Voll Gger 2016

Journal of Structural Geology 85 (2016) 168e187
Contents lists available at ScienceDirect
Journal of Structural Geology

journal homepage: www.elsevier.com/locate/jsg
Mapping folds and fractures in basement and cover rocks using UAV
photogrammetry, Cape Liptrap and Cape Paterson, Victoria, Australia
Stefan A. Vollgger*, Alexander R. Cruden
School of Earth, Atmosphere and Environment, Monash University, Australia
a r t i c l e i n f o a b s t r a c t
Article history: Brittle and ductile deformation of alternating layers of Devonian sandstone and mudstone at Cape
Received 1 December 2015 Liptrap, Victoria, Australia, resulted in upright folds with associated fold accommodation faults and
Received in revised form multiple fracture sets. Structures were mapped at the Fold Stack locality at Cape Liptrap using high-
22 February 2016
resolution aerial photographs acquired by a digital camera mounted on an unmanned aerial vehicle
Accepted 26 February 2016
(UAV). Subsequent photogrammetric modelling resulted in georeferenced spatial datasets (point cloud,
Available online 3 March 2016
digital elevation model and orthophotograph) with sub-cm resolution and cm accuracy, which were used
to extract brittle and ductile structure orientation data. An extensive dataset of bedding measurements
Keywords:
Folds
derived from the dense point cloud was used to compute a 3D implicit structural trend model to visualise
Fractures along-strike changes of Devonian (Tabberabberan) folds at the Fold Stack locality and to estimate bulk
Inheritance shortening strain. This model and newly collected data indicate that first generation shallowly south-
UAV southwest plunging upright folds were gently refolded about a steeply plunging/subvertical fold axis
Digital photogrammetry during a Devonian low-strain northesouth shortening event. This also led to the local tightening of first
Point cloud generation folds and possibly strike-slip movement along regional scale faults. In order to distinguish
fractures associated with Devonian compression from those that formed during Cretaceous extension
and later inversion, we compared the five fracture sets defined at Cape Liptrap to previously mapped
joints and faults within the overlying sedimentary cover rocks of the Cretaceous Strzelecki Group
(Gippsland Basin), which crop out nearby. An east-southeast trending fracture set that is not evident in
the Strzelecki Group can be linked to the formation of Devonian folds. Additionally, hinge line traces
extracted from the Fold Stack dataset are aligned parallel to a dominant fracture set within the overlying
cover sediments. This suggests that basement structures (folds and coeval parallel faults) have an
important influence on fault and joint orientations within Cretaceous cover rocks.
© 2016 Elsevier Ltd. All rights reserved.
1. Introduction Extension fractures can be precursors to faults (Blenkinsop,

2008) and develop when the tensile strength of a rock is excee-
Geological fractures are discrete discontinuities within a rock ded. Many coalescing tensile fractures (microcracks) are under-
mass that are caused by brittle failure. They include faults, joints stood to form shear fractures (Lockner et al., 1992; Paterson and
and filled fractures such as veins and dykes (Gillespie et al., 1993). Wong, 2005; Scholz, 2002). Anderson's theory of faulting uses the
Their importance in the accumulation and production of oil and Mohr-Coulomb failure criterion to explain the orientation of con-
gas, as well as for the formation of hydrothermal ore deposits (e.g., jugate sets of brittle shear fractures in situations where two of the
orogenic gold deposits) has long been recognised (Cox and three principal compressive stresses (s1 > s2 > s3) are parallel to
Etheridge, 1987; Cox et al., 1991; Healy et al., 2012; Newhouse, Earth's surface (Anderson, 1942, 1905). According to this theory,
1942; Sibson, 1996; Towse, 1980). shear failure initiates along planes containing the s2 axis and lying
at approximately 30 to the maximum compression s1 direction,
resulting in three fundamental classes of faults (thrust, normal,
wrench/strike-slip faults) (Anderson, 1905). However, pre-existing
* Corresponding author. School of Earth, Atmosphere and Environment, Monash fabrics and structures, such as bedding surfaces in fold limbs, will
University, Clayton, VIC, 3800, Australia.
preferentially fail in shear and hybrid shear-extension failure at
E-mail addresses: [email protected] (S.A. Vollgger), sandy.cruden@
monash.edu (A.R. Cruden). angles up to 60 to the maximum compressive stress, especially in
http://dx.doi.org/10.1016/j.jsg.2016.02.012
0191-8141/© 2016 Elsevier Ltd. All rights reserved.
S.A. Vollgger, A.R. Cruden / Journal of Structural Geology 85 (2016) 168e187 169
the presence of fluid overpressure (Sibson, 2012). This means that models, triangulated irregular networks of points (TINs) and
when reactivation is a dominant failure mechanism, the funda- orthorectified photographs (orthophotographs). These datasets
mental relationship between fracture orientation and far-field enable digital extraction of orientation measurements of structures
stress can deviate significantly from theoretical predictions. such as fractures, bedding and foliation. While terrestrial laser
In this study, we aim to elucidate the role of reactivation of scanning is currently a more established method for collecting
Devonian aged structures that formed during the Tabberabberan similar point clouds, e.g., in the field of petroleum geology and rock
Orogeny and their relationship to Cretaceous structures within the slope characterisation (Abell an et al., 2014; Gigli et al., 2014;
overlying Gippsland Basin. To do so, we compare structures within Hodgetts, 2013; Jones et al., 2008, 2004; McCaffrey et al., 2010,
Devonian rocks Cape Liptrap, Victoria, Australia to those in over- 2005; Seers and Hodgetts, 2013a; Trinks et al., 2005), the use of a
lying, unfolded Cretaceous sedimentary rocks at Cape Paterson UAV combined with digital photogrammetry is comparatively less
(Fig. 1). expensive and easier to deploy. Furthermore, this approach ensures
We employ a novel approach that uses high-resolution aerial a direct connection between imagery and the resulting 3D model.
photographs that have been acquired by a small, unmanned aerial Several studies have demonstrated that digital photogrammetry
vehicle (UAV), commonly referred to as a drone. UAVs are can deliver high quality datasets with accuracies similar to tradi-
increasingly used in the geosciences (e.g., Bemis et al., 2014; tional laser scanning (Clapuyt et al., 2015; Harwin and Lucieer,
Clapuyt et al., 2015; Gomez et al., 2015; Jordan, 2015; Lucieer 2012; Leberl et al., 2010; Müller et al., 2014; Strecha et al., 2015;
et al., 2014; Niethammer et al., 2012; Nishar et al., 2016; Sto €cker Thiele et al., 2015), meaning that accurate spatial measurements
et al., 2015) due to their ability to acquire aerial photographs at can be extracted from such datasets. At the moment, the use of UAV
resolutions that cannot be achieved with traditional airborne or photogrammetry in the field of structural geology is still in its in-
satellite-based imaging techniques. The resulting aerial photo- fancy, as indicated by the limited number of related publications
graphs can resolve cm to mm scale features, thereby opening new (Bemis et al., 2014; Johnson et al., 2014; Micklethwaite et al., 2012;
possibilities for the analysis of small-scale geological structures and Vasuki et al., 2014).
textures. Advances in the field of computer vision (i.e., Structure- In this paper, we also elaborate on the possibilities and limita-
from-Motion) have facilitated the computation of geo-referenced tions of UAV photogrammetry for applications in structural geology.
dense three-dimensional (3D) point clouds from a sequence of We discuss and explain the steps of 1) data acquisition using a
overlapping two-dimensional (2D) images. Other spatial datasets small, commercially available UAV and digital camera; 2) data
can be derived from the point cloud, such as digital elevation processing using commercial digital photogrammetry software;
Fig. 1. Geological map of southeastern Victoria, with field areas labelled in red. Cretaceous rocks of the Gippsland Basin (Strzelecki Group), which crop out at the Caves and Eagles
Nest localities are underlain by the Devonian Liptrap Formation that is exposed around the coast of Cape Liptrap. Data source: Seamless Geology Victoria-2014 edition. (For
interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
170 S.A. Vollgger, A.R. Cruden / Journal of Structural Geology 85 (2016) 168e187
and 3) the extraction of orientation data from dense point clouds This tectonic event overprinted the structural framework of Cape
and orthophotographs using open source software programmes. Liptrap and led to the development of the Gippsland Basin, which
This workflow is followed to acquire and analyse orientation crops out at Cape Paterson.
measurements of bedding and fractures in Devonian basement
rocks (Liptrap Formation; Fold Stack locality) and overlying Creta- 2.2. Cape Paterson
ceous cover sequences (Strzelecki Group; Eagles Nest and Caves
localities) exposed along coastal outcrops in south central Victoria, Unfolded sedimentary rocks of the Cretaceous Strzelecki Group
Australia. A 3D implicit structural trend model is also derived from crop out along wave-cut platforms at the Eagles Nest and the Caves
digital bedding measurements from the Fold Stack locality in order localities in the east part of the Cape Paterson study area (Fig. 1).
