toppling mechanisms on the other fold limb, depending on their orientation with
RI
130 respect to a given slope face (Figure 4i-iii). Steeply plunging folds (Figure 4iv-vii)
SC
131 may result in complex translational and rotational slope displacements in both planar
132 and toppling modes (Brideau, 2005; Yan, 2008). The bluntness and tightness of a
U
133 fold will not only influence the kinematics of slope failure but may also be indicative
134
135
AN
of the degree of tectonic damage (reduced rock mass quality) sustained by the rock
mass in the form of both fracturing and interbed slip (Figure 5). Open folds may show
M
136 the typical transverse, longitudinal, and conjugate discontinuities associated with
137 folding, whereas tight and isoclinal folds may develop foliation and cleavage, a sign
D
139 similar/shear type will also influence kinematics and potential failure block geometry.
141 combined effect on rock mass kinematics and strength. Folded rock units with
142 plunging hinge lines, for example, have likely sustained more deformation events
C
143 than units with horizontal hinges. Folding in two directions and the presence of
AC
144 interference folds has been shown to influence the complexity of slope deformations
145 within landslides and, in combination with natural release surfaces, may explain
146 commonly observed multiple block-type movement. Figure 6, for example, shows the
147 sliding scar morphology of the 1963 Vajont Slide in Italy. This surface is affected not
148 only by two generations of regional folds, producing complex, metre- to decametre-
ACCEPTED MANUSCRIPT
150 (Thiessen and Means, 1980) type interference patterns (Figure 6b; Massironi et al.,
151 2013), but also by concretions on the centimetre scale. Folding plays a vital role in
152 providing what are referred to as geologic release surfaces for rock slope failures on
PT
153 the sides (lateral release), at the back (rear release), and on the base (basal plane)
154 of failure blocks. Discontinuity sets associated with folding (i.e. transverse,
RI
155 longitudinal and conjugate) may play a major role in releasing the blocks that slide
SC
156 along folded bedding surfaces – such release surfaces provide the necessary
157 kinematic freedom for slope failure. The association of kinematic rock slope failure
U
158 modes with anticlinal folding has been discussed by Badger (2002), who described
159
160
AN
the relationship between failure (planar/toppling) and the relative location of the rock
slope and discontinuity sets commonly associated with folded, bedded units. Figure
M
161 7 illustrates how three possible configurations of the folded bedding (S0) or primary
162 discontinuity set (S1) and secondary discontinuity set (S2) can produce different
D
163 sliding failures. Another example of structural analysis associated with natural slope
TE
164 failures is that conducted by Humair et al. (2013) on the 1903 Frank Slide area in
165 Alberta, Canada (Figure 8); they demonstrated the importance of the Turtle Mountain
EP
166 Anticline orientation, dimensions, and associated fractures to slope stability. Relative
167 slope and fold orientations, as well as damage intensity, affect slope stability and the
C
169 Fault planes often form sliding surfaces or release surfaces and can be associated
170 with the steepening of bedding (drag folding) that may induce slope instability. For
171 example, Figure 9a shows a failure in a coal mine, where a fault acted as a rear
172 release plane for a translational failure that occurred along drag-folded strata
173 adjacent to the fault. Figure 9b shows the fault exposed within the mine and the drag
ACCEPTED MANUSCRIPT
174 folding. As discussed previously, damage zones associated with faults may similarly
175 result in a reduction in rock mass quality influencing slope stability. Limited work in
176 rock slope engineering has been undertaken to date considering fault contact
177 relationships and their influence on rock slope instability; this includes not only
PT
178 damage but also fault linkage, drag mechanisms, fault seals, shale smear, the nature
RI
180 Roughness exerts a primary control on rock slope failures and exists at all scales
SC
181 from primary and secondary small-scale asperities to large-scale undulations. Focus
182 in rock engineering has been on the use of roughness measures such as the Joint
U
183 Roughness Coefficient, JRC (Barton 1976, 2011; Barton and Choubey 1977). These
184 AN
roughness measures have been linked to empirical approaches for assigning shear
185 strength to joints. The JRC, for example, is used in the following equation:
M
tan log
D
186 where τ is shear strength along the discontinuity plane, σn is the normal stress on the
TE
187 plane, φr is the residual friction angle, and JCS is the Joint Compressive Strength.
EP
188 The importance of sedimentary and tectonic structures on discontinuity surfaces has
190 roughness, such as slickensides and polishing, have been recognised as indicators
AC
191 of residual strength and the need for testing the shear strength in the appropriate
192 direction with respect to surface structures noted. Cruden and Krahn (1978)
193 described the presence of slickensides on folded bedding surfaces, as well as minor
195 We suggest that sedimentary structures, such as sole marks, flute casts,
196 concretions, and ripple marks, also may have a significant influence on local or
197 micro- to meso-scale roughness, and hence affect global rock slope behaviour. For
198 example, concretions or ripple marks may increase bedding plane frictional
PT
199 resistance and in some cases even induce dilation and the need for shearing
200 through asperities during failure. Such features may act as rock bridges along
RI
201 potential sliding surfaces. Rock bridges are zones of intact rock classified as either i)
SC
202 in-plane rock bridges on a discontinuity surface, or ii) out-of-plane rock bridges
U
204 2.3. Importance of structures in igneous rock slopes
205
AN
Rock slope failures in igneous rock types are highly varied, depending on the
206 topography, in situ stress conditions, and degree of weathering or alteration. Given
M
207 the high strength of granites and similar igneous rock masses, it is not surprising that
D
208 it is invariably the structures within the igneous rock that form the weak link and
TE
209 control slope instability. One example of structural controls in igneous rock types is
210 the effect of sheet, or exfoliation, joints. Hencher and Richards (1982), Hencher et al.
EP
211 (2011), and Ziegler et al. (2013, 2014) discussed the importance of sheet joints,
213 focussing on joint initiation and propagation in slopes, as well as their effects on
AC
214 slope stability and determining in situ stresses from fracture patterns. As sheet joints
215 generally parallel the topography, instability may involve undercutting by excavation,
216 failure by linkage with other structures and, in the case of high-stress environments,
218 When considering structurally controlled stability of igneous rock slopes the
221 Non-persistent structures must be taken into account as the strength of rock bridges
PT
222 between discontinuities is high and therefore even small percentages of rock bridges
223 will increase the overall strength of the slope resulting in a higher factor of safety.
RI
224 Consideration should be given to step-path-type failures involving blocks moving on
SC
225 multiple fractures with possible intact brittle rock fracture. Characterisation of fracture
226 networks in igneous rocks for use in subsequent numerical models (see below)
U
227 requires considerable care and often the combined use of field mapping, remote
231 They also influence groundwater flow, which is an important factor in the alteration of
D
232 igneous rock masses as well as in pore pressure conditions in slopes. At Diavik
TE
233 Diamond Mines in Northwest Territories, Canada, for example, a major fault zone
234 dissects two kimberlite pipes and directs groundwater flow into the existing open pit
EP
235 (Tuckey, 2012; Vivas Becerra, 2014). Considerable work has been undertaken
236 recognising the influence of structure on hydrogeology and pore water pressure at
C
238 Although usually not a dominant factor in slope stability of igneous rock masses,
239 micro-scale features should also be considered. The different strengths of various
240 minerals included in a particular rock mass, as well as grain and lattice distortions
241 and deformations, affect rock mass weathering and strength. Stress-induced fracture
242 propagation and coalescence in high slopes may ultimately form persistent
ACCEPTED MANUSCRIPT
243 structures along which sliding can occur. Considerable research is ongoing,
PT
247 Numerous large, catastrophic rock slope failures have occurred in metamorphic
RI
248 lithologies. Foliations, cleavage, and other deformational features strongly influence
249 stability; the metamorphic facies level of a given rock mass can also have a
SC
250 significant influence on failure by altering rock mass strength. Catastrophic failures
251 have occurred in both high-grade metamorphic rocks such as gneisses and in highly
U
252 foliated slates, phyllites, and schist. Dominant foliations can be parallel, oblique, or
253
AN
opposite to the sliding direction, with mechanisms consequently involving planar
256 Complex folding of foliations in the dip and strike directions of a slope adds
TE
258 An example of a major landslide in metamorphic gneiss terrain is the 1991 Randa
EP
259 rockslide sequence in Switzerland (Figure 10) (Eberhardt et al., 2004; Willenberg et
260 al., 2008; Gischig et al., 2009; Löw et al., 2012). This rockslide has been the subject
C
261 of extensive research over the last decade and is an excellent example of the
AC
262 influence of geological structure on rock slope failure. Geological structures form the
264 Another example is the Åknes rock slope in Norway (Table 1) (Ganerød et al.,
265 2008, Grøneng et al., 2011; Blikra, 2012). This active slope instability is located in a
266 highly deformed metamorphic region comprising foliated schistose rock types on the
ACCEPTED MANUSCRIPT
267 western slope of the Storfjorden fjord and presents a major risk of landslide-induced
268 tsunami. The influence of the metamorphic structures in controlling the active
269 movement of the rock slope has been shown through geophysical and borehole
270 methods, geotechnical instrumentation and numerical modelling. The basal surface
PT
271 of the landslide is stepped, following discontinuities separated by intact rock bridges.
