Content uploaded by Jennifer J. Day
Author content
All content in this area was uploaded by Jennifer J. Day on Sep 30, 2021
Content may be subject to copyright.
Structural and geomechanical analysis of the 2016
Elephant Rock failure at Hopewell Rocks Provincial
Park, New Brunswick
Amanda Hyslop & Jennifer J. Day
Department of Geological Sciences and Geological Engineering – Queen’s
University, Kingston, ON, Canada
Stefan Kruse
Terrane Geoscience Inc., Fredericton, NB, Canada
Kevin Snair
Hopewell Rocks Provincial Park, Department of Tourism, Heritage and Culture, New Brunswick, Canada
ABSTRACT
Hopewell Rocks Provincial Park, on the Bay of Fundy, is a popular attraction for geotourism. On March 14th, 2016, a partial
failure occurred of the Elephant Rock sea stack formation. This study presents a geomechanical interpretation of this failure
supported by analyses of a 3D photogrammetry model of the failure surface, photographs taken before and after the failure,
historic tidal, climate, and weather data, and the historic rockfall database from the park. Major contributing factors to the
failure include tidal erosion near the base of the formation, weakening of intact rock thorugh marine exposure in the
intertidal zone, and freeze-thaw effects on fracture propagation. The results of this study provide insight to rockfalls and
sea stack failures for evaluating risk of future potential failures that may impact geotourism.
RÉSUMÉ
Le parc provincial Hopewell Rocks, sur la baie de Fundy, est une attraction populaire pour le géotourisme. Le 14 mars
2016, une défaillance partielle s'est produite dans la formation de l'empilement marin d'Elephant Rock. Cette étude
présente une interprétation géomécanique de cette défaillance supporté par des analyses d'un modèle photogrammétrique
3D de la surface de rupture, des photographies prises avant et après la défaillance, des données historiques sur les
marées, le climat et la météo, et la base de données historique des chutes de pierres du parc. Les principaux facteurs
contribuant à l'échec comprennent l'érosion par les marées près de la base de la formation, l'affaiblissement de la roche
intacte à travers l'exposition marine dans la zone intertidale et les effets du gel-dégel sur la propagation des fractures. Les
résultats de cette étude donnent un aperçu des chutes de pierres et des ruptures de cheminées marines pour évaluer le
risque de futures pannes potentielles susceptibles d'avoir un impact sur le géotourisme.
1 INTRODUCTION
Sea stacks and shore stacks are geological landforms
consisting of steep vertical columns of rock detached by
wave action from a cliff-lined shore. They are among the
world's most recognizable landscape features, and as
such, they are popular destinations for geotourism. These
structures are produced when pre-existing rock joints,
fractures, or other structural weaknesses are preferentially
eroded by mechanical or hydraulic wave erosion, and more
resistant rocks are left standing, eventually eroding to a
stack (Gardner 2019). Due to the relatively high erosion
rates and the tall and narrow geometry of these formations,
sea stacks are prone to structural instability, such as partial
toppling failures and collapses (Galea et al. 2018). Famous
examples of these formations are found at the Hopewell
Rocks Provincial Park in the Bay of Fundy, New Brunswick,
Canada.
There have been numerous partial rockfall
failures or complete collapses of sea stacks and sea cliffs
over the years at the Hopewell Rocks Provincial Park. A
recent major failure occurred on March 14th, 2016, of a
partial failure of the formation known as the Elephant Rock
(Figure 1). Filocomo (2016) conducted a preliminary study
of this collapse, which consisted of collecting structural and
geotechnical data, constructing a full 3D point-cloud
photogrammetry model of the Elephant Rock (Figure 2)
using photographs collected from an unmanned aerial
vehicle (UAV), and conducting geotechnical laboratory
tests on rock samples to measure the density, porosity, and
unconfined compressive strength (UCS) of the intact rock.
In this study, the 3D photogrammetry model developed by
Filocomo (2016) is used to conduct a detailed analysis of
the failure surface. The failure surface is divided into ten
segments based on orientation to analyze the curvature of
each segment and interpret which portions of the failure
surface were likely pre-existing joints or intact rock bridges
prior to failure. The strength of both weathered and fresh
intact rocks, tidal erosion, freeze-thaw effects, and rain
accumulation prior to the Elephant Rock failure and other
recorded rockfalls in the park are analyzed in this study and
contribute to the presented interpretation of the 2016
Elephant Rock failure.