to visualise and assess the spatial variation of first and second order The Strzelecki Group is dominated by volcanogenic clastic sedi-
folds and to estimate bulk shortening strain. The formation of ments and makes up the lowermost unit of the Gippsland Basin,
brittle structures and folds in the Liptrap Formation and Strzelecki which unconformably overlies Palaeozoic basement rocks (Duddy,
Group is then linked to the regional evolution of the Lachlan Oro- 2003a; Norvick et al., 2001). The Gippsland Basin generally trends
gen and later modifications during the Cretaceous break-up of eastewest and is 400 km long and 80 km wide. The basin devel-
Gondwana and subsequent Tertiary inversion. oped as part of the southern Australian rift-related graben system;
significant subsidence started in the Early Cretaceous, ultimately
2. Geological setting leading to the accumulation of up to 14 km of Cretaceous and
Cainozoic sediments (Duddy, 2003a; Wilcox et al., 1992). Willcox
The Cape Liptrap and Cape Paterson study areas are located on et al. (1992) suggest four main stages of rifting based on fault ori-
the coast of Victoria, Australia, around 150 km and 115 km south- entations and crosscutting relationships in the onshore part of the
east of Melbourne, respectively (Australia) (Fig. 1). At Cape Liptrap, basin. The oldest structures related to stage 1 are normal faults that
the Devonian basement rocks that underlie the Gippsland Basin are were active during deposition of the Strzlecki Group and strike
exposed along wave-cut platforms and cliffs. The Strzelecki Group between northeast-southwest and eastewest (Willcox et al., 1992).
of the Cretaceous Gippsland Basin crops out at and around Cape Stage 2 faults strike northwest-southeast, have normal displace-
Paterson, approximately 33 km northwest of Cape Liptrap. ments of typically less than 5 m and are sometimes intruded by Late
Cretaceous mafic dykes (Willcox et al., 1992). Stage 3 faults
2.1. Cape Liptrap comprise northesouth trending faults with normal displacements
of up to 120 m and they offset the earlier two fault populations
The Cape Liptrap study area is situated in the eastern part of the (Willcox et al., 1992). Later north-northwest south-southeast
Melbourne Zone, which is one of ten tectonostratigraphic zones of compression (stage 4) led to two groups of reverse faults, striking
the Palaeozoic Lachlan Orogen (VandenBerg et al., 2000). The northwest and northeast respectively that make up the majority of
Lachlan Orogen makes up the south central part of the Tasmanides onshore faults (Willcox et al., 1992). Structural maps of the Latrobe
in eastern Australia and was developed during arc-continent col- Group indicate that axial trends of anticlines range from generally
lisions along the south and east margins of the Gondwana super- north-northeast in the onshore area of the Gippsland Basin to east-
continent (Glen, 2013; Gray et al., 2006; VandenBerg et al., 2000). northeast in the offshore part (Colwell and Willcox, 1993; Duddy,
The Melbourne Zone is characterised by a thick Lower Silurian to 2003b).
Middle Devonian sedimentary sequence (Murrindindi Supergroup) Previous work by Valenta (2014) near Cape Paterson focussed on
(VandenBerg et al., 2000). The Murrindindi Supergroup conform- fractures and fault systems within the Strzlecki Group that devel-
ably overlies Upper Ordovician sedimentary rocks and was later oped during the Cretaceous rifting between Australia and
intruded by Middle and Upper Devonian granites, some of which Antarctica and later Tertiary inversion. Detailed mapping of these
are exposed at Wilson's Promontory (Fig. 1) (VandenBerg et al., structures was carried out using high-resolution orthophotographs
2000). The Liptrap Formation is part of the Murrindindi Super- that were acquired and processed by the first author of this study
group and consists of 3200 m of finely laminated quartz-rich silt- using a similar UAV and digital photogrammetry workflow to that
stones and shales with interbedded sandstone (turbidites) outlined below (Section 3.1).
(Lindner, 1953; O'Connor, 1978; Talent, 1965). During dominantly
eastewest shortening associated with the Middle Devonian Tab- 3. Methodology
berabberan Orogeny, the Liptrap Formation was deformed and
metamorphosed to lower greenschist facies and developed a weak 3.1. UAV photogrammetry
axial plane cleavage (Gray et al., 1999; Lennox and Golding, 1989;
VandenBerg et al., 2000). At the Fold Stack locality at Cape Lip- The term UAV photogrammetry was first introduced by
trap, this tectonic event formed a 100 m-scale anticlinal structure Eisenbeiss (2009) to describe the digital photogrammetric pro-
with associated parasitic folds (O'Connor, 1978). Deformation and cessing of images that have been acquired by a remotely
related metamorphic fluids that were derived from within the controlled, semi-autonomously or autonomously operating
sedimentary succession have been linked to several sets of quartz airborne platform. Digital photogrammetry is used to compute
veins in the study area, which occur predominantly in sandstone georeferenced and coloured dense point clouds, digital elevation
layers (Lennox and Golding, 1989). In the eastern part of Cape models and orthophotographs from a sequence of overlapping
Liptrap, the major steeply dipping, northeast-southwest trending images that have been taken from various viewpoints. With UAV
Waratah Fault juxtaposes the Devonian Liptrap Formation in the photogrammetry, these images are collected with a camera that is
west against Devonian carbonates in the east (Gardner et al., 2009; mounted to a moving UAV. This setup bridges the gap between
VandenBerg et al., 2000) (Fig. 1). Kinematic indicators within brittle terrestrial and classical aerial photography from manned aircraft
fault rocks suggest significant sinistral displacement along the or satellites. One essential advantage of UAVs is their ability to
Waratah Fault (VandenBerg et al., 2000). systematically and inexpensively acquire high-resolution aerial
About 230 million years after the Devonian Tabberabberan photographic datasets at altitudes between ~10 to a few 100 m.
Orogeny, Cretaceous rifting led to the breakup of Gondwana and They can also be used to acquire data in areas that are difficult or
sea floor spreading between Antarctica and Australia (Webb, 1991). dangerous to access, such as coastal cliffs, wavecut platforms,
Fig. 2. A sequence of 211 overlapping aerial photographs were acquired by a UAV over an area of approximately 13,000 m2. A total of 9 parallel, southwest-northeast trending flight
lines with a line spacing of 13 m were flown. Images were taken automatically at one-second intervals at a horizontal flight speed of 4 m per second. These survey settings allowed
sufficient image overlap for the majority of the survey area. Ground control points were surveyed before the flight and used for geo-referencing purposes.
landslide/rock fall areas or steep mine pit and quarry walls. apparent displacement of the position of a body with respect to a
Depending on the camera setup and the distance to the object of reference point caused by a shift in the point of observation (i.e.,
interest, aerial photographs captured from a UAV deliver cm to
mm spatial resolution. The associated ground sampling distance
(GSD) describes the distance between pixels measured on the Table 1
ground and has an inverse relationship to the spatial resolution, Calculated survey parameters and camera properties/settings for UAV photogram-
metry at the Fold Stack locality. Note that altitude above ground (AGL) is not equal to
hence a lower GSD relates to a higher resolution. For nadir aerial altitude of the UAV mentioned in Fig. 2. The latter uses mean sea level as reference
photographs (camera pointing vertically downwards), the GSD datum.
can be calculated using the following equation:
Camera Image resolution 24 Megapixel
Image width (Wimage) 6000 pixel
AGL Wsensor Image height (Himage) 4000 pixel
GSD ¼ (1)
Wimage f Sensor width (Wsensor) 23.5 mm
Sensor height 15.6 mm
Focal length (f) 16 mm
where GSD is the ground sampling distance in meters per pixel, AGL Shutter speed (Tshutter) 1/320 s
is the altitude above ground level in meters, Wsensor is the physical ISO 200
width of the camera sensor in mm, Wimage is the width of the digital Aperture f/5.6
image in pixels and f is the focal length in mm. GSD is an important UAV survey Altitude above ground (AGL) 35 m
Ground sampling distance (GSD) 8.6 mm/pixel
survey parameter because it dictates the smallest possible features
Image side overlap (OLside) 75%
that can be identified and measured in aerial photographs. It fol- Image forward overlap (OLforward) 85%
lows from Eqn. (1) that GSD can be improved by decreasing AGL (by Flight line spacing (Sflightline) 13 m
flying closer to the ground or object) or by increasing the sensor Horizontal speed (Vhorizontal) 4 m/s
resolution and f. Shutter interval (Tinterval) 1.3 s
Motion blur (MB) 1.3 pixel
Digital photogrammetry makes use of parallax, which is the
Table 2 (Lowe, 2004, 1999) are employed to identify and match features in
Coordinates and estimated horizontal (HRMS) and vertical (VRMS) position errors all images. Next, during the so-called bundle adjustment, recog-
(in meters) for the 12 ground control points (GCPs) at the Fold Stack locality. For each
GCP, 30 readings have been taken and averaged using a ProMark 500 RTK GPS. All
nised features are used to compute camera positions, camera ori-
coordinates in GDA94/MGA zone 55. For location of GCPs, refer to Fig. 2. entations and the scene geometry, resulting in a sparse point cloud.