272 An interesting aspect of this slide is the multiple blocks involved in the movement
RI
273 with different rates and movement directions; these may be related to three-
SC
274 dimensional variations in the geological structure along the basal sliding surface.
275 Such movement of blocks is not uncommon and shows the importance of
U
276 understanding three-dimensional geological structure.
277 AN
The Downie Slide in British Columbia, Canada, is the largest prehistoric landslide
278 in North America with an estimated volume of 1.5 billion m3. The toe of this landslide
M
279 was submerged by a reservoir impounded by the Revelstoke Dam, and hence this
280 landslide has been extensively investigated. A major slope dewatering scheme was
D
282 modelling of this landslide is presented in Kalenchuk (2010). Like Åknes, this slide
283 comprises multiple blocks and is located in metamorphic mica schist, gneiss, and
EP
284 quartzite.
C
286 provided in Newman (2013) and Clayton (2014). An excellent discussion on toppling
289 consider both primary and secondary slope failure mechanisms. A useful example is
290 presented by Brideau et al. (2006) at the East Gate slope in Glacier National Park,
291 British Columbia, Canada. The upper unstable rock slope is situated within phyllites
ACCEPTED MANUSCRIPT
292 and involved pseudo-circular rock mass failure influenced by foliations including
293 crenulation cleavage. Although this failure was at considerable distance from a major
294 highway (Trans-Canada Highway) at the foot of the mountain slope, it led to the
295 gradual accumulation of friable phyllitic failure debris. Over time, this debris has
PT
296 remobilised as debris flows that have travelled downslope and caused significant
RI
298 Deep-seated gravitational displacements are an important variant of deforming
SC
299 rock slopes, which occur in multiple rock types and are particularly common in post-
U
301 topographic features such as antiscarps, half-grabens, double-crested ridges,
302 AN
trenches, and slope toe bulging, providing evidence of previous or active movement.
303 Although beyond the scope of the current paper, it is important to emphasise that
M
304 DSGSDs are often located in foliated rock slopes; mechanisms of displacement may
305 be highly variable, including translational sliding, circular and toppling failures. The
D
306 La Clapière landslide in France, for example, is part of a larger DSGSD complex
TE
307 (Figure 11) and has shown varied behaviour over the past century (Vengeon et al.,
310 Various techniques have been applied to rock slopes to characterise and analyse
AC
311 them, ranging from field surveys to remote sensing. Traditional rock slope
312 investigations often commence with field observations of the rock mass, including
314 Lithology, degree of weathering, and intact rock strength are determined through
315 observation and in situ or laboratory tests. Stress conditions may be measured on
ACCEPTED MANUSCRIPT
316 site using overcoring, flat jack, or hydraulic fracturing tests to provide direction and
317 magnitude of stress, or using earthquake focal mechanism analysis, breakout and
318 failure plane orientation observations, or stress relief measurements to determine the
PT
320 rare and remains a challenge, as a wide variation in results are common.
RI
321 Discontinuity surveys using line surveys or circular/rectangular windows allow the
SC
323 characteristics, such as persistence, intensity, roughness, aperture, infill and
324 seepage; and the delineation of structural domains. Recently, Hencher (2013) stated
U
325 that, although the method of discontinuity surveying is simple and straightforward,
326 AN
the definition of discontinuities is not. He suggested a revised classification scheme
327 of discontinuities that includes an estimation of their tensile strength. This indicates
M
328 explicitly that not all discontinuities have the same effect on rock mass strength and
330 By considering specific slope (or design sector) orientations, the potential stability
332 These methods have changed little since their introduction, and are summarised in a
333 large number of publications including Richards et al. (1978), Priest (1980), Hoek
C
334 and Bray (1981), and Wyllie and Mah (2004). Computer codes such as DIPS
AC
335 (Rocscience, 2014) are now routinely used in kinematic analysis of rock slopes to
336 assess potential simple planar, wedge and toppling (flexural and direct) instability
337 modes (Figure 12). Limitations of this method can include insufficient consideration
338 of discontinuity persistence, spacing, and other factors controlling discontinuity shear
340 produce either conservative or, more importantly, unsafe stability assessments. A
ACCEPTED MANUSCRIPT
342 structures, such as faults/shear zones and persistent bedding planes, on slope
343 stability and then incorporate instability mechanisms related to joint sets. Ground-
PT
345 suggested. Most kinematic analyses are restricted to simple planar, toppling or
346 tetrahedral (+/- tension crack) geometries. Often potential failures will have more
RI
347 complex geometries involving multiple basal, rear and lateral release surfaces. As an
SC
348 example, the truncation of a tetrahedral wedge by basal surfaces such as bedding
U
350 analysis of rock slopes, with the exception of direct toppling, is also generally limited
351
352
AN
to consideration of translational failure mechanisms with limited consideration of
rotational moments or intact rock fracture (Hungr and Amann, 2011). The most
M
353 recent versions of rock slope wedge analysis programs (e.g., Swedge (Rocscience,
354 2014)) allow for limit equilibrium analysis of pentahedral wedge geometries. Figure
D
355 13 indicates two possible geometries, each with basal and lateral release planes and
TE
356 Figure 13b with a rear tension crack. The use of kinematic analysis methods in slope
357 stability assessments are suggested to be in general more suited to surface mine
EP
358 bench scale or highway road cuts than to large-scale landslides or open pit mine
361 scanning, digital photogrammetry, InSAR and thermal imaging are allowing, in
363 slopes (Jaboyedoff et al. (2007), Lato et al. (2009), Sturzenegger and Stead (2009),
364 and Petley (2012)). Using such ground-based, airborne and satellite-based methods
365 provides the ability to map previously inaccessible rock slope areas, often with sub-
ACCEPTED MANUSCRIPT
367 geological mapping, is the use of Unmanned Airborne Vehicle (UAV) platforms.
369 persistence, spacing, and even discontinuity roughness (see, for example Haneberg,
PT
370 2007; Tatone and Grasselli, 2010); ductile structural features such as folds may also
RI
372 and interference patterns may be quantitatively assessed using remote sensing, with
SC
373 implications for failure initiation and propagation. If possible, field validation of remote
U
375 Finally, subsurface studies, where conducted, contribute greatly to the improved
376 AN
spatial characterisation of rock slope masses, and may include both borehole
377 logging and geophysical surveys. For example, geophysical borehole logging, 1D
M
378 borehole radar reflection surveys, 3D surface georadar, and seismic refraction
379 surveys, as well as inclinometer and televiewer data were all used in the
D
380 characterisation of subsurface structures at the Randa slope (Löw et al., 2012).