2 GEOLOGICAL SETTING
The Hopewell Rocks Provincial Park is located on the
shores of the Bay of Fundy at Hopewell Cape near multiple
coastal tourism centers, including Fundy National Park
(Figure 3). The Bay of Fundy has an average tidal range of
10.7 m, which is the world's largest tidal range (Trenhaile
et al. 1998). The sea stack formations at the Hopewell
Rocks are composed of coarse-grained, poorly sorted,
polymictic conglomerates and arkosic sandstone with thin
interbeds of silt and sandstone (Figure 4). These rocks
belong to the Hopewell Conglomerate and Maringouin
Formation of the Carboniferous Hopewell Group (Trenhaile
et al. 1998).
Figure 1. The Elephant rock formation before (left) and after (right) the March 14th, 2016 partial collapse
Figure 2. 3D photogrammetry model of the Elephant Rock
collected in May 2016 after the March 2016 collapse,
looking approximately northeast
Figure 3. Map of Hopewell Rocks Provincial Park showing
Elephant Rock and some other landmarks
Figure 4. Coarse (top) and fine interbeds (bottom) of the
Hopewell Conglomerate
2.1 Structure Orientations
Structural orientation observations and measurements of
bedding and joints from Filocomo (2016), Jacques,
Whitford and Associates Ltd. (1987), Hanson (1977), and
data collected by the authors have been collectively
interpreted in this study. The results consist of one bedding
set, two joint sets, and some outlier planar structures
(Figure 5). The groups of structural sets were determined
by including data within a 30° radius of the maximum
concentration of data points.
3 INTACT ROCK STRENGTH
A Schmidt hammer was used on the Elephant Rock to
measure the compressive wall strength of clasts and matrix
components of the conglomerate in the intertidal zone, on
both the seaward failure surface and on the weathered
landward wall. The results are shown in Figure 6, where
unconfined compressive strength (UCS) values were
obtained using a dry unit weight of 23.5 kN/m3 calculated
from the average dry density and porosity reported by
Filocomo (2016). The average UCS values of the seaward
failure surface and the weathered landward wall were
determined to be 48 MPa and 37 MPa, respectively.
Seawater exposure during high tide explains the lesser
strength of the weathered landward wall compared to the
younger seaward failure surface. This is supported by field
observations made when collecting loose rock block
samples from above and below the tidal zone. The samples
from below the tidal zone crumbled under their weight when
picked up, whereas the samples from above the tidal zone
remained intact. Aqueous dissolution of the sedimentary
cements within the rock is hypothesized to be the driving
mechanism of this weakening.
Figure 5. Orientations of bedding and joint rockmass
structures at Hopewell Rocks Provincial Park
Figure 6. Unconfined Compressive Strength (UCS) of
clasts and matrix of the conglomerate, measured in 2019
with a Schmidt hammer on the Elephant Rock formation
4 FAILURE SURFACE ANALYSIS
The curvature of the failure surface was analyzed using
multiple methods to distinguish between portions of the
failure surface that were likely pre-existing fractures or
intact rock bridges.
4.1 Curvature Quantification
The Elephant Rock failure surface was divided into ten
segments based on orientation in the photogrammetry
model (Figure 7) using CloudCompare’s FACETS Plugin
Kd-tree approach (CloudCompare 2021). This feature
recursively divides the 3D point cloud into quarter cells until
the points within the cell all fit the best-fitting plane with a
specified maximum deviation (maximum distance @ 99%).
The subdivision is complete when the cells are empty or
have less than 6 points; otherwise, the root mean square is
incorrect. The minimum number of points for a facet to be
included is also specified.
Figure 7. Detail of failure surface point cloud model divided
into ten segments based on orientation, looking northwest
A sensitivity analysis was completed to determine the
best input parameters for the Kd-tree cell fusion algorithm
and facet definition to subdivide the failure surface
accurately. The maximum angle for the Kd-tree fusion was
set at 20° due to the large variability in the joint sets. The
maximum deviation from the best-fit plane was set at 0.5 m,
determined using the approach outlined by Dewez et al.
(2016). A value of 1 m was used for the maximum relative
distance between the merged patches and the current facet
center because smaller values reduced the size of the
individual facets, making the general orientation of different
parts of the failure surface difficult to distinguish. The
maximum edge length, which is used to extract the
(concave) contour at the end of the facet extraction
process, was chosen to be 1 m to decrease noise and
computation time and still have the contour relatively close
to the points. Finally, the minimum number points below
which a facet should be discarded was selected to be 162
points, which is approximately 0.01% of the failure
surface’s point cloud (Figure 7).