Following the bundle adjustment, a dense point cloud can then be
GCP X/Easting Y/Northing Z/Elevation HRMS VRMS
computed based on multi-view stereo algorithms (MVS). 3D
15 405697.6331 5693974.8635 5.3455 0.0120 0.0183 meshes (including texture), digital elevation models (digital terrain
16 405680.3875 5693974.8394 4.6720 0.0121 0.0183
models and digital surface models) and orthophotographs can also
17 405690.8863 5693991.2189 4.6076 0.0184 0.0274
18 405673.7670 5693962.4593 6.0838 0.0127 0.0189 be derived from the point cloud.
19 405664.2564 5693967.3838 5.3168 0.0119 0.0182 An image overlap of 60%e85% (Agisoft LLC, 2015) is recom-
20 405653.6977 5693944.7051 7.4039 0.0147 0.0229 mended to ensure homography between images that are used for
21 405657.8191 5693927.0414 8.1831 0.0193 0.0219
digital photogrammetry. This is also important when images are
22 405646.2387 5693903.8964 6.6875 0.0127 0.0191
23 405633.1107 5693910.4497 6.0729 0.0130 0.0190
captured with a digital camera that is mounted to a moving UAV.
24 405650.6835 5693890.6896 7.1724 0.0131 0.0188 For a given camera setup and desired GSD (Eqn. (1)), the shutter
25 405636.5859 5693880.3917 8.8789 0.0142 0.0213 interval Tinterval in seconds (also referred to as image acquisition
26 405662.5336 5693895.2577 7.4596 0.0307 0.0463 frequency) can be calculated from:
camera location) (Allum, 1966). The Structure-from-Motion algo- Himage GSD 100 OLforward
Tinterval ¼ (2)
rithms that are employed in digital photogrammetry programmes Vhorizontal 100
compute the scene geometry based on parallax. In a first step,
feature detectors such as the scale invariant feature transform where Himage is the height of the digital image in pixels (with
camera sensor width oriented perpendicular to flight direction),
Fig. 3. The main steps of a digital photogrammetry workflow: a) First, the overlapping digital aerial photographs and ground control points are imported into the software package.
Features within images are then automatically detected, matched and used to calculate locations and orientations of camera stations. b) Tie points that have been used for the image
alignment form an initial sparse point cloud (10,429 points). Because of the incorporation of ground control points, the sparse point cloud is already geo-referenced. c) After point
densification (multi-view stereo algorithm), points that were manually classified as representing the ground surface (grey, water in blue) were exported as a dense point cloud (142
Mio points) for further analyses (here using CloudCompare). Additional datasets such as an orthophotograph and digital elevation model were computed for fracture mapping in a
GIS application (here QGIS). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 4. High-resolution orthophotograph and digital elevation model (DEM) resulting from the digital photogrammetry workflow. Note the two 1e1.5 m thick folded sandstone
marker beds in the core of Fold Stack overlain by shale-dominated units. Small inset shows photograph of the ~30 m high Fold Stack, which is surrounded by wave-cut platforms.
OLforward is the percentage of the desired image overlap in the flight

direction (e.g., 75) and Vhorizontal is the horizontal flight speed in the 100 OLside
Sflightline ¼ GSD Wimage (4)
flight direction in m/s. Commonly, calculated Tinterval is used as an 100
upper constraint, but when possible, pictures are taken at smaller
where Wimage is the width of the digital image in pixels (with
intervals for redundancy purposes.
camera sensor width oriented perpendicular to flight direction)
The quality of the aerial photographs themselves will have a
and OLside is the percentage of image overlap between flight lines
major influence on the resulting photogrammetric model. Motion
(e.g., 75%).
blur, which is the apparent streaking caused by the motion of the
camera when taking a photograph, therefore needs to be mini-
mised. For a given Vhorizontal and GSD, the motion blur MB in pixels 3.1.1. UAV e unmanned aerial vehicle
for objects moving across the frame can be estimated from: Unmanned aerial vehicles are remotely piloted aircraft which
are generally classified into fixed wing, multi-rotor, hybrid and
blimp/airship platforms. In this study, we employ a multi-rotor UAV
with a total take-off weight below 10 kg, which can be used for
Vhorizontal Tshutter small and medium scale (tens of m2 to up to several km2) aerial
MB ¼ (3) surveys. For such multi-rotor UAVs, flight times vary between 10
GSD
and 25 min per standard lithium polymer (LiPo) battery. Compared
where Tshutter is the camera exposure time in seconds. Motion blur to fixed wing UAVs, multi-rotors can operate in confined environ-
can therefore be minimised by either reducing the horizontal flight ments, take-off and land vertically, hover over one spot and fly at
speed and exposure time, or by increasing GSD. The latter translates low horizontal speeds. The latter is important to maintain optimal
to flying at a higher altitude for a given camera setup and will result camera settings (shutter speed, ISO, aperture) to minimises motion
in a lower spatial resolution. blur and ensure high quality of images captured at low altitudes.
For aerial surveys where photographs are acquired along par- There is currently a large number of consumer and commercial
allel flight lines, the required flight line spacing Sflightline in m is UAV manufacturers. In this study, we employed a DJI S800 EVO
given by: hexacopter that was equipped with a DJI A2 flight controller and
(small aperture) to ensure a large depth of field. Lenses with fixed

focal lengths are lighter than their zoom counterparts, which is
why they are commonly used for UAV photogrammetry. Such len-
ses typically have better optical performance (i.e., sharpness) than
zoom lenses because they are optimised to a specific focal length. In
this study, we used a 16 mm f/2.8 lens attached to the Sony Nex 7
camera (c.f., Strecha et al., 2015).
4. Acquisition of aerial images and ground control points
Prior to the acquisition of aerial photographs by UAV, an A4

sized stencil and high-visibility removable spray paint were used to
mark 12 randomly distributed ground control points at the Fold
Stack locality (Fig. 2). The coordinates of these points were sur-
veyed for subsequent georeferencing purposes using a dual fre-
quency ProMark 500 RTK GPS device with an accuracy of 1e5 cm
(Table 2).
On the day our survey was flown, overcast weather and light to
gentle winds provided ideal and safe operating conditions. The UAV
take-off time was set to coincide with low tide to ensure maximum
exposure of outcrops along the wave-cut platforms that surround
the Fold Stack locality. Before the start of the UAV survey, the
necessary camera parameters and survey settings were calculated
based on Eqns. (1)e(4) (Table 1). A total number of nine parallel,
northeast trending flight lines with a nominal spacing of 13 m were
flown with eight flight lines at an average altitude of 38 m and an
additional flight line at 53 m in the topographically highest central
Fig. 5. Oblique view of part of the coloured dense point cloud (RGB values derived
from the original aerial photographs). Hinge points are manually picked and a plane- part of the survey area (Fig. 2). The horizontal flight speed was
fitting tool was used to compute the axial plane orientation (dip and dip direction) limited to 4 m/s to minimise motion blur. Such a low horizontal
using CloudCompare software. The same workflow was used for the extraction of flight speed was also necessary to achieve well-balanced camera
bedding measurements. See Fig. 3 for location. settings (exposure time, ISO, aperture) and ensure sharp and
correctly exposed images with low noise. The camera was triggered
automatically during the whole flight at one-second intervals,
wirelessly connected to a DJI iPad ground station (www.dji.com).
giving a flight-parallel camera station spacing along each flight line
This setup allowed to pre-program a sequence of waypoints on a
of ~4 m. The total flight time was 25 min (3 autonomous flights),
tablet computer so that the UAV will fly along a grid at a set speed
including travelling approximately 700 m (one way) between the
and altitude. This ensures the systematic acquisition of aerial
take-off location and the survey area. After survey completion, all
photographs with a constant line spacing and sufficient forward
images (24 megapixels, JPEG format) were downloaded from the
and side overlap. A damped gimbal attached to the hexacopter
camera and sorted. Pictures acquired outside the area of interest
automatically stabilises the camera along three axes and removes
were omitted, resulting in 211 aerial photographs for the Fold Stack
vibrations inherited from the motors and the airframe. Further-
locality. These images were subsequently fed to a digital photo-
more, the gimbal enables the collection of perfect nadir images
grammetry workflow (Fig. 3a).
(camera pointing vertically down), even when the UAV pitches or
rolls during the flight. The gimbal also allows remote adjustment of
5. Data processing and digital structural mapping
the orientation of the camera, facilitating the acquisition of high
oblique (images that include the horizon) and low oblique (images
5.1. Digital photogrammetry
that exclude the horizon) aerial photographs.