TE
381 Clayton (2014) showed the use of borehole Time-Domain Reflectometry (TDR) in
383 One of the most extensively and intensively monitored sites is the Åknes slope
C
384 (Grøneng et al., 2011; Blikra, 2012). Here, borehole data sets, such as core logs,
AC
385 measurements of water pressure and flow and penetration velocity, resistivity
386 profiles, and P-wave velocity profiles, have allowed the determination of hydrologic
387 conditions and rock mass quality, and the identification of large fractures and sliding
390 rock slope instabilities in remote mountainous areas such data are rare and often
PT
393 4.1. Limit equilibrium analysis
394 The simplest and most common form of slope analysis involves force and moment
RI
395 equilibrium and uses iterative methods of slices to calculate the factor of safety (FS =
SC
396 resisting forces (or moments) / driving forces (or moments)). The most accurate
397 methods determine the factor of safety for both moment and force equilibrium and
U
398 use different assumptions for the interslice forces (Bromhead, 1992; Duncan and
399
400
AN
Wright, 2005). The methods incorporate a searching algorithm that, in its simplest
form, determines the critical circular slip surface for a slope (lowest factor of safety),
M
401 which is in reality only valid for extremely weak rock masses or soil-like materials. To
402 allow analyses of more complex slope failure mechanisms, composite or block
D
403 searching algorithms have been developed that search for critical, non-circular block
TE
404 geometries. Facilities exist to import groundwater conditions varying from simple
406 providing the ability to consider transient slope instability. Seismic disturbance may
407 be included using a pseudo-static approach and a wide range of support may be
C
408 applied in the software. In practice the influence of geological structure often
AC
409 determines the location of the critical sliding surface which may be:
410 • Along major discrete structures, such as bedding planes (usually adversely
411 dipping or folded), faults, and shear zones. These may lead to single-plane
413 • Through the rock mass and controlled by anisotropy within the rock mass
415 • A combination of failure along discrete structures and failure through intact
416 rock.
PT
417 Recent developments have seen an increasing use of anisotropic functions in limit
RI
418 equilibrium analyses to allow for directional weakness created by pre-existing
419 geological structure (Rocscience, 2014). Modifications have also allowed the
SC
420 inclusion of rock bridges by the use of apparent cohesion and friction values to
421 represent percentage of intact rock that must be sheared through for failure to occur
U
422 (Jennings, 1970; Baczynski, 2000). Analyses have in the past been predominantly
423 AN
deterministic, providing one factor of safety for the slope, and have often been
427 methods are mostly applied in the analysis of simple tetrahedral wedge geometries.
TE
428 Elegant methods of columns have been developed to allow three-dimensional limit
429 equilibrium analyses of more complex failure surfaces, but have seen less frequent
EP
430 use in the analysis of landslides. Probabilistic analyses have become increasingly
431 common, with the need to consider the risk and consequence of slope instability and
C
432 both model and parameter uncertainty. Although limit equilibrium methods provide
AC
433 useful preliminary analysis and may be adequate for simple slope failure
434 mechanisms, they consider only forces or moments and not slope displacements,
435 and as such are unable to fully capture complex slope failure mechanisms.
438 approach, where the rock mass is treated as an equivalent medium. The rock mass
440 constitutive criterion allows simulation of varied stress-strain behaviour from simple
PT
441 elastic to elasto-plastic or time-dependent creep. The strength of the modelled rock
RI
443 compression, shear and tension. The strength is hence inherently scale-dependent
SC
444 and it is here that the influence of structure is paramount. A Geological Strength
445 Index (GSI) ranging from 0 to 100 (Figure 14) was introduced by Hoek et al. (1995)
U
446 and modified by Cai et al. (2004) and Hoek et al. (2013), and is currently used as the
447
448
AN
main method for deriving equivalent media properties of a rock mass suitable for
450 and, indirectly, the rock mass strength; both of these input criteria are controlled by
D
451 the structural geology of an area. A modification in the GSI approach was proposed
TE
452 for tectonically disturbed flysch rock masses (Marinos and Hoek, 2000) (Figure 15).
453 It is essential always to consider a slope rock mass using a structural geology
EP
454 domain approach recognizing characteristic structures within an area; such a domain
456 both again often a function of the tectonic history and structural geology. In open pit
AC
457 mining geomechanics, the approach adopted is to consider the structural geological
458 model of the slope, the hydrogeological model, and the rock mass model. These
459 models are combined in a geotechnical model of the slope or open pit (Read and
460 Stacey, 2009). An adequate structural model of the pit slope is hence an essential
462 The most common continuum numerical modelling approaches for rock slope
463 analysis are the finite element and the finite difference methods. The most popular
464 codes for each approach are, respectively, Phase2 and RS3 (Rocscience, 2014),
465 and FLAC2D and FLAC3D (Itasca, 2014). Both finite element and finite difference
PT
466 methods have the ability to incorporate discrete structures such as faults but these
467 methods are not optimal in the analysis of jointed media. Recent developments in
RI
468 the finite element code Phase2 have allowed incorporation of jointed media (e.g.
SC
469 bedding, discrete fracture networks and Voronoi tessellation) and the modelling of
470 structurally controlled rock slope failures, such as toppling, planar and biplanar
U
471 modes. Figure 16 shows examples of the use of 2D and 3D continuum codes in the
472
473
AN
modelling of landslides and rock slopes. Where discrete major structures control rock
slope instability, the use of the GSI in deriving the rock mass strength is not
M
474 recommended as the shear strength of the structures, not the rock mass, will control
475 the kinematics of slope failure. In such slopes, discontinuum methods are used to
D
478 To accurately represent jointed media and large displacements of rock slopes,
EP
480 numerical modelling methods exist and are usually grouped under “Discrete Element
AC
481 Methods”. The most commonly used in rock engineering are the Distinct Element
482 and Discontinuous Deformation Analysis methods. Stead and Coggan (2012)
483 provide a more detailed description of these methods; here, we will focus on the
484 Distinct Element Method, which is the most commonly used for modelling rock
485 slopes. A rock slope is discretized into joint-bounded blocks; the joints are assigned
486 a shear strength constitutive criterion and shear and normal stiffnesses (controlling
ACCEPTED MANUSCRIPT
487 displacement along joints). The intact rock blocks can be rigid or deformable. If the
488 blocks are treated as deformable they can be assigned a wide range of constitutive
490 options. Modelling of rock slopes using Distinct Element methods can thus simulate
PT
491 movement along major structures, such as faults, as well as the translation and
RI
493 exists, considerable challenges remain in simulating the influence of geological
SC
494 structures on rock slope deformation, including:
495 • The persistence of structures, such as joints, both along strike and dip is
U
496 usually difficult to quantify in the field. Most models routinely assume
497 AN
persistent, through-going structures – an over-simplification in practice.
498 • Where non-persistent structures exist, the engineer must consider rock
M
499 bridges.
500 • Rock masses are typically heterogeneous. Properties such as strength and
D
501 discontinuity intensity vary temporally and spatially. Shear zones, damage
TE
502 zones, and folds should be noted and treated as separate domains in
503 analysis.
EP
506 to determine which structures are important for the specific problem at hand.
AC
508 Four levels of discontinuum modelling of rock slopes, from the simplest to the
510 i. Consideration of the dominant discrete failure surfaces only. Figure 17a
512 modelling the Vajont landslide. The discrete surface in this analysis follows
513 folded bedding with a low-angle seat and high-angle rear forming a chair-
PT
514 shaped failure surface. Such an analysis allows identification of the
515 interaction mechanisms between the upper driving block and the lower
RI
516 passive block, with the formation of a transitional Prandtl prism or yield
SC
517 zone. Figure 17b shows a similar approach in the 3D Distinct Element
518 code 3DEC (Itasca, 2014), applied to the Beauregard landslide in Italy
U
519 (Kalenchuk, 2010). This model involved consideration of the primary basal
520
521
AN
failure surface, and the lateral and rear release surfaces. It is important to
524 model of the Mitchell Creek landslide in British Columbia has recently
TE
531 ii. Incorporation of the full geological structure – that is, inclusion of both
532 major structures (faults and folds) and discontinuity sets (bedding and joint
533 sets). Joint sets within Distinct Element models are simulated by defining
534 their dip (and dip direction if in 3D), spacing, and strength properties (e.g.