Each segment’s curvature was quantified using two
different methods described by Sturzenegger and Stead
(2009). Curvature is defined as "surface irregularities with
a wavelength greater than about 100 mm" by Priest (1993).
Greater curvature relates to higher shear strength, which
reduces the probability of failure. Fractures such as joints
tend to be relatively planar features that are smooth with
small curvature; therefore, the lower curvature segments
would more likely represent pre-existing fractures. It is
important to note that the joints in the area are quite
undulating and very rough due to the large distribution of
grain and clast sizes in the conglomerate.
4.1.1 Curvature Quantification - Method 1
The first method used to quantify the curvature of the ten
segments involved fitting an average plane to each surface
and developing a curvature map defined by the orthogonal
distance between each average plane and the actual
surface of each segment. To understand the distribution of
the deviation from the best fit plane for each segment, the
resulting differences were extracted to compute the
standard deviations (Figure 8). Segments 4 and 9 have the
largest standard deviation and segments 1 and 6 have the
smallest. All best fit planes of the failure surface segments
have similar orientations to those of the regional joint set
data, as shown in Figure 9.
Figure 8. Curvature maps for method 1 and the standard
deviation of each segment
Figure 9. Stereonet showing best fit orientations of the ten
failure surface segments and known regional joint set data
4.1.2 Curvature Quantification - Method 2
The second method involves quantifying the curvature of
the ten failure surface segments using the CloudCompare
(2021) Compass Tool to obtain orientation measurements
at numerous places on each segment in windows of
progressively increasing size (Sturzenegger and Stead
2009). The Compass Tool is used to select points within a
circle of the photogrammetry model and calculates the dip
and dip direction of a best-fit plane to those points. First, a
circle with a radius of 0.10 m was used. The circle radius
was then increased to 0.20 m, 0.40 m, and 0.80 m. The
poles of the average plane orientation of each circular
window were plotted on strereonets (Figure 10) to interpret
the distribution of the orientation measurements. In this
analysis, a larger or more sporadic distribution represents
greater curvature. Segments 1 2, 3, 5 and 6 have the
smallest distribution radii. Segment 3 also has two areas of
pole concentrations indicating that the segment is
undulating. Segments 4, 7, 8, 9 and 10 have a more
sporadic distribution.
4.1.3 Differences in Curvature
Intact rock bridges are interpreted to be a potential reason
for the difference in curvature between segments. Intact rock
Figure 10. Stereonet plots for failure surface segments 1 through 10 showing pole orientation distributions for four
window sizes and overall radii of pole distributions
bridges are defined as the portion of intact rock separating
open fractures (e.g. Diederichs 2003; Tuckey and Stead
2016). Rock bridges on the Elephant Rock failure surface
are interpreted to have existed where steps of rock aligned
with bedding and joint set 2 are now visible (segment 4 and
part of segments 9 and 10), as shown in Figure 11. These
steps are interpreted to have been zones of redirection for
propagating fractures during failure.
Photos of the Elephant Rock taken before the collapse
show a fracture trace from the bottom to the top of the
formation that aligns with the 2016 failure surface (Figure
12). The top and bottom portion of this fracture trace,
represented by segments 1 and 6, appear more open,
whereas the portion represented by segment 4 is tighter.
Figure 11. 3D photogrammetry model of failure surface
showing steps of bedding and joint set 2
5 DISTRIBUTION OF FAILURE DEBRIS
Based on the distribution of failure debris after the 2016
collapse, the failure appears to have occurred in at least
two different blocks or stages (Figure 13). The southern
block includes segments 1 to 6, and the northern block
includes segments 7 to 10. The debris of the southern
portion is concentrated in two different areas and may
therefore be comprised of two sub-blocks. The northern
portion is more fragmented than the southern portion. This
is likely attributed to the fracture density of this portion
being higher. Further study using geomechanical
numerical modelling tools is underway to analyze the
failure sequence in more detail.