The nadir aerial photographs of the Fold Stack locality acquired

3.1.2. Digital camera and lens (payload) by UAV were processed using Agisoft PhotoScan Professional digital
A detailed description and comparison of digital cameras and photogrammetry software (version 1.1.6) (Agisoft LLC, 2015). The
lenses is outside the scope of this study. However, it is important to focal length and image dimensions were automatically detected
note that digital cameras with larger sensors (i.e., APS-C or full after digital images were imported by the program and then used
frame) generally produce better images due to higher signal-to- for the subsequent calibration of the intrinsic parameters of the
noise ratios. Besides the physical size of the sensor (Wsensor), the camera (principal point coordinates, distortion coefficients). An
sensor resolution (proportional to Wimage) is crucial because it initial low-quality image alignment was then performed in order to
controls the GSD, as demonstrated in Eqn. (1). Therefore, high- spatially locate the non-geotagged images within a relative coor-
resolution camera sensors are preferred. For this study, we dinate frame to assist with manual identification of ground control
employed a Sony Nex 7 mirrorless digital camera, which has an points. The 12 surveyed ground control points were marked in all
APS-C sized sensor (23.6 15.6 mm) and a resolution of 24 meg- the images in which they occurred. The photogrammetry software
apixels. The camera was automatically triggered at a set time in- uses these ground control points for both geo-referencing purposes
terval using an infrared intervalometer. and to optimise extrinsic parameters such as estimated camera
The optical performance of the camera lens also dictates the locations and orientations. This improves the accuracy of the final
quality of the acquired images. A suitable lens for digital photo- model. After the image alignment process, the locations and ori-
grammetry does not require a wide aperture (¼low f-number, e.g. f/ entations of the images were established and a sparse point cloud
1.4) because images are generally captured at higher f-numbers was computed. Non-static and reflective objects such as water
Fig. 6. a) Fracture sets and axial traces at the Fold Stack locality mapped digitally on the high-resolution orthophotograph. b) Circular histogram (weighted by fracture segment
length) for 791 joints and faults mapped at a larger scale. c) Circular histogram for 27 mapped axial traces (dashed brown lines) weighted by segment line length. d) Fractures and
faults were classified into five main sets, based on orientation peaks on circular histograms, and colour-coded based on their orientation (for set #2, see Fig. 7). The bin size for
circular histograms is 5 . For fractures mapped in Area A, refer to Fig. 7. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version
of this article.)
bodies or waves that are captured in the images are problematic for the point cloud affected by water movement (i.e., ocean side of Fold
digital photogrammetry algorithms, resulting in some erroneous Stack) displayed non-static artefacts, which were manually classi-
tie point positions, which can have a negative effect on the overall fied as water and ignored in later processing steps. The final dense
accuracy of the final model. In our case, these were manually point cloud for the Fold Stack location consists of ~100 million
removed, resulting in a sparse point cloud with a total of 10,429 points. Holes appear in the dense point cloud in topographically
points (Fig. 3b). Optimisation of the camera locations and orienta- complex areas (e.g., overhanging rocks) that were not covered by
tions based on the ground control points reduced the total spatial nadir aerial photographs as well as in regions with physical oc-
error within the sparse point cloud to ~0.01 m (x ¼ 0.0072 m, clusions or low image correspondence. Adding an additional set of
y ¼ 0.0055 m, z ¼ 0.0058 m). Subsequent point cloud densification low or high oblique aerial photographs could have circumvented
generated a final 3D model with 142 million points for an area of such issues.
approximately 13,000 m2 (Fig. 3c). Every point within the dense A digital elevation model and an orthophotograph with ground
point cloud is associated with the following 9 attributes; x,y,z co- resolutions of 7.7 mm/pixel were derived from the dense point
ordinates within a defined coordinate system, colour channel (red, cloud (Fig. 4). Unlike the dense point cloud, there are no holes
green, blue digital numbers) information from the original photo- visible in the digital elevation model and orthophotograph because
graphs, and orientation data described by the x, y and z compo- they are vertical projections derived from the original 3D point
nents of the vertex normal vector (nx, ny, nz) for each point. Areas of cloud, hence all areas are covered by nadir aerial photographs.
Fig. 7. a) Detailed mapping of fractures within a sandstone layer in the western part of the Fold Stack locality (see Fig. 6 for location). c) The circular histogram (weighted by joint
segment length) highlights four dominant joint orientations based on a total of 1876 joints. c) Fractures are colour-coded based on the fracture set classification. The bin size for
circular histograms is 5 . (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Finally, the point cloud, orthophotograph and digital elevation orthophotograph.

model were exported in ASCII and geoTIFF file formats respectively
for further structural analysis and interpretation.
5.2.1. Bedding plane measurements
Extraction of bedding orientation data from the dense 3D point
5.2. Digital structural mapping cloud has an advantage over triangulated meshes with projective
textures. The use of large texture maps on multi-million vertex
Digital mapping on high-resolution spatial datasets permits the models is a well-known bottleneck in computer graphics and
extraction of accurate structural measurements from difficult to makes the visualisation and rendering of textured meshes slow and
access areas in a fraction of the time it takes to perform traditional impractical on conventional hardware. In contrast, coloured point
fieldwork (García-Selle s et al., 2011; Hodgetts, 2013; McCaffrey clouds with even hundreds of millions of points show a much
et al., 2005; Trinks et al., 2005). Digital mapping on a computer better performance on the same hardware. Moreover, depending
screen also permits rapid changes of view, which assists the on the magnification level and point size, dense point clouds look
interpreter in assessing spatial and structural relationships in the similar to meshed surfaces and are therefore well suited for digital
dataset. Bedding, S0, and axial plane orientations at the Fold Stack mapping.
locality were digitally extracted from the dense point cloud. Joints In this study, we used the open-source cloud editing program
and faults were mapped directly on the high-resolution CloudCompare (EDF R&D and Telecom Paristech, 2015; Girardeau-
Fig. 8. 3D implicit structural trend model derived from bedding measurements (yellow discs) that were digitally extracted from the dense point cloud. Structural trend surfaces
visualise the general geometry of the Fold Stack locality (green and blue structural trend surfaces were clipped to make central parts visible). Fold axis orientations extracted from
the implicit model are in agreement with fold axis orientations calculated directly from bedding measurements (Fig. 12).
Fig. 9. The 3D implicit structural trend model was used to acquire wavelength l and arc length L measurements for the first-order anticline, which forms the Fold Stack. A bulk
shortening strain of ~26% was estimated from these data.
Montaut et al., 2005) to extract structural data from the geo- advantages of an alternative plane fitting algorithm that was pro-
referenced dense point cloud. CloudCompare applies a least posed by Jones et al. (2016), which is well suited for planar frac-
squares method to fit a plane to a group of selected points from tures, but not ideal for folded layers and beds. In contrast to
which the user can extract planar structural measurements (Fig. 5). bedding measurements, orientations for axial planes were
This plane-fitting method employs the eigenvalues and eigenvec- computed from 27 digitised hinge lines of F1 folds that were traced
tors of the covariance matrix of selected points, following an in areas with sufficient topography within the dense point cloud.
approach outlined by Woodcock (1977). The smallest of the three This allowed to minimise collinearity and guarantee reliable mea-
eigenvectors describes the pole to the best fit plane from which surements (see Fern andez, 2005 and Seers and Hodgetts, 2016 for
strike and dip values can be calculated (Cracknell et al., 2013). more details about collinearity and coplanarity). Additionally, all
A total of 2146 bedding measurements were extracted from the digitally extracted measurements were visually assessed in 3D in
~13,000 m2 Fold Stack survey locality. Bedding measurements were order to identify and eliminate outliers. Bedding and axial planar
solely extracted from planar areas within the dense point cloud by data were subsequently plotted and analysed on lower hemisphere
selecting 4 to 9 evenly distributed points within a small circular equal area stereographic projections (stereonets). Bedding mea-
area along a bedding plane. Due to the curved nature of these surements were also used to compute a 3D implicit structural trend
planes at the Fold Stack locality, qualitative properties for these model (Section 5.3).
measurements such as coplanarity were not really meaningful.
Additionally, the curvature of these planes diminishes the
orthophotograph, hence smaller-scale features were not mapped in

the dataset. Due to the textural complexity of the orthophotograph,
automated or semi-automated edge detection algorithms (e.g.,
Vasuki et al., 2014) generated a number of false positives that
required additional modifications, whereas manual digitisation
resulted in a more consistent dataset. For the Eagles Nest and the
Caves localities, fractures and faults were mapped or manually
traced on orthophotographs by Valenta (2014).
Circular histograms of fracture orientations for all localities were
calculated using the QGIS plugin ‘line direction histogram’ by Tveite
(2015). Fractures were grouped into 5 azimuth class intervals and
weighted based on their segment line length to highlight the ori-
entations of dominant fracture sets. Joints and minor faults at the
Fold Stack locality were mapped at a larger scale over the entire
~13,000 m2 outcrop area (Fig. 6) and in detail along a well exposed
~1 m thick sandstone bed located in the core of the main fold
structure (Fig. 7).