ACCEPTED MANUSCRIPT
538 persistent throughout the model (Figure 16a shows the different
PT
539 assumptions between continuous and discontinuous joints). In order to
RI
541 values such as dip, dip direction, joint stiffness, cohesion, friction angle,
SC
542 and dilation angle. Complex, continuously folded bedding surfaces can be
543 simulated through the use of arc commands as seen in the simulation of
U
544 the Frank Slide in Figure 17c (Benko and Stead, 1998); this figure shows
545
546
AN
the incorporation of two joint sets, discrete structures, and a major fold. An
548 bridges.
D
549 iii. Inclusion of rock bridges along discontinuity surfaces. Rock bridges are
TE
552 bridge strength to the overall strength of the discontinuity (for example,
553 Franz (2009); Brideau et al. (2012)). Variants of this approach are used in
C
554 many model types whereby the strength of the discontinuity is adjusted to
AC
556 bridges.
557 iv. Explicit incorporation of rock bridges using a Discrete Fracture Network
558 (DFN). DFNs are increasingly being used in all fields of rock engineering,
563 Associates, 2012), Move4D (Midland Valley, 2014), and FracSim3D (Xu
PT
564 and Dowd, 2010). The principal inputs required from structural mapping for
565 DFN generation are statistical data on joint trace length, joint intensity,
RI
566 joint orientation, and joint termination. Discrete fracture networks are then
SC
567 imported into geomechanical models to allow explicit representation of
U
569 important to consider the structural geology and tectonic history of an area
570
571
AN
to ensure realistic geomechanical models. It is possible to incorporate
573 wherever feasible and care must always be taken to ensure that
D
580 this subject has been extensively researched, insufficient attention has
581 been given to the structural geological aspects. Tuckey (2012) and Tuckey
583 rock bridges in large open pit rock slopes. Dowd et al. (2007) and
586 networks. Work such as the Leeds Rock Fracture Experiment on fracture
588 Hencher (2013) and Hencher and Richards (2014) described key research
PT
589 in the constraint of fracture geometry in rock masses. Their work
590 emphasised the need to consider the fundamentals of joint genesis in rock
RI
591 engineering investigations with associated implications for geomechanical
SC
592 modelling.
U
594 Recent developments in geomechanical codes have provided an impetus for
595
AN
consideration of the roles of brittle fracture and damage in large rock slopes.
596 Continuum and discontinuum codes are able to simulate damage in different ways.
M
597 In continuum codes, damage within a rock slope may be characterised implicitly by
D
598 considering the change in number of yielded elements compared to the total number
TE
599 of elements. In discontinuum codes, brittle fracture and damage may be simulated
600 explicitly using a variety of approaches (Table 2). Methods include hybrid Finite-
EP
601 Discrete Element codes (FDEM), Distinct Element codes (DEM) incorporating a
602 Voronoi or Trigon mesh, Particle Flow Codes (PFC), and lattice spring codes. In
C
603 FDEM approaches, the rock slope is discretized into finite elements and, by
AC
606 fragments, enabling the full failure process to be captured. The simulation
608 fracturing. It is also possible to incorporate non-persistent joints into the rock slope
609 either as step-path features or DFNs. Figure 18a shows an example of an open pit
ACCEPTED MANUSCRIPT
610 slope where fracturing of rock bridges within a DFN are induced due to stresses
612 Distinct Element codes may be used to model brittle fracture in both 2D and 3D
613 using four methods – Voronoi, Trigon, PFC, and lattice spring methods. In the first
PT
614 method, intact rock properties are given to the sides of the polygonal Voronoi mesh;
RI
615 cracks develop when the induced stresses cause failure of contacts according to
616 predefined shear and tensile strength properties. A variant of the Voronoi code,
SC
617 Trigon, has recently been introduced where the rock slope is discretized into
618 triangular, not polygonal, blocks; as before, new stress-induced cracks are
U
619 developed when the intact rock properties on the Trigon elements are exceeded.
620 AN
Figure 18b shows the use of Trigon in the 2D UDEC code to simulate brittle fracture.
621 A DFN can be incorporated within the Voronoi or Trigon meshes to allow simulation
M
622 of rock bridge failure between pre-existing joints. A 3D Voronoi/Trigon method has
623 been developed by Gao and Stead (2014) to allow simulation of 3D fracture in rock
D
624 slopes, but has yet to be used for detailed rock slope analyses. The third method of
TE
625 simulating brittle fracture in rock slopes is through the use of PFC either in 2D or 3D.
626 The rock slope is simulated using particles that are bonded together; stress-induced
EP
627 failure of the particle bonds simulates the development of cracks in the rock. Wang
628 et al. (2003) describe the successful use of PFC (Itasca, 2014) to model footwall
C
629 slopes (Figure 18c), while Lorig et al. (2009) show the use of PFC with an
AC
630 incorporated DFN to model large rock slopes (Figure 18d). The combination of a
631 PFC with an incorporated DFN is referred to as a synthetic rock mass (SRM)
632 (Sainsbury et al. 2008). In this method, 3D SRM models are used to define the
633 strength of jointed rock masses up to 100 m in size, and, based on these strengths,
634 continuum 3D codes are used to model large open pit slopes up to 1 km in height.
ACCEPTED MANUSCRIPT
635 Finally, a recent development has been the lattice spring code approach, Slope
636 Model (Itasca, 2014). In this approach the particles and bonds within a PFC code are
637 replaced by nodes and springs. The method has been used both in the simulation of
638 natural rock slopes, such as the Vajont Slide in Italy (Figure 19), and large open pit
PT
639 slopes incorporating DFNs. Havaej et al. (2014), for example, show the use of
640 damage ellipsoids and Flinn plots in the analysis of brittle fracture modelling data
RI
641 from large open pits and landslides.
SC
642 A wide range of sophisticated geomechanical models exist for both 2D and 3D
643 modelling of rock slopes; however, their successful application requires careful
U
644 consideration of structural geology and tectonics:
645 •
AN
Highly tectonized weak rock mass may be amenable to continuum modelling,
646 whereas major discrete faults may dictate the use of discontinuum codes.
M
647 Highly foliated rocks may necessitate the use of special strength models
D
649 kinematics and selection of the appropriate model, as well as the choice of
650 peak and residual shear strengths and the directional shear strength
EP
652 • Geological structures may control not only the basal sliding surface but also
C
654 • Rock mass strength and kinematics will vary and models must consider
656 • Discrete fracture networks will dictate the use of DFN model generators that
657 are able to simulate the true variation of fracture intensity and orientation
661 • The presence of faults or large-scale structures may have important controls
662 on slope failures in seismically active areas. Faults may act as sources of
PT
663 dynamic disturbance leading to failure; major discrete structures, if present in
RI
665 5. Conclusions
SC
666 There are many similarities between structural geology and geotechnical
667 investigations. Both disciplines utilise field characterisation, geophysical, and remote
U
668 sensing methods to analyse features and investigate kinematics and mechanics. The
669
AN
aims of each discipline differ, however, one focussing on the history of the Earth’s
670 deformation and the other concerning soil and rock materials and how they interact
M
671 with the engineered world. Nonetheless, structural geology concepts and principles
D
672 are extremely valuable in geotechnical studies. Currently, common practice in slope
TE
673 design often neglects the investigation of the evolution of structural conditions
675 features.
676 The inheritance of structural settings and features controls slope stability and
C
677 failure behaviour from the micro-scale to macro-scale (Figure 20). Structure can and
AC
678 often does affect fundamental rock mass properties, such as apparent cohesion,
679 frictional resistance, and shear strength, as well as failure kinematics and
681 fractures, and folds affect local behaviour of rock masses. Large features such as
682 decametre-scale folds and faults influence the geometry of rock failures, as well as
ACCEPTED MANUSCRIPT
683 their kinematics. In general, the orientations and dimensions of structures determine
685 The locations of planes of weakness, such as faults or fold hinges, which increase
686 rock mass damage and thus degrade rock mass strength, can control failure location
PT
687 and sliding surface geometry and also influence landslide runout. In addition,
RI
689 stability, acting as either water conduits or aquitards.