6 METEOROLOGICAL EFFECTS
Freeze-thaw cycles and heavy rain fall events are known
triggers of rockfalls (e.g. Matsuoka 2019). A database of 24
rockfalls in the Hopewell Rocks Park with exact dates was
used to analyze meteorological data leading up to recorded
rockfall events. The meteorological data was obtained
through the Government of Canada’s (2021) historical
weather database. The rockfall database includes more
recent rockfalls recorded by author K. Snair, Public
Relations and Marketing at Hopewell Rocks Park, and
Figure 12. Before collapse photo of Elephant Rock showing
a fracture trace along 3 segments of 2016 failure surface
Figure 13. Photo from above Elephant Rock after collapse
showing the distribution of failure debris in two blocks
rockfalls reported in rock scaling reports from 1977 to 2003
(Tourism New Brunswick 2003).
The park has a rock scaling contract, which involves a
scaling team removing any detached rock blocks from the
cliff face and sea stack formations before they fall on their
own. The scaling team visits annually before the opening
of the park for approximately four days to clear the cliffs
and sea stacks of potential rockfall hazards, which likely
reduces the number of rockfalls that may otherwise occur.
They also identify any areas of higher rockfall risk that
should be cordoned off to the public. In this analysis, some
rockfalls during the off-season (mid-October to mid-May)
are excluded as their precise dates were not recorded.
6.1 Freeze-Thaw Cycles: 2016 Elephant Rock Failure
Freeze-thaw occurs when water infiltrates into fractures
and micropores in a rockmass and freezes at low
temperatures, expanding about 9% of the original volume.
This expansion induces tensile stresses and fracture
propagation in the intact rock. When the ice melts, water
flows through any fractures increasing the damage (Chen
et al. 2004). Freeze-thaw would be most prominent
between the maximum and minimum tidal elevations,
where the rock is saturated from the tide and near the top
of the formation where precipitation infiltrates the rock. The
salinity in the Bay of Fundy ranges from approximately 29.5
to 31 ppm (Fisheries and Oceans Canada 2013); therefore,
the seawater's freezing point is approximately -1.6°C
(Fujino et al. 1974). The daily maximum and minimum
temperatures recorded from January 1 to March 31, 2016
were analyzed to determine any periods when freeze-thaw
processes occurred. 49 daily freeze-thaw cycles were
recorded during that period. Nine consecutive daily freeze-
thaw cycles occurred immediately before the 2016
Elephant Rock failure, which was the greatest number of
consecutive daily freeze-thaw cycles recorded in that
period. The extended period of freeze-thaw likely
contributed to destabilizing the rockmass.
6.2 Freeze-Thaw Cycles: All Recorded Rockfalls
Of the 24 rockfalls in the park recorded with exact dates,
54% occurred between January and May, when freeze-
thaw would have the greatest effect. Nearly all rockfalls
during this period experienced numerous freeze-thaw
cycles in the month before failure except for two in mid to
late May which experienced none; however, all
experienced multiple in the two months before failure. No
coherent pattern developed, with some rockfalls occurring
after an extended period with no freeze-thaw and some
occurring after an extended period with many cycles. The
scaling reports from 1977 to 2013 state multiple times that
most major rock falls occur during late winter to early
spring. In 1992, fewer rockfalls were reported than in the
previous years. This was attributed to a very cold winter,
resulting in a low number of freeze thaw cycles (Tourism
New Brunswick 2003).
Seasonal thawing from late spring to early summer has
also been attributed to the release of rock blocks
(Matsuoka, 2019). At Hopewell Rocks Park, 46% of the
recorded falls occur between March and May. Many of the
larger rockfall events at the park have been recorded
during these months (Tourism New Brunswick 2003). Public
safety risks are therefore elevated, but the park is closed.
6.3 Rain Accumulation and Wind Speeds
Rain accumulation leading up to recorded rockfalls was
also investigated in this study. The time of failure is
unknown for most rockfalls; therefore, only daily rain
accumulation was considered. Of the 24 recorded rockfalls,
83% experienced more than 11 mm of rain within the week
before failure; of those, 58% experienced more than 14 mm
within 4 days before failure and 38% experienced over
15.9 mm the day of or the day prior to failure.
Only 14 rockfall records had information about wind
speed for the day of failures (Government of Canada
2021). Of the 14 rockfalls, 43% of the recorded rockfalls
experienced a maximum wind gust over 32 km/h on the day
of failure; however, hourly wind speed records show no
failures experienced significant wind speeds (over 20 km/h)
during the day of failure.