5.3. Computation of a 3D implicit structural trend model
3D implicit modelling interpolates spatial data by computing a

scalar field (also referred to as potential field), which offers a
number of advantages over traditional explicit modelling tech-
niques that are based on manual digitisation of maps and cross
Fig. 10. Poles to axial planes derived from the dense point cloud plotted on an equal sections (Calcagno et al., 2008, 2006; Carr et al., 1997; Caumon
area lower hemisphere stereographic projection. Note that axial planes are generally et al., 2013; Chiles et al., 2004; Cowan et al., 2003; Frank, 2006;
dip steeply east or west. Axial planes for the central anticlines and syncline of the Fold
Jessell et al., 2014; Lane et al., 2007; Ledez, 2003; Maxelon, 2004;
Stack locality are highlighted in red (for location, see Fig. 6). (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of
Mcinerney et al., 2005; Vollgger et al., 2015). The implicit struc-
this article.) tural trend model of the Fold Stack locality encompasses a scalar
field that is described by a spatial interpolation function (i.e., radial
basis function). This function was computed from 2146 digitally
5.2.2. Fracture mapping extracted bedding measurements using the geological modelling
Joints and faults were manually traced on the computed high- programme Leapfrog Mining 2.6 (ARANZ Geo, 2014) (Fig. 8). The
resolution orthophotograph of the Fold Stack locality using a isosurfaces of the resulting scalar field were used to visualise the
Geographic Information System (GIS) programme (QGIS internal 3D geometry of first and second order folds at the Fold
Development team, 2015). Mapped joints and faults had to be at Stack locality. Implicit modelling works best in areas with high data
least 8 mm in width or length to be resolved in the high-resolution point density where isosurfaces can be sufficiently constrained
Fig. 11. Variations in fold geometry and development of hinge line curvature in pure shear analogue modelling experiments (modified after Dubey and Cobbold, 1977) at a) 9%
shortening, and b) 36% shortening. Note that the general curvature of fold axial traces decreases with increasing shortening strain. Direction of compression shown by arrows.
6. Results and interpretation
6.1. D1-related folding and faulting
Eastewest shortening related to the Devonian Tabberabberan

Orogeny (VandenBerg et al., 2000) is responsible for the formation
of shallowly south-southwest plunging F1 folds. The core of the Fold
Stack is dominated by two 1e1.5 m thick sandstone layers that have
been deformed into a pair of upright, slightly asymmetric anticlines
joined by a close syncline (Figs. 4 and 9). These are second-order
folds to a major F1 anticline that represents the Fold Stack
outcrop. The wavelength l and arc length L of this anticline were
measured from the 3D implicit structural trend model (l/2e65 m,
L/2e88 m) (Fig. 9) and used to estimate the shortening strain e
following Hudleston and Treagus (2010):
l
e¼ 1 (5)
L
Assuming there is no significant displacement in the hinge di-
rection, Eqn. (5) gives a shortening of approximately 26% for the
main anticlinal structure. Well-developed quartz fibre slickenlines
oriented orthogonal to hinge lines and associated steps observed
along S0 in F1 limbs indicate reverse shear and are consistent with
upright flexural-slip folding (Donath, 1962). Sandstone layers also
maintain a constant orthogonal thickness throughout the fold
profile, except in areas where fold accommodation faults are
developed. The main anticline of the Fold Stack locality can there-
fore be classified as a type 1B (parallel fold, after Ramsay, 1967)
which developed by a flexural-slip mechanism. This mechanism is
also in agreement with fold accommodation faults described below,
which developed progressively during F1 fold amplification (Price,
1965).
The interlimb angles of second-order anticlines and synclines
within the core of the Fold Stack increase from west to east indi-
cating a change from close to open folds. Axial planes of these folds
dip steeply west (76 e85 ), but other lower-order folds at the Fold
Stack locality have steeply east-dipping axial planes (Figs. 4e10). A
component of simple shear associated with D1 west-over-east
thrusting is also in agreement with net tectonic transport to the
east and the dominant east vergence in the Melbourne Zone, as
suggested by VandenBerg et al. (2000).
Folds in mudstone-dominated packages which crop out in the
southernmost part of the Fold Stack locality share the same to-
pology as their underlying sandstone packages (two anticlines
connected by a syncline). However, their profile shapes vary along
the hinge line, from rounded in the sandstone layers to chevron-
shaped in mudstone-dominated rock packages. Additionally, the
wavelength decreases from 12 m in the north to 5 m approximately
60 m farther south (Figs. 6 and 8). This along strike change in
wavelength causes the two anticlines to converge towards the
south. According to the 3D implicit structural trend model, the
associated strike variations of axial traces are ~6 . The slight cur-
vature in the hinge lines of the two central, second order anticlines
Fig. 12. Poles to bedding for the northern and southern part of the Fold Stack locality
plotted on an equal area lower hemisphere stereographic projections (see Fig. 4 for
is interpreted to be associated with D1 eastewest shortening. This
locations). Fold axes orientations were derived from Bingham axial distribution cal- interpretation is based on Dubey and Cobbold's (1977) analogue
culations. Fold axes swing by 14 to the east from north to south. modelling results, which showed comparable variations in 3D
shape of folds caused by initial layer-parallel shortening (Fig. 11a).
Hinge lines deviate to a much larger degree in the northernmost
(Vollgger et al., 2015). The abundance of bedding measurements parts of the survey area (Fig. 8). This is highlighted by digitally
allowed the generation of a 3D implicit structural trend model extracted bedding measurements from the northern and southern
without manual sectional digitisation or adjustments, therefore areas of the Fold Stack locality, which are compared in Fig. 12.
minimising the modelling bias. Calculated Bingham axial distributions for 1835 bedding mea-
surements in the south give a fold axis trend and plunge of 201/11,
while 311 bedding measurements from the north give a fold axis
Fig. 13. Thrust faults within the hinge region and along the eastern limb of a F1 syncline located in the core of Fold Stack (see Fig. 6 for location). Note that these faults formed during
progressive D1 shortening (fold accommodation faults) and were probably reactivated during D2. Small insets modified after Mitra (2002).
trend and plunge of 215/06. Such variation can be explained by fold located primarily within hinge regions of F1 folds, which is not the
linkage and fold propagation processes for small shortening values case at the Fold Stack locality. Hence, we interpret these faults as
of ~10% (Dubey and Cobbold, 1977) (Fig. 11a). However, the interlimb wedge thrusts, hinge wedge thrusts and out-of-syncline
limb angles that are observed at the Fold Stack locality require thrusts, which are common types of fold accommodation faults
significantly higher shortening values. A larger amount of short- as classified by Mitra (2002) (Figs. 4e13). Such faults are known to
ening, as inferred for the Fold Stack locality, should have straight- be kinematically related to flexural-slip folding (Price, 1965). Thrust
ened fold trains and reduced the variation in axial trace faults within limbs and hinges of second order F1 folds may, at a
orientations, as shown by analogue experiments (Dubey and larger scale, be kinematically linked to bedding-parallel faults
Cobbold (1977) (Fig. 11b). We therefore suggest that the hinge (detachment faults) that were initiated during east-directed
line deviations and fold axis variations of almost 14 over a distance thrusting, as reported by VandenBerg et al. (2000).
of 10e15 m cannot be related to non-cylindrical folding alone, but At the Fold Stack locality, the largest fault offsets are observed in
also require a second D2 deformation event. fold accommodation faults that are related to the formation of F1
Ubiquitous low angle thrust faults are evident in sandstone folds. No major offsets along strike-slip faults were discernible in
layers and they generally have west-over-east movement, which the dense 3D point cloud or high-resolution orthophotograph.
leads to small scale, local repetition of stratigraphy. Thrust-related Small-scale dextral strike-slip movement can be inferred from en-
strain is accommodated by ductile deformation in the mudstone echelon extensional quartz veins (Fig. 14a). A mapped sinistral
layers. Interestingly, low angle thrust faults are found along fold strike-slip fault parallel to fracture set #5 offsets the hinge line of
limbs as well as in the hinge regions of second order F1 folds the eastern second-order anticline by ~30 cm (Fig. 14b) and
(Fig. 13). If thrust faults had formed prior to F1 folding, they should therefore postdates D1 folding. However, this fault terminates to
have acted as initial heterogeneities and controlled the nucleation the west within a mudstone layer and is only of local importance.
of folds during layer-parallel shortening, as demonstrated by nu- The regional scale Waratah Fault located ~4 km east of the Fold
merical modelling and scaled analogue experiments (Abbassi and Stack locality is thought to have formed initially as reverse fault
Mancktelow, 1990; Cobbold, 1975; Fernandez and Kaus, 2014; (Janssen et al., 1998). The Waratah Fault strikes parallel to north-
Frehner, 2014; Grasemann and Schmalholz, 2012; Mancktelow, northeast trending axial traces of F1 upright folds at the Fold
1999; Schmid et al., 2008). Consequently, thrust faults should be Stack locality. Furthermore, F1 axial traces are also parallel to other
Fig. 14. Small-scale structures visible in the high-resolution aerial images. a) En-echelon quartz extension veins indicate dextral brittle-ductile shear displacement in the core of the
Fold Stack. b) Minor sinistral offset of the hinge line of the eastern anticline located in the core of the Fold Stack. c) In the eastern part of the survey area, bedding is gently folded by
D2. Fold style is quite different compared to F1.