SC
690 Whether it is from the consideration of stress and strain techniques or the
691 delineation of major structures and their influence on the rock mass, structural
U
692 geology is of fundamental importance in rock slope investigations. Numerous studies
693 AN
have shown the role of major structures in defining the limits of landslides.
696 strength variations in high mountain slopes. We believe that integrated structural
D
698 mechanisms. Recent landslide studies have shown the immense value of multi-
699 disciplinary studies involving rock engineers, engineering geologists, and structural
EP
701 References
AC
702 Alghalandis, Y.F., Dowd, P.A., Xu, C., 2014. Connectivity field: A measure for
703 characterising fracture networks. Mathematical Geoscience, DOI:
704 10.1007/s11004-014-9520-7.
705 Agliardi, F., Crosta, G.B., Frattini, P., Malusà, M.G., 2013. Giant non-catastrophic
706 landslides and the long-term exhumation of the European Alps. Earth and
707 Planetary Science Letters 365, 263-274.
708 Ambrosi, C., Crosta, G.B., 2011. Valley slope influence on deformation mechanisms
709 of rock slopes. In: Jaboyedoff, M. (Ed.), Slope Tectonics. Geological Society of
710 London, London, 215-233.
ACCEPTED MANUSCRIPT
711 Baczynski, N.R.P., 2000. STEPSIM4 “step-path” method for slope risks. In:
712 Proceedings of the International Conference on Geotechnical and Geological
713 Engineering (GeoEng 2000), Melbourne, Australia, November 2000. Technomic
714 Publishing Co., Lancaster, 86-92.
715 Badger, T.C., 2002. Fracturing within anticlines and its kinematic control on slope
716 stability. Environmental and Engineering Geoscience 8, 19-33.
717 Barton, N.R., 1976. The shear strength of rock and rock joints. International Journal
718 of Rock Mechanics and Mining Sciences & Geomechanics Abstracts 13, 1-24.
PT
719 Barton, N.R., 2011. From empiricism, through theory, to problem solving in rock
720 engineering. In: Qian, Q., Zhou, Y. (Eds.), Harmonising Rock Engineering and
721 the Environment. Taylor & Francis, London, 3-14.
RI
722 Barton, N.R., Choubey, V., 1977. The shear strength of rock joints in theory and
723 practice. Rock Mechanics 10, 1-54.
724 Benko, B., 1997. Numerical modelling of complex slope deformations. Ph.D. thesis,
SC
725 University of Saskatchewan.
726 Benko, B., Stead, D., 1998. The Frank slide: A re-examination of the failure
727 mechanism. Canadian Geotechnical Journal 35, 299-311.
728 Blikra, L.H., 2012. The Åknes rockslide, Norway. In: Clague, J.J., Stead, D. (Eds.),
U
729 Landslides: Types, Mechanisms and Modelling. Cambridge University Press,
730 New York, 323-335.
731
732
733
AN
Brian, M.J., Rosser, N.J., Norman, E.C., Petley, D.N., 2014. Are microseismic
ground displacements a significant geomorphic agent? Geomorphology 207, 161-
173.
734 Brideau, M.-A., 2005. The influence of tectonic structures on rock mass quality and
M
735 implications for rock slope stability. Master’s thesis, Simon Fraser University.
736 Brideau, M.-A., 2010. Three-dimensional kinematic controls on rock slope stability
737 conditions. Ph.D. thesis, Simon Fraser University.
738 Brideau, M.-A., McDougall, S., Stead, D., Evans, S.G., Couture, R., Turner, K., 2012.
D
745 Brideau, M.-A., Yan, M., Stead, D., 2009. The role of tectonic damage and brittle
746 rock fracture in the development of large rock slope failures. Geomorphology
747 103, 30-49.
748 Bromhead, E.N., 1992. The Stability of Slopes. Blackie Academic & Professional,
C
749 Glasgow.
750 Cai, M., Kaiser, P.K., Uno, H., Tasaka, Y., Minami, M., 2004. Estimation of rock
AC
751 mass deformation modulus and strength of jointed hard rock masses using the
752 GSI system. International Journal of Rock Mechanics and Mining Sciences 41, 3-
753 19.
754 Cappa, F., Guglielmi, Y., Soukatchoff, V.M., Mudry, J., Bertrand, C., Charmoille, A.,
755 2004. Hydromechanical modelling of a large moving rock slope inferred from
756 slope levelling coupled to spring long-term hydrochemical monitoring: example of
757 the La Clapière landslide (Southern Alps, France). Journal of Hydrology 291, 67-
758 90.
ACCEPTED MANUSCRIPT
759 Clayton, M.A., 2014. Characterization and analysis of the Mitchell Creek landslide: A
760 large-scale rock slope instability in north-western British Columbia. Master’s
761 thesis, Simon Fraser University.
762 Cruden, D.M., Krahn, J., 1978. Frank Rockslide, Alberta, Canada. In: Voight, B.
763 (Ed.), Rockslides and Avalanches, Elsevier, Amsterdam, 97-112.
764 Damjanac, B., Varun, Lorig, L., 2013. Seismic stability of large open pit slopes and
765 pseudo-static analysis. In: Dight, P.M. (Ed.), Slope Stability 2013. Australian
766 Centre for Geomechanics, Perth, 1203-1216.
PT
767 Diederichs, M.S., 2003. Manuel Rocha medal recipient rock fracture and collapse
768 under low confinement conditions. Rock Mechanics and Rock Engineering 36,
769 339-381.
RI
770 Dowd, .A., Xu, C., Mardia, K.V., Fowell, R.J., 2007. A comparison of methods for the
771 stochastic simulation of rock fractures. Journal of Mathematical Geology 39, 697–
772 714.
SC
773 Duncan, J.M., Wright, S.G., 2005. Soil Strength and Slope Stability. John Wiley &
774 Sons, Inc., Hoboken, New Jersey.
775 Eberhardt, E., Stead, D., Coggan, J.S., 2004. Numerical analysis of initiation and
776 progressive failure in natural rock slopes – the 1991 Randa rockslide.
U
777 International Journal of Rock Mechanics and Mining 41, 69-87.
778 Fleuty, M.J., 1964. The description of folds. Proceedings of the Geologists’
779
780
781
Association 75, 461-492.
AN
Fossen, H., 2010. Structural Geology. Cambridge University Press, Cambridge.
Franz, J., 2009. An investigation of combined failure mechanisms in large scale open
782 pit slopes. Ph.D. thesis, University of New South Wales.
M
783 Ganerød, G.V., Grøneng, G., Rønning, J.S., Dalsegg, E., Elvebakk, H., Tønnesen,
784 J.F., Kveldsvik, V., Eiken, T., Blikra, L.H., Braathen, A., 2008. Geological model
785 of the Åknes rockslide, western Norway. Engineering Geology 102, 1-18.
786 Gao, F.Q., Stead, D., 2014. The application of a modified Voronoi logic to brittle
D
787 fracture modelling at the laboratory and field scale. International Journal of Rock
788 Mechanics and Mining Sciences 68, 1-14.
TE
789 Gischig, V., Löw, S., Kos, A., Moore, J.R., Raetzo, H., Lemy, F., 2009. Identification
790 of active release planes using ground-based differential InSAR at the Randa rock
791 slope instability, Switzerland. Natural Hazards and Earth System Sciences 9,
792 2027-2038.
EP
793 Gischig, V., Amann, F., Moore, J.R., Löw, S., Eisenbeiss, H., Stempfhuber, W.,
794 2011. Composite rock slope kinematics at the current Randa instability,
795 Switzerland, based on remote sensing and numerical modelling. Engineering
796 Geology 118, 37-53.
C
797 Glastonbury, J., Fell, R., 2000. Report on the analysis of “rapid” natural rock slope
798 failures. University of New South Wales, School of Civil and Environmental
AC
809 Hamdi, P., Stead, D., Elmo, D., 2013. Numerical simulation of damage during
810 laboratory testing on rock using a 3D-FEM/DEM approach. In: Proceedings
811 ARMA 2013, ARMA Paper 13-583.
812 Hammah, R.E., Yacoub, T., Curran, J.H., 2009. Variation of failure mechanisms of
813 slopes in jointed rock masses with changing scale. In: Diederichs, M., Grasselli,
814 G. (Eds.), Proceedings of the 3rd CANUS Rock Mechanics Symposium, Toronto,
815 ON. Canadian Association of Rock Mechanics, Paper 3956.
816 Haneberg, W.C., 2007. Directional roughness profiles from three-dimensional
PT
817 photogrammetric or laser scanner point clouds. In: Eberhardt, E., Stead, D.,
818 Morrison, T. (Eds.), Rock Mechanics: Meeting Society’s Challenges and
819 Demands. Proceedings of the 1st Canada-US Rock Mechanics Symposium,
RI
820 Vancouver, BC, Canada, 101-106.
821 Havaej, M., Stead., D., Mayer, J., Wolter, A., 2014. Modelling the relationship
822 between failure kinematics and slope damage in high rock slopes using a lattice
SC
823 scheme approach. ARMA 2014.
824 Havaej, M., Wolter, A., Stead, D., Tuckey, Z., Lorig, L., Eberhardt, E., 2013.
825 Incorporating brittle fracture in three-dimensional modelling of high rock slopes.
826 In: Dight, P.M. (Ed.), Slope Stability 2013, Brisbane, September, 2013. Australian
U
827 Centre for Geomechanics, Perth.
828 Hencher, S.R., 2013. Characterizing discontinuities in naturally fractured outcrop
829
830
831
AN
analogues and rock core: The need to consider fracture development over
geological time. In: Spence, G.H., Redfern, J., Aguilera, R., Bevan, T.G.,
Cosgrove, J.W., Couples, G.D., Daniel, J.-M. (Eds.), Geological Society, London,
832 Special Publications 374, 113-123.
M
833 Hencher, S.R., Richards, L.R., 1982. The basic frictional resistance of sheeting joints
834 in Hong Kong Granite. Hong Kong Engineering 11, 21-25.
835 Hencher, S.R., Richards, L.R., 2014. Assessing the shear strength of rock
836 discontinuities at laboratory and field scales. Rock Mechanics and Rock
D
839 Characterisation, shear strength and engineering. Rock Mechanics and Rock
840 Engineering 44, 1-22.
841 Hoek, E., Bray, J. W., 1981. Rock slope engineering. Taylor & Francis.
842 Hoek, E., Brown, E.T., 1980. Empirical strength criterion for rock masses. Journal of
EP
849 of in situ stresses on open pit design. In: Read, J., Stacey, P. (Eds.), Guidelines
850 for Open Pit Slope Design. CSIRO Publishing, Collingwood, Australia, 437-445.
851 Hoek, E., Kaiser, P.K., Bawden, W.F., 1995. Support of Underground Excavations in
852 Hard Rock. A.A. Balkema, Rotterdam, Netherlands.
853 Humair, F., Pedrazzini, A., Epard, J.-L., Froese, C.R., Jaboyedoff, M., 2013.
854 Structural characterization of Turtle Mountain anticline (Alberta, Canada) and
855 impact on rock slope failure. Tectonophysics 605, 133-148.
856 Hungr, O., Amann, F., 2011. Limit equilibrium of asymmetric laterally constrained
857 rockslides. International Journal of Rock Mechanics and Mining Sciences 48,
858 748-758.
ACCEPTED MANUSCRIPT
859 Itasca, 2014. 3DEC (v. 5.0), FLAC2D (v. 7.0), FLAC3D (v. 5.0), PFC (v. 5.0), Slope
860 Model (v. 2.9), UDEC (v. 6.0). Itasca Consulting Group. Minneapolis, MN, USA.
861 Jaboyedoff, M., Metzger, R., Oppikofer, T., Couture, R., Derron, M.-H., Locat, J.,
862 Turmel, D., 2007. New insight techniques to analyze rock-slope relief using DEM
863 and 3D-imaging cloud points: COLTOP-3D software. In: Eberhardt, E., Stead, D.,
864 and Morrison, T. (Eds.) Rock Mechanics: Meeting Society’s Challenges and
865 Demands. Taylor & Francis, London, 61-68.
866 Jennings, J.E.B., 1970. A mathematical theory for the calculation of the stability of
PT
867 slopes in open cast mines. In: Proceedings of the Symposium on the Planning of
868 Open Pit Mines, Johannesburg, South Africa, August/September 1970. Balkema,
869 Cape Town, 87–102.
RI
870 Kalenchuk, K.S., 2010. Multi-dimensional analysis of large, complex slope instability.
871 Ph.D. thesis, Queen’s University.
872 Kinakin, D., Stead, D., 2005. Analysis of the distributions of stress in natural ridge
SC
873 forms: Implications for the deformation mechanisms of rock slopes and the
874 formation of sackung. Geomorphology 65, 85-100.
875 Lato, M., Diederichs, M.S., Hutchinson, D.J., Harrap, R., 2009. Optimization of
876 LiDAR scanning and processing for automated structural evaluation of
U
877 discontinuities in rock masses. International Journal of Rock Mechanics and
878 Mining, 46, 194-199.
879
880
881
AN
Lebourg, T., Mickael, H., Hervé, J., Samyr, E.B.B., Thomas, B., Swann, Z.,
Emmanuel, T., Maurin Jr., V., 2010. Temporal evolution of weathered cataclastic
material in gravitational faults of the La Clapiere deep-seated landslide by
882 mechanical approach. Landslides 8, 241-252.
M
883 Leith, K., Moore, J.R., Amann, F., Löw, S., 2014a. Subglacial extensional fracture
884 development and implications for Alpine Valley evolution. Journal of Geophysical
885 Research: Earth Surface 119, 62-81.
886 Leith, K., Moore, J.R., Amann, F., Löw, S., 2014b. In situ stress control on
D
889 Lorig, L., Stacey, P., Read, J., 2009. Slope design methods. In: Read, J., Stacey, P.
890 (Eds.), Guidelines for Open Pit Slope Design. CSIRO Publishing, Collingwood,
891 237-264.
892 Löw, S., Gischig, V., Willenberg, H., Alpiger, A., Moore, J.R., 2012. Randa:
EP
893 Kinematics and driving mechanisms of a large complex rockslide. In: Clague,
894 J.J., Stead, D. (Eds.), Landslides: Types, Mechanisms and Modelling. Cambridge
895 University Press, New York, 297-309.
896 Marinos, P., Hoek, E., 2000. GSI: A geologically friendly tool for rock mass strength
C
908 Moore, J.R., Gischig, V., Burjanek, J., Löw, S., Fäh, D., 2011. Site effects in unstable
909 rock slopes: Dynamic behaviour of the Randa instability (Switzerland). Bulletin of
910 the Seismological Society of America 101, 3110-3116.
911 Newman, S., 2013. Deep-seated gravitational slope deformations near the Trans-
912 Alaska Pipeline, east-central Alaska Range. Master’s thesis, Simon Fraser
913 University.
914 Nicksiar, M., Martin, C.D., 2014. Factors affecting crack initiation in low porosity
915 crystalline rocks. Rock Mechanics and Rock Engineering 47, 1165-1181.
PT
916 Pedrazzini, A., Jaboyedoff, M., Froese, C.R., Langenberg, C.W., Moreno, F., 2011.
917 Structural analysis of Turtle Mountain: Origin and influence of fractures in the
918 development of rock slope failures. Geological Society, London, Special
RI
919 Publications 351, 163-183.
920 Petley, D., 2012. Remote sensing techniques and landslides. In: Clague, J.J., Stead,
921 D. (Eds.), Landslides: Types, Mechanisms and Modelling. Cambridge University
SC
922 Press, New York, 159-171.
923 Priest, S.D., 1980. The use of inclined hemisphere projection methods for the
924 determination of kinematic feasibility, slide direction and volume of rock blocks.
925 International Journal of Rock Mechanics and Mining Sciences & Geomechanics
U
926 Abstracts 17, 1-23.
927 Ramsay, J.G., 1967. Fold and fracturing of rocks. McGraw-Hill, New York.
928
929
930
Publishing, Collingwood.
AN
Read, J., Stacey, P. (Eds.), 2009. Guidelines for Open Pit Slope Design. CSIRO
Richards, L.R., Leg., G.M.M., Whittle, R.A., 1978. Appraisal of stability conditions in
931 rock slopes. In: Bell, F.G. (Ed.), Foundation Engineering in Difficult Ground,
M
932 Butterworths, London, 449-512.
933 Rocscience, 2014. DIPS (v. 6.0), Phase2 (v. 9.0), RS3 (v. 1.0), Swedge (v. 6.0).
934 Rocscience Inc., Toronto, ON, Canada.
935 Sainsbury, B., Pierce, M.E., Mas Ivars, D., 2008. Analysis of caving behaviour using
D
936 a synthetic rock mass-ubiquitous joint rock mass modelling technique. In:
937 Proceedings of the 1st Southern Hemisphere International Rock Mechanics
TE
938 Symposium (SHIRMS 2008), Perth, Australia. International Society for Rock
939 Mechanics, 243-253.
940 Shipton, Z.K., Cowie, P.A., 2003. A conceptual model for the origin of fault damage
941 zone structures in high-porosity sandstone. Journal of Structural Geology 25,
EP
942 333-344.
943 Shipton, Z.K., Cowie, P.A., 2008. Damage zone and slip-surface evolution over µm
944 to km scales in high-porosity Navajo sandstone, Utah. Journal of Structural
945 Geology 23, 1825-1844.
C
946 Stead, D., Coggan, J.S., 2012. Numerical modelling of rock-slope instability. In:
947 Clague, J.J., Stead, D. (Eds.), Landslides: Types, Mechanisms and Modelling.
AC
957 remote sensing techniques. Natural Hazards and Earth System Sciences 9, 267-
958 287.
959 Styles, T.D., 2009. Numerical modelling and analysis of slope stability within fracture
960 dominated rock masses. Ph.D. thesis, Camborne School of Mines, University of
961 Exeter.
962 Tatone, B.S.A., Grasselli, G., 2010. A new 2D discontinuity roughness parameter
963 and its correlation with JRC. International Journal of Rock Mechanics and Mining
964 Sciences 47, 1391-1400.
PT
965 Thiessen, R., Means, W.D., 1980. Classification of fold interference patterns: A re-
966 examination. Journal of Structural Geology 2, 311-326.
967 Tuckey, Z., 2012. An integrated field mapping-numerical modelling approach to
RI
968 characterising discontinuity persistence and intact rock bridges in large open pit
969 slopes. Master’s thesis, Simon Fraser University.
970 Tuckey, Z., Stead, D., Eberhardt, E., 2013. An integrated approach for
SC
971 understanding uncertainty of discontinuity persistence and intact rock bridges in
972 large open pit slopes, Slope Stability 2013, Brisbane, Sept 2013.
973 Vengeon, J.-M., Giraud, A., Antoine, P., Rochet, L., 1999. Contribution à l’analyse de
974 la deformation et de la rupture des grands versants rocheux en terrain
U
975 cristallophyllien. Canadian Geotechnical Journal 36, 1123-1136.
976 Vivas Becerra, J., 2014. Groundwater characterization and modelling in natural and
977
978
979
AN
open pit rock slopes. Master’s thesis, Simon Fraser University.
Vyazmensky, A., Stead, D., Elmo, D., Moss, A., 2010. Numerical analysis of block
caving-induced instability in large open pit slopes: A finite element/discrete
980 element approach. Rock Mechanics and Rock Engineering 43, 21-39.
M
981 Wang, C., Tannant, D.D., Lilly, P.A., 2003. Numerical analysis of the stability of
982 heavily jointed rock slopes using PFC2D. International Journal of Rock
983 Mechanics and Mining Sciences 3, 415-424.
984 Willenberg, H., Löw, S., Eberhardt, E., Evans, K.F., Spillman, T., Heinke, B., Maurer,
D
985 H., Green, A.G., 2008. Internal structure and deformation of an unstable
986 crystalline rock mass above Randa (Switzerland): Part I – internal structure from
TE
995 Wyllie, D.C., Mah, C.W., 2004. Rock Slope Engineering: Civil and Mining (4th Ed.).
996 Spon Press, New York.
AC
997 Xu, C., Dowd, P., 2010. A new computer code for discrete fracture network
998 modelling. Computers and Geosciences 36, 292-301.
999 Yan, M., 2008. Numerical modelling of brittle fracture and step-path failure: from
1000 laboratory to rock slope scale. Ph.D. thesis, Simon Fraser University.
1001 Ziegler, M., Löw, S., Bahat, D., 2014. Growth of exfoliation joints and near-surface
1002 stress orientations inferred from fractographic markings observed in the upper
1003 Aar valley (Swiss Alps). Tectonophysics 626, 1-20.
1004 Ziegler, M., Löw, S., Moore, J.R., 2013. Distribution and inferred age of exfoliation
1005 joints in the Aar Granite of the central Swiss Alps and relationship to Quaternary
1006 landscape evolution. Geomorphology 201, 344-362.
ACCEPTED MANUSCRIPT
1007 Zoback, M.L., 1992. First- and second-order patterns of stress in the lithosphere:
1008 The World Stress Map Project. Journal of Geophysical Research: Solid Earth 97,
1009 11703-11728.
1010
PT
RI
U SC
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT
1012 Figure 1. Key rock failure modes considered in slope stability analysis: a)
1013 planar/translational sliding, b) toppling, and c) wedge sliding. d) Multi-planar
1014 translational failure (Palliser rockslide), demonstrating the complexity slope
1015 engineers may encounter.
PT
1016 Figure 2. Classification of failure modes related to structures. a) Large rock glide, b)
1017 rough translational slide, c) planar translational slide, d) toe buckling translational
1018 slide, e) biplanar compound slide, f) curved compound slide, g) toppling failure, h)
1019 irregular compound slide (modified after Glastonbury and Fell, 2000).
RI
1020 Figure 3. Open pit slope showing influence of structural geology, with translational
1021 failure mechanism and low-angle thrust faulting influencing the footwall instability
SC
1022 (along bedding). Note the seepage at the fold hinge.
1023 Figure 4. Modification of Fleuty’s (1964) fold classification diagram to indicate slope
1024 stability and failure complexity. Kinematic complexity is highest in the centre of the
U
1025 diagram, where multiple modes of failure may be combined. Inset shows an example
1026 of the interaction between slopes and tectonic structure at Mount Kidd, Canada.
1027
1028
AN
Figure 5. Modification of fold tightness classification, showing how damage increases
with increasing fold tightness (after Fossen, 2010). Transverse, longitudinal, and
1029 converse discontinuities may be present in open folds, whereas tight folds may
1030 develop foliation and cleavage.
M
1031 Figure 6. Scales of roughness at the Vajont Slide, ranging from a) global (macro)
1032 scale synclines (MS = Massalezza Syncline with hinge trending N, ES = Erto
D
1033 Syncline trending E) to meso-scale interference patterns (b), parasitic folds (c), and
1034 micro-scale concretions (d). CT = Col Tramontin Fault, and CE = Col delle Erghene
1035 Fault in a).
TE
1036 Figure 7. Fold and slope interaction creating slope instability (indicated by grey
1037 areas). S0 = bedding, S1 = discontinuity set 1, S2 = discontinuity set 2. a) Bedding
1038 dips parallel to slope and orthogonal discontinuity set acts as rear release, b)
EP
1039 bedding dips into slope and orthogonal discontinuity set acts as main sliding surface,
1040 and c) slope face intersects with fold hinge and bedding and orthogonal discontinuity
1041 set create a biplanar failure surface (modified from Badger, 2002).
C
1042 Figure 8. The 1903 Frank Slide in Alberta, Canada. The South Peak area is still
1043 unstable and subject to ongoing monitoring.
AC
1044 Figure 9. a) Example of drag folding-induced slope failure at a former UK coal mine.
1045 Fault outcrops in the trees at the back of the failure, and dips parallel to general
1046 sliding direction indicated by the white arrow. Note steepening of bedding adjacent to
1047 fault in b), which resulted in additional driving forces for failure and ultimately slope
1048 instability.