7 TIDAL EROSION
Tidal erosion is the primary mechanism for the
development of sea stacks. Hydraulic wave erosion occurs
at the base of the formation where the tide contacts the rock
(Trenhaile et al. 1998). This undercuts the rock and can
cause instability. An example of tidal erosion at the park is
shown in Figure 14, where the minimum span of the
Sentinel formation in the intertidal zone decreased by
approximately 20% over 50 to 60 years, at an estimated
maximum erosion rate of 6 mm/year. Comparable
estimates of sea stack erosion rates in the park of 5 to
52.5 mm/year were reported by Trenhaile et al. (1998).
Figure 14. Sentinel formation sea stack at Hopewell Rocks
Park showing accelerated erosion in the intertidal zone
8 DISCUSSION
Multiple factors influence the stability of sea stack
formations at Hopewell Rocks Provincial Park, including
tidal marine exposure, freeze-thaw cycles, rainfall, and high
winds. Tidal marine exposure reduces the intact strength of
the rock and increases erosion rates in the intertidal zone,
near the base of the formations. The accelerated tidal
erosion increases ground stress near the base by
undercutting the formation. From late winter to early
summer, freeze-thaw cycles induce fracture propagation
and reduce the shear strength of discontinuities,
particularly those located in the intertidal zone and near the
top of the formations that experience greater saturation
from marine exposure and precipitation. Heavy rainfall
events also reduce the shear strength of discontinuities
(e.g. Matsuoka 2019) and have been attributed to triggering
at least 38% of the rockfalls recorded in the park. Regional
joint sets are steeply dipping and oriented approximately
parallel and perpendicular to the coast, forming semi-
rectangular blocks. Bedding is also moderately inclined
(~30-35° dip). Toppling and sliding wedge failures are
possible failure mechanisms for rockfalls in the park. The
collapse of the Elephant Rock is interpreted to be a
combination of these two mechanisms. Segments 1, 2, 3,
5, 6 and 8, which comprise approximately 22% of the
Elephant Rock failure surface had the smallest standard
deviation from their best fit planes; therefore, they are
interpreted to have been pre-existing open fractures,
whereas segments 4, 7, 9, and 10 are interpreted to have
been tight fractures with rock bridges. The failure may have
occurred in at least two distinct blocks, with the larger
southern block being the leading component of the failure
because of the increased stresses at the base caused by
its greater volume and smaller base width, followed by the
northern block. The southern block failure is interpreted to
be a toppling mechanism initiated by freeze-thaw opening
along segments 1, 2, and 3 (precipitation), combined with
tidal erosion and marine freeze-thaw along segments 5 and
6, and developing fracture propagation through segment 4.
Similar failure mechanisms are interpreted to have
occurred in the north block, with the primary factor being
undercutting in the intertidal zone and fracture propagation
through segment 10.
9 CONCLUSIONS
This study presents a geomechanical interpretation of the
2016 Elephant Rock failure at Hopewell Rocks Provincial
Park through analyses of pre- and post-failure
photographs, a 3D photogrammetry model of the failure
surface, field measurements of local intact rock strength
and rockmass structures, historic rockfall database
information from the park, local sea level and tidal elevation
records, and climate records of temperature (for freeze-
thaw cycles), precipitation, and wind speed.
The failure surface was divided into 10 segments based
on orientation and curvature and is interpreted to have
been a combination of pre-existing open fractures
(segments 1, 2, 3, 5, 6 and 8) and tight fractures with intact
rock bridges (segments 4, 7, 9, and 10) shortly before
failure. Intact rock strength (UCS based on Schmidt
hammer data) is reduced with prolonged marine exposure
near the base of the formation. Freeze-thaw cycles and
rainfall are contributing factors to rockfalls in the park, and
most of rockfalls occur from late winter to early summer
when freeze-thaw is prominent. Heavy rainfall events
appeared to have triggered 32% of the recorded failures.
Wind speed does not appear to influence the timing of
rockfalls. In addition, maximum erosion near the high tide
elevation by hydraulic wave action and potential dissolution
of cements is estimated to be 6 mm/year at the park. The
failure mechanism of the Elephant Rock is interpreted to be
a multi-stage combination of toppling and sliding failures,
driven by tidal erosion that increased stresses at the base,
freeze-thaw cycles that opened fractures at the base
(marine) and top (precipitation). Ongoing research aims to
use geomechanical numerical models to investigate the
failure mechanism and progression in more detail.
The results of this study provide insight to rockfalls and
sea stack failures at Hopewell Rocks Park for evaluating
risk of future potential failures that may impact geotourism.