Fig. 15. Oblique view of the digital outcrop model (¼orthophotograph draped over digital elevation model). Shallowly south plunging F1 folds, formed by D1 eastewest shortening,
are represented by the structural trend surfaces. Subsequent northesouth D2 shortening gently refolded F1 folds, leading to a change in fold axes orientations by up to 14 . Yellow
discs are bedding measurements from the Fold Stack locality.
regional scale faults such as Doomburrim and Fish Creek Faults trending folds and faults during Tabberabberan eastewest short-
(Fig. 1). This points towards a coeval formation of north-northeast ening and suggests that they are kinematically linked.
Fig. 16. Relationship between fracture orientation (x-axis ¼ strike) and normalised fracture length (y-axis) for the three study areas. a) Five fracture are distinguished in Devonian
basement rocks at the Fold Stack locality. Note the scale dependency of fracture distributions. Joints from the Fold Stack Area A (grey line) mapped at a much smaller scale differ
from the fracture population measured from the whole area (black line). Orientations of F1 axial traces do not show an obvious relationship with any of the dominant fracture sets.
b) Three main fracture sets were distinguished in the Cretaceous sediments of the Strzelecki Group (#2, #3, #5). c) Faults mapped along the coast adjacent to the Caves and Eagles
Nest localities coincide with three of the main joint sets at the Fold Stack. However, the least dominant joint set #2 coincides with the most dominant fault orientation. Fracture set
#2 also coincides with the orientation of F1 axial traces mapped in the basement (grey field).
#1 to #5, were identified at the Fold Stack locality (Figs. 6 and 7).
The north-northwest trending fracture set #2 is weakly developed
on a larger scale, but becomes apparent at a smaller scale within
Area A of the Fold Stack locality (Fig. 16a). It is oriented parallel to
the north-northwest trending axial traces of F1 folds and is there-
fore interpreted to be related to extensional bec joints (strike
joints). The east-southeast trending fracture set #4 is oriented
perpendicular to set #2 and F1 axial traces, consistent with an
origin as extensional aec joints (dip joints) that formed during D1
eastewest compression, in response to minor elongation in the
plunge direction (Stearns and Friedman, 1972). Therefore, set #2
and set #4 are likely to have been generated during the Devonian
Tabberabberan Orogeny. It is tempting to interpret fracture sets #3
and #5 as conjugate shear joints related to the same overall east-
ewest shortening event that formed F1 folds. However, at the scale
of the data, no indication of shear movement was visible. Addi-
tionally, the angle between fracture sets #3 and #5 is ~90 , which
disqualifies them as conjugate shear joints for which acute angles
of ~60 would be expected. These fractures are likely to be related
to later extensional events, as discussed below in Section 6.4.
North trending fractures of set #1 lack sense-of-movement in-
dicators at the scale of observation and are interpreted here as
tensile fractures that formed during northesouth shortening
related to D2.
6.4. Cape Paterson fracture populations
Only three main fracture set orientations were mapped within

the Strzelecki Group at the Eagles Nest and the Caves localities
Fig. 17. Faults and joints traced by Valenta (2014) in a high-resolution aerial ortho- (Figs. 17 and 18), compared to five dominant fracture sets at the
photograph at the Eagles Nest locality (Strzelecki Group, Gippsland Basin). Two Fold Stack locality. The dominant east-northeast trending set #3
dominant joint sets are evident in the circular histogram (weighted by joint segment
and the orthogonal north-northwest trending set #5 are approxi-
length). Fractures are colour-coded based on the fracture set classification used at the
Fold Stack locality (Figs. 6 and 7). (For interpretation of the references to colour in this mately parallel to joints at the Fold Stack locality (c.f., Fig. 16a and
figure legend, the reader is referred to the web version of this article.) b). Fractures of set #3 are parallel to normal faults within the
Cretaceous Gippsland Basin, which were interpreted to be related
6.2. D2-related folding and faulting to Early Cretaceous extension (Etheridge et al., 1987; Willcox et al.,
1992). On the other hand, north-northwest trending fractures of set
We propose that a minor D2 northesouth shortening event #5 may be associated with Middle to Late Cretaceous northeast-
gently refolded F1 folds along a sub-vertical to steeply-plunging southwest trending extension (Willcox et al., 1992).
fold axis (Fig. 15). Such a structural history is supported by the The dominant fracture set #4 defined at the Fold Stack locality is
presence of gently folded, moderately east-dipping mudstone- absent in the overlying Strzelecki Group, which supports the
dominated rock packages within the F2 hinge region, which are interpretation above that set #4 is Devonian in age and related to
visible in the eastern region of the high-resolution orthophoto- the Tabberabberan Orogeny. The absence of set #4 in the Strzelecki
graph (Fig. 14c). Low-strain northesouth shortening is recorded in Group suggests that pre-existing, east-southeast trending base-
several domains within the Melbourne Zone of the Lachlan Orogen ment structures were not reactivated during Cretaceous extension
(VandenBerg et al., 2000). Many Melbourne Zone gold deposits are or Tertiary inversion and subsequent relaxation. This is noteworthy,
controlled by structures that have a major component of north- as fracture set #4 is well developed in the underlying basement
esouth shortening, with the main period of gold mineralisation sequences at the Fold Stack locality (Fig. 16a). However, its unfav-
constrained to between 385 and 365 Ma (Bierlein et al., 1999), ourable orientation in respect to the stress fields acting during
which postdates the major phase of the Tabberabberan Orogeny Cretaceous extension events may have prevented reactivation.
(385e380 Ma) (VandenBerg et al., 2000). Such timing relationships Instead, two new fracture sets #3 and #5 formed as discussed
are in agreement with our proposed structural evolution. D2 might above, and these are both oriented at ~45 to set #4.
also have caused local tightening of second order F1 folds, which Fracture set #2 in the Strzelecki Group trends north-northeast,
could explain the observed eastewest variation of interlimb angles parallel to a large number of faults mapped along the coast
within the core of Fold Stack. Additionally, D2 may have locally including the Eagles Nest and Caves localities (Fig. 19). Conversely,
reactivated earlier fold accommodation faults along the limbs of F1 this fault orientation is not evident in the 100 m-scale data at the
folds close to or within F2 hinge regions. However, these fault off- Fold Stack locality (Fig. 16a and c), even though the spacing of these
sets are expected to be small due to the low-strain nature of D2, as faults is typically less than 100 m at Cape Paterson (Fig. 19).
inferred from the large interlimb angle of F2 folds. D2 is also Nevertheless, fracture set #2 coincides with digitally mapped fold
consistent with a sinistral overprint recorded in the Waratah Fault, hinge line trends at the Fold Stack locality (Figs. 6 and 16c) and
which was reactivated by Devonian northesouth shortening. regional scale basement faults (Fig. 1). This suggests that north-
northeast trending Devonian basement folds and faults were
6.3. Cape Liptrap fracture populations reactivated during Cretaceous extension or Middle to Late Eocene
compression that affected the Gippsland Basin (Smith, 1988).
Five systematic fracture sets (joints and minor faults), numbered Reactivation therefore controlled the formation of fracture set #2 in
Fig. 18. Faults and joints traced by Valenta (2014) in a high-resolution aerial orthophotograph at the Caves locality (Strzelecki Group, Gippsland Basin). Three dominant joint sets are
evident in the circular histogram (weighted by joint segment length). Fractures are colour-coded based on the fracture set classification used at the Fold Stack locality (Figs. 6 and 7).
(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
overlying sedimentary rocks. Such a structural inheritance might resolution could be collected from selected smaller areas to opti-
lead to a misalignment between fractures and the existing regional mise and accelerate such surveys. According to Eqn. (1), an AGL of
stress field at the time of formation. Pre-existing fracture orienta- 10 m in combination with a 50 mm lens attached to a camera with
tions may induce local stress field perturbation and therefore in- APS-C sized sensor would deliver a GSD of 0.8 mm/pixel. Such ultra-
fluence the attitude of newly developing fractures. Hence, the high-resolution imagery would enable mapping of other properties
estimation of the stress field orientation for some of these younger such as the thickness of small veins. In the aerial photographs ac-
structures has to be viewed with caution. This might also explain quired in this study (GSD of 8 mm/pixel), veins with thicknesses of
the wide spectrum of inferred regional extensional directions 1 cm are only ~1 pixel wide, which makes them optically difficult to
during the Cretaceous, from northwest-southeast to northeast- distinguish from the brightly coloured sandstone layers in which
southwest, that have been proposed for the Gippsland Basin (Hill they dominantly occur, which excluded them from this study.
et al., 1994; Perineck et al., 1994; Willcox and Stagg, 1990). In addition to limitations related to image resolution, the digital
extraction of structural datasets (Section 5.2) still remains a chal-
lenge, primarily due to the lack of software for viewing and per-
6.5. Limitations of UAV photogrammetry forming geological/structural assessments of very large point
clouds or textured surfaces, which are common outputs from dig-
In spite of technological (e.g., battery life/flight time), natural ital photogrammetry workflows. CloudCompare, which was used in
(e.g., weather conditions) and legal (e.g., flying within visual line of this study has been able to handle large point clouds but needs
sight) limitations, UAV photogrammetry has great potential for some further development of structural tools. As an alternative,
applications in structural geology. Even though traditional struc- Virtual Reality Geological Studio (VRGS) as used by Seers and
tural fieldwork was deliberately kept to a minimum in this study, Hodgetts (2013b) is free to academia and also allows to analyse
we were able to remotely collect a comprehensive brittle and large point clouds.
ductile structural dataset that was analysed, interpreted and used
for 3D implicit modelling. One limitation we encountered was
difficulty in extracting reliable sense-of-shear indicators from the 7. Conclusions
aerial photographs at the working scale of this study, which are
important for the kinematic and dynamic reconstruction of the We have demonstrated that spatial datasets generated from
tectonic history. Striations on fault planes or slickenlines were not UAV photogrammetry combined with digital structural mapping
resolved in the high-resolution aerial photographs, and their and 3D implicit modelling can be used for detailed analysis and
measurements had to be acquired manually in the field. While interpretation of brittle and ductile structures. At the Devonian Fold
aerial photographs with even higher resolution would permit Stack locality at Cape Liptrap, bedding measurements extracted
mapping of even the smaller features, as shown by Thiele et al. from a dense point cloud were employed to visualise the complex
(2015), this would require a lower AGL or a higher f, which trans- internal 3D structure of an outcrop and aided the quantification of
lates to increases in the UAV survey and processing times. Alter- spatial variations in Tabberabberan fold geometries and orienta-
natively, ultra-high-resolution aerial photographs with sub-mm tions. Joints and faults traced in high-resolution orthophotographs
Fig. 19. Faults mapped by Valenta (2014) along the coast adjacent to the Eagles Nest and the Caves localities (Strzelecki Group, Gippsland Basin). Three dominant fault sets are
evident in the circular histogram (weighted by fault segment length). Faults are colour-coded based on the fracture set classification used at the Fold Stack locality (Figs. 6 and 7).
Image is courtesy of Google Earth. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
of the Fold Stack outcrop were compared to structures within selected basement structures influenced the structural framework
outcrops of unconformably overlying Cretaceous Strzelecki group of the Gippsland Basin.
sedimentary rocks exposed ~33 km to the northwest at Cape
Paterson. This comparison allowed us to quantify the relationships Acknowledgements
between Devonian (Tabberabberan) compressional and Cretaceous
extensional and Tertiary inversion structures. S.A.V. and A.R.C. acknowledge Monash University for scholar-
Eastewest shortening related to the Devonian Tabberabberan ship and research support. Additionally, we acknowledge the work
Orogeny at the Fold Stack locality was accommodated by the of Daniel Girardeau-Montaut who is the main developer of
development of upright F1 folds and associated fold accommoda- CloudCompare. Steven Micklethwaite, Tom Blenkinsop and Tobias
tion faults (hinge wedge faults and limb wedge faults). The F1 folds Bauer are thanked for their constructive feedback. Anindita Samsu
were subsequently gently refolded by a Devonian low-strain, is thanked for proof-reading this manuscript. We are grateful for
northesouth D2 shortening event that caused fold axes to swing constructive feedback by Thomas Seers and one anonymous
by up to 14 at the 10e15 m scale. D2 is also interpreted to have reviewer which improved the quality of this paper. Stereonets were
reactivated earlier thrust faults that cut F1 limbs at low angles. generated using Richard Allmendinger's programme Stereonet 9
Five systematic fracture sets were identified at the Fold Stack (Allmendinger et al., 2013; Cardozo and Allmendinger, 2013). Aca-
locality, while three were identified in the overlying Strzelecki demic licenses for the Leapfrog Mining software package were
Group with common orientations. The two remaining fracture sets, kindly provided by ARANZ Geo Pty. Ltd.
which do not occur in the Strzelecki Group are interpreted to be
related to Devonian shortening events. North-northeast trending Appendix A. Supplementary data
axial traces of F1 upright basement folds as well as regional scale
basement faults that both formed during the Devonian are oriented Supplementary data related to this article can be found at http://
parallel to one of the dominant joint and fault sets in the overlying dx.doi.org/10.1016/j.jsg.2016.02.012.
Strzelecki Group rocks. However, a dominant Devonian fracture set
that is oriented orthogonal to F1 axial traces is absent in Cretaceous
References
cover rocks. This suggests that later tectonic processes have reac-
tivated only favourably oriented basement structures and that only Abbassi, M.R., Mancktelow, N.S., 1990. The effect of initial perturbation shape and
symmetry on fold development. J. Struct. Geol. 12, 273e282. http://dx.doi.org/ continental extension. In: Beaumont, C., Tankard, A.J. (Eds.), Sedimentary Basins
10.1016/0191-8141(90)90011-M. and Basin-Forming Mechanisms. Canadian Society of Petroleum Geologists,
Abellan, A., Oppikofer, T., Jaboyedoff, M., Rosser, N.J., Lim, M., Lato, M.J., 2014. pp. 147e162.
Terrestrial laser scanning of rock slope instabilities. Earth Surf. Process. Land- Fernandez, N., Kaus, B.J.P., 2014. Influence of pre-existing salt diapirs on 3D folding
forms 39, 80e97. http://dx.doi.org/10.1002/esp.3493. patterns. Tectonophysics 637, 354e369. http://dx.doi.org/10.1016/
Agisoft LLC, 2015. Agisoft Photoscan Professional [WWW Document]. URL. www. j.tecto.2014.10.021.
agisoft.com (accessed 16.07.15.). Fern andez, O., 2005. Obtaining a best fitting plane through 3D georeferenced data.
Allmendinger, R.W., Cardozo, N.C., Fisher, D., 2013. Structural Geology Algorithms: J. Struct. Geol. 27, 855e858. http://dx.doi.org/10.1016/j.jsg.2004.12.004.
Vectors & Tensors. Cambridge University Press, Cambridge, England. Frank, T., 2006. Advanced Visualization and Modeling of Tetrahedral Meshes. Th.
Allum, J.A.E., 1966. Photogeology and Regional Mapping. Pergamon Press. Doct. Inst. Natl. Polytech. Lorraine. Technische Universitaet Bergakademie
Anderson, E.M., 1905. The dynamics of faulting. Trans. Edinb. Geol. Soc. 8, 387e402. Freiberg.
Anderson, E.M., 1942. The Dynamics of Faulting and Dyke Formation with Appli- Frehner, M., 2014. 3D fold growth rates. Terra Nova 26, 417e424. http://dx.doi.org/
cations to Britain. Oliver and Boyd, Edinburgh. 10.1111/ter.12116.
ARANZ Geo, 2014. Leapfrog Mining [WWW Document]. URL. http://www. García-Selle s, D., Falivene, O., Arbues, P., Gratacos, O., Tavani, S., Mun~ oz, J.A., 2011.
leapfrog3d.com/products/leapfrog-mining (accessed 31.10.14.). Supervised identification and reconstruction of near-planar geological surfaces
Bemis, S., Micklethwaite, S., Turner, D., James, M.R., Akciz, S., Thiele, S., from terrestrial laser scanning. Comput. Geosci. 37, 1584e1594. http://
Bangash, H.A., 2014. Ground-based and UAV-based photogrammetry: a multi- dx.doi.org/10.1016/j.cageo.2011.03.007.
scale, high-resolution mapping tool for structural geology and paleoseismol- Gardner, T., Webb, J., Pezzia, C., Amborn, T., Tunnell, R., Flanagan, S., Merritts, D.,
ogy. J. Struct. Geol. 69, 163e178. http://dx.doi.org/10.1016/j.jsg.2014.10.007. Marshall, J., Fabel, D., Cupper, M.L., 2009. Episodic intraplate deformation of
Bierlein, F.P., Foster, D.A., Mcknight, S., Arne, D.C., 1999. Timing of gold minerali- stable continental margins: evidence from Late Neogene and Quaternary ma-
sation in the Ballarat goldfields, central Victoria: constraints from 40Ar/39Ar rine terraces, Cape Liptrap, Southeastern Australia. Quat. Sci. Rev. 28, 39e53.
results. Aust. J. Earth Sci. 46, 301e309. http://dx.doi.org/10.1046/j.1440- http://dx.doi.org/10.1016/j.quascirev.2008.10.004.
0952.1999.00708.x. Gigli, G., Morelli, S., Fornera, S., Casagli, N., 2014. Terrestrial laser scanner and
Blenkinsop, T.G., 2008. Relationships between faults, extension fractures and veins, geomechanical surveys for the rapid evaluation of rock fall susceptibility sce-
and stress. J. Struct. Geol. 30, 622e632. http://dx.doi.org/10.1016/ narios. Landslides 11, 1e14. http://dx.doi.org/10.1007/s10346-012-0374-0.
j.jsg.2008.01.008. Gillespie, P.A., Howard, C.B., Walsh, J.J., Watterson, J., 1993. Measurement and
Calcagno, P., Chiles, J.P., Courrioux, G., Guillen, A., 2008. Geological modelling from characterisation of spatial distributions of fractures. Tectonophysics 226,
field data and geological knowledge Part I. Modelling method coupling 3D 113e141. http://dx.doi.org/10.1016/0040-1951(93)90114-Y.
potential-field interpolation and geological rules. Phys. Earth Planet. Inter. 171, Girardeau-Montaut, D., Roux, M., Marc, R., Thibault, G., 2005. Change detection on
147e157. points cloud data acquired with a ground laser scanner. In: ISPRS Workshop e
Calcagno, P., Courrioux, G., Guillen, A., Fitzgerald, D., Mcinerney, P., 2006. How 3D Laser Scanning 2005 e Enschede, The Netherlands, September 12-14, 2005.
implicit geometric modelling helps to understand geology: the 3DGeoModeller Glen, R.A., 2013. Refining accretionary orogen models for the Tasmanides of eastern
methodology. In: Society for Mathematical Geology XIth International Australia. Aust. J. Earth Sci. 60, 315e370. http://dx.doi.org/10.1080/
Congress. 08120099.2013.772537.
Cardozo, N., Allmendinger, R.W., 2013. Spherical projections with OSXStereonet. Gomez, C., Hayakawa, Y., Obanawa, H., 2015. A study of Japanese landscapes using
Comput. Geosci. 51, 193e205. http://dx.doi.org/10.1016/j.cageo.2012.07.021. structure from motion derived DSMs and DEMs based on historical aerial
Carr, J.C., Beatson, R.K., McCallum, B.C., Fright, W.R., McLennan, T.J., Mitchell, T.J., photographs: new opportunities for vegetation monitoring and diachronic
1997. Smooth surface reconstruction from noisy range data. In: GRAPHITE '03 geomorphology. Geomorphology 242, 11e20. http://dx.doi.org/10.1016/
Proceedings of the 1st International Conference on Computer Graphics and j.geomorph.2015.02.021.
Interactive Techniques in Australasia and South East Asia. http://dx.doi.org/ Grasemann, B., Schmalholz, S.M., 2012. Lateral fold growth and fold linkage. Geol-
10.1145/604471.604495. ogy 40, 1039e1042. http://dx.doi.org/10.1130/G33613.1.
Caumon, G., Gray, G., Antoine, C., Titeux, M.-O., 2013. Three-dimensional implicit Gray, D.R., Foster, D.A., Korsch, R.J., Spaggiari, C.V., 2006. Structural style and crustal
stratigraphic model building from remote sensing data on tetrahedral meshes: architecture of the Tasmanides of eastern Australia, example of a composite
theory and application to a regional model of La Popa Basin, NE Mexico. IEEE accretionary orogen. Geol. Soc. Am. SP 2414, 1e33. http://dx.doi.org/10.1130/
Trans. Geosci. Remote Sens. 51, 1613e1621. http://dx.doi.org/10.1109/ 2006.2414(07).
TGRS.2012.2207727. Gray, D.R., Janssen, C., Vapnik, Y., 1999. Deformation character and palaeo-fluid flow
s, J.P., Aug, C., Guillen, A., Lees, T., 2004. Modelling the geometry of geological
Chile across a wrench fault within a Palaeozoic subduction-accretion system: War-
units and its uncertainty in 3D from structural data: the potential-field method. atah Fault Zone, southeastern Australia. J. Struct. Geol. 21, 191e196.
In: Orebody Modelling and Strategic Mine Planning. Perth, Western Australia, Harwin, S., Lucieer, A., 2012. Assessing the accuracy of georeferenced point clouds
pp. 355e362. produced via multi-view stereopsis from unmanned aerial vehicle (UAV) im-
Clapuyt, F., Vanacker, V., Van Oost, K., 2015. Reproducibility of UAV-based Earth agery. Remote Sens. 4, 1573e1599. http://dx.doi.org/10.3390/rs4061573.
topography reconstructions based on structure-from-motion algorithms. Geo- Healy, D., Sibson, R.H., Shipton, Z., Butler, R., 2012. Stress, faulting, fracturing and
morphology. http://dx.doi.org/10.1016/j.geomorph.2015.05.011 (in press). seismicity: the legacy of Ernest Masson Anderson. Geol. Soc. Lond. Spec. Publ.
Cobbold, P., 1975. Fold propagation in single embedded layers. Tectonophysics 27, 367, 1e6. http://dx.doi.org/10.1144/SP367.1.
333e351. Hill, K.C., Cooper, G.T., Richardson, M.J., Lavin, C.J., 1994. Structural Framework of the
Colwell, J.B., Willcox, J.B., 1993. Regional Structure of the Gippsland Basin: Inter- eastern Otway-Torquay Basins: inversion and interaction between to major
pretation and Mapping of Deep Seismic Data Set. Report. structural provinces. Geophysics 25, 79e87.
Cowan, E.J., Beatson, R.K., Ross, H.J., Fright, W.R., Mclennan, T.J., Evans, T.R., Carr, J.C., Hodgetts, D., 2013. Laser scanning and digital outcrop geology in the petroleum
Lane, R.G., Bright, D.V., Gillman, A.J., Oshust, P.A., Titley, M., 2003. Practical industry: a review. Mar. Pet. Geol. 46, 335e354. http://dx.doi.org/10.1016/
implicit geological modelling. In: Dominy, S. (Ed.), Fifth International Mining j.marpetgeo.2013.02.014.
Geology Conference Proceedings, AusIMM Publication Series No 8/2003, Hudleston, P.J., Treagus, S.H., 2010. Information from folds: a review. J. Struct. Geol.
pp. 89e99. 32, 2042e2071. http://dx.doi.org/10.1016/j.jsg.2010.08.011.
Cox, S., Etheridge, M., 1987. The role of fluids in syntectonic mass transport, and the Janssen, C., Laube, N., Bau, M., Gray, D.R., 1998. Fluid regime in faulting deformation
localization of metamorphic vein-type ore deposits. Ore Geol. Rev. 2, 65e86. of the Waratah Fault Zone, Australia, as inferred from major and minor element
Cox, S., Wall, V., Etheridge, M., 1991. Deformational and metamorphic processes in analyses and stable isotopic signatures. Tectonophysics 294, 109e130. http://
the formation of mesothermal vein-hosted gold deposits e examples from the dx.doi.org/10.1016/S0040-1951(98)00127-9.
Lachlan Fold Belt in central Victoria. Ore Geol. Rev. 6, 391e423. res, L., Kemp, E.De, Lindsay, M., Wellmann, F., Hillier, M., Laurent, G.,
Jessell, M., Aille
Cracknell, M.J., Roach, M., Green, D., Lucieer, A., 2013. Estimating bedding orienta- Carmichael, T., Martin, R., 2014. Next Generation Three-dimensional Geologic
tion from high-resolution digital elevation models. IEEE Trans. Geosci. Remote Modeling and Inversion, vol. 18. SEG Special Publication, pp. 261e272.
Sens. 51, 2949e2959. Johnson, K., Nissen, E., Saripalli, S., Arrowsmith, J.R., McGarey, P., Scharer, K.,
Donath, F.A., 1962. Role of layering in geologic deformation. Trans. N. Y. Acad. Sci. 24, Williams, P., Blisniuk, K., 2014. Rapid mapping of ultrafine fault zone topog-
236e249. http://dx.doi.org/10.1111/j.2164-0947.1962.tb00770.x. raphy with structure from motion. Geosphere 10, 969e986. http://dx.doi.org/
Dubey, A., Cobbold, P., 1977. Noncylindrical flexural slip folds in nature and exper- 10.1130/GES01017.1.
iment. Tectonophysics 38, 223e239. Jones, R.R., McCaffrey, K.J.W., Imber, J., Wightman, R., Smith, S.A.F., Holdsworth, R.E.,
Duddy, I.R., 2003a. Mesozoic: a time of change in tectonic regime. In: Birch, W.D. Clegg, P., De Paola, N., Healy, D., Wilson, R.W., 2008. Calibration and validation
(Ed.), Geology of Victoria. Geological Society of Australia (Victoria Division), of reservoir models: the importance of high resolution, quantitative outcrop
pp. 239e288. analogues. Geol. Soc. Lond. Spec. Publ. 309, 87e98.
Duddy, I.R., 2003b. Tertiary: a period of transition to marine basin environments. In: Jones, R.R., McCaffrey, K.J.W., Wilson, R.W., Holdsworth, R.E., 2004. Digital field data
Birch, W. (Ed.), Geology of Victoria. Geological Society of Australia (Victoria acquisition: towards increased quantification of uncertainty during geological
Division), Australia, pp. 289e335. mapping. Geol. Soc. Lond. Spec. Publ. 239, 43e56. http://dx.doi.org/10.1144/
EDF R&D Telecom