1049 Figure 10. The 1991 Randa rock slope failure in Switzerland. Foliation planes dip
1050 into the slope (Gischig et al., 2011).
ACCEPTED MANUSCRIPT
1051 Figure 11. La Clapière landslide, France (inset), an active part of a larger DSGSD
1052 complex (from Lebourg et al., 2010).
1053 Figure 12. Kinematic analysis conducted in the stereographic software DIPS
1054 (Rocscience, 2014), showing daylight envelope containing poles to planar failures.
1055 Friction cone delimits joints with dip greater than assumed friction angle, and hence
1056 potential sliding.
PT
1057 Figure 13. Two pentahedral wedge geometries in Swedge (Rocscience, 2014).
1058 Figure 14. Geological Strength Index (GSI) chart, based on block size and joint
1059 condition (from Hoek et al., 1995).
RI
1060 Figure 15. GSI chart for heterogeneous rock masses such as flysch (from Marinos
1061 and Hoek, 2000).
SC
1062 Figure 16. Examples of continuum analysis. a) Phase2 simulations of a 60-m-high
1063 slope using different joint networks – continuous (left), non-persistent (centre), and
1064 Voronoi (right) (modified after Hammah et al., 2009). b) RS3 model of an open pit
1065 (from Rocscience website, http://www.rocscience.com). c) FLAC2D model of the
U
1066 Frank Slide showing deformation along a presumed failure surface (after Benko and
1067 Stead, 1998). d) FLAC3D model geometry of the Åknes, Norway, rock slope
1068
1069
AN
instability (after Grøneng et al., 2010).
Figure 17. Examples of discontinuum analysis. a) UDEC model of the Vajont Slide,
1070 Italy (after Wolter, 2014). Note that the sliding surface follows folded bedding. b)
M
1071 3DEC model of the Beauregard, Italy, landslide, showing discrete movement blocks
1072 (after Kalenchuk, 2010). c) UDEC model of the Frank Slide, Canada (after Benko
1073 and Stead, 1998).
D
1074 Figure 18. Examples of brittle fracture and damage modelling. a) Progressive
1075 development of failure due to excavation of an open pit in an FDEM simulation (after
TE
1076 Vyazmensky et al., 2010). b) Brittle fracture within a Trigon model of the 1963 Vajont
1077 Slide, Italy, in UDEC, where red lines indicate tensile failure, and green lines indicate
1078 shear failure (after Wolter, 2014). c) Failure of a slope with 70% fracture persistence
1079 in PFC2D (after Wang et al., 2003). d) Combination of particle flow code and DFN in
EP
1084 Figure 20. Summary of the influence of structure on rock slope stability and failure
1085 mechanisms.
1086 Table Captions
1087 Table 1. Summary of case studies presented in the text.
1088 Table 2. Summary of numerical modelling approaches used to simulate brittle
1089 fracture.
ACCEPTED MANUSCRIPT
Table 1
Name Dominant Type Volume (m3) Lithology Structural Control Mechanisms
PT
Åknes, Norway DSGSD 35,000,000 gneiss foliation, 6 DS, faults, folds sliding
> 1960 - present rockslide biotite schist rear release - tension crack intact rock fracture
gouge lateral release - faults creep
RI
Beauregard, Italy DSGSD 663,000,000 mica schist sub-vertical DS sliding
1951 - present paragneiss release - fractures, trough toppling
SC
Downie, British Columbia DSGSD 1,500,000,000 micaschist 4 main DS, foliation, folds, fault creep
10 000 yr ago gneiss rear release - ridge crest sliding
U
quartzite lateral release - joint sets
East Gate, British Columbia rockslide n/a phyllite 4 main DS, foliation, cleavage sliding
AN
1997 debris flows release - joint sets flow
M
Jul. 13, 2006 release - joint sets rockfall
D
Flims, Switzerland rockslide 8.5- limestone faults, folds, foliation sliding
8500 yr ago 12,000,000,000 release - joint sets, faults toppling
TE
Frank, Alberta rock avalanche 30,000,000 limestone 3 DS, anticline, faults sliding
Apr. 29, 1903 clastics rear release - fold hinge toppling
EP
Hope, British Columbia rockslide 47,000,000 greenstone foliation, DS, shear zones, faults sliding
1965 felsite rear release - fault flow
lateral release - joint sets, faults
ACCEPTED MANUSCRIPT
PT
8700 yr ago schist release - foliation, joint sets flow
RI
1900 - present DSGSD schist fold alters foliation orientation sliding
diorite rear release - fractures creep
SC
lateral release - fractures, fold
Madison, Montana rockslide 20,000,000 gneiss 3-4 DS, foliation, regional faults composite sliding
Aug. 17, 1959 schist rear release - ridge crest wedge failure
U
dolostone lateral release - gullies (creep)
earthquake
AN
Mt Steele, Yukon rock/ice avalanche 27-80,000,000 granodiorite fault, 3 main DS sliding
Jul. 24, 2007 (5/95%) diorite release - joint sets flow
gabbro earthquake
M
Palliser, Alberta rockslide 8,000,000 cherty carbonates 5 DS, regional thrust, syncline sliding
prehistoric (2 events) 20,000,000 syncline curves bedding as base wedge failure
D
release - joint sets
Randa, Switzerland rockslide (2-stage) 30,000,000 ortho/paragneiss
TE 3 main DS, fault, schistosity sliding
Apr. 18, 1991 DSGSD rear release - joint sets toppling
May 9, 1991 3-9,000,000 lateral release - joint sets, fault
present
EP
Vajont, Italy rockslide 270,000,000 carbonates 9 DS, folds, faults sliding, rotation
Oct. 9, 1963 clay rear release - tension crack, fault wedge failure
C
Jul. 28, 1987 quartzite rear release - joint sets, faults sliding
gneiss, lateral release - fault, joint sets
amphibolite
ACCEPTED MANUSCRIPT
1090 Table 2
Approach Advantages Limitations Comments
Finite-Discrete • simulates intact rock fracture in rock • slope modelling mainly Successfully used in
Element Hybrid slopes from continuum to discontinuum confined to 2D to-date with modelling both natural
(Elfen/YGEO) • possible to fracture across or around future development in 3D and engineering rock
triangular elements pending slopes (Eberhardt et
• possible to include influence of • simulation of groundwater al., 2004; Stead et al.
groundwater table (Styles 2009) pressure dissipation in slopes 2006).
PT
• DFNs can be incorporated within rock varies in different FDEM
slope models models
• directional weakness may be simulated
Distinct Element • simulates intact rock fracture associated • careful calibration required of Used successfully in
RI
+ Voronoi/Trigon with slope failure, can incorporate Voronoi polygon/Trigon contact modelling of natural
(UDEC 2D, 3DEC polygons within bedded/through-going properties rock slopes and
3D) jointed rock masses • polygon size may influence engineered rock
• possible to incorporate DFNs, but adds kinematics of movement slopes. Further
SC
complexity • may need to edit DFNs to research required on
• triangular elements allow increased allow solution calibration of model
kinematic freedom • limited experience in use of properties.
• possible to incorporate influence of Trigon modelling
U
groundwater pressure in UDEC Voronoi • long model run times for large-
and Trigon simulation of slopes (Vivas scale problems especially for
2014) 3D problems which may not be
Particle Flow
Codes
AN
• brittle fracture of slopes can be modelled
through breaking of bonds between
feasible in some cases
• require careful calibration
• large rock slope models
Used to successfully
model footwall failures
(PFC2D/PFC3D) particles in 2D or spheres in 3D possible in 2D but in mines, large rock
M
• ability to include DFNs and groundwater computationally expensive in slopes and subsequent
pressures 3D runout and large open
• SRM testing possible • limited use in 3D models to pits slope.
date
D
SRM approach • derivation of large scale rock mass • SRM testing programs Successful use
(PFC3D/DFN/ properties through simulated brittle required for important predominantly for major
FLAC3D) fracture testing of large scale uniaxial, lithological units within rock mine slopes to date.
TE
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT
PT
RI
U SC
AN
M
D
TE
EP
C
AC