ACKNOWLEDGMENTS
This research was financially supported by the Queen’s
University Catalyst Fund and the Natural Sciences and
Engineering Research Council of Canada.
REFERENCES
Chen, T.C., Yeung, M.R. and Mori, N. 2004. Effect of water
saturation on deterioration of welded tuff due to freeze-thaw
action. Cold Reg. Sci. Tech., 38(2-3): 127-136. DOI:
10.1016/j.coldregions.2003.10.001.
CloudCompare (version 2.12.alpha) [GPL software]. 2021.
Available from http://www.cloudcompare.org/.
Dewez, T.J.B., Girardeau-Montaut, D., Allanic, C. and
Rohmer, J. 2016. FACETS: A CloudCompare plugin to
extract geological planes from unstructured 3D point clouds.
Int. Archives of Photogrammetry, Remote Sensing and
Spatial Info. Sciences, XLI-B5: 799-804.
doi:10.5194/isprsarchives-XLI-B5-799-2016.
Diederichs, M.S. 2003. Rock fracture and collapse under low
confinement conditions. Rock Mech. Rock Engng., 36(5):
339-381. DOI: 10.1007/s00603-003-0015-y.
Filocomo A. 2016. Structural and Geomechanical Analysis of
the 2016 Elephant Rock failure, Hopewell Rocks Provincial
Park, New Brunswick, Canada. University of New
Brunswick, Fredericton.
Fisheries and Oceans Canada. 2013. Important Marine and
Estuarine Habitat of Inner Bay of Fundy Atlantic Salmon. DFO
Can. Sci. Advis. Sec. Sci. Advis. Rep. 2013/054.
https://waves-vagues.dfo-mpo.gc.ca/.
Fujino, K., Lewis, E. L., and Perkin, R. G. 1974, The freezing
point of seawater at pressures up to 100 bars, J. Geophys.
Res., 79(12), 1792-1797, doi:10.1029/JC079i012p01792.
Galea, P., Bozionelos, G., D'Amico, S., Drago, A., and Colica,
E. 2018. Seismic Signature of the Azure Window Collapse,
Gozo, Central Mediterranean. Seismological Res. Letters,
89(3): 1108-1117, doi: 10.1785/0220170115.
Gardner, J. 2019. How water, wind, waves and ice shape
landscapes and landforms: Historical contributions to
geomorphic science. Geomorph.,366(2020)106687.
doi.org/10.1016/j.geomorph.2019.02.031.
Government of Canada. 2021. Historic climate and weather
data: Fundy, NB. Environment and natural resources.
https://climate.weather.gc.ca/climate_data/daily_data_e.ht
ml?StationID=6919 [accessed 24 May 2021].
Hanson, R.L. 1977. The Rock Provincial Park Study on
Potential Threat to Public Safety. Department of Tourism.
Jacques, Whitford and Associates Ltd. 1989. Hopewell Rocks
Site Visit, New Brunswick Department of Tourism,
Recreation and Heritage, project M1305, 8 pp.
Matsuoka, N. 2019. A multi-method monitoring of timing,
magnitude and origin of rockfall activity in the Japanese
Alps. Geomorphology, 336: 65-76. DOI:
10.1016/j.geomorph.2019.03.023.
Priest, S.D. 1993. Discontinuity Analysis for Rock
Engineering. Chapman and Hall, London. 473 pp.
Sturzenegger, M. and Stead, D. 2009. Close-range terrestrial
digital photogrammetry and terrestrial laser scanning for
discontinuity characterization on rock cuts. Eng. Geol., 106:
163-182. DOI: 10.1016/j.enggeo.2009.03.004.
Tourism New Brunswick. 2003. Rock Scaling Reports 1977-
2003. Tourism, Recreation and Heritage.
Trenhaile, A.S., Pepper, D.A., Trenhaile, R.W. and Dalimonte,
M. 1998. Stacks and notches at Hopewell Rocks, New
Brunswick, Canada. Earth Surf. Procs. Land., 23: 975-988.
doi: 10.1002/(SICI)1096-9837(1998110)23:11<975::AID-
ESP916>3.0.CO;2-K.
Tuckey, Z. and Stead, D. 2016. Improvements to field and
remote sensing methods for mapping discontinuity
persistence and intact rock bridges in rock slopes. Eng.
Geol., 208: 136-153. doi:10.1016/j.enggeo.2016.05.001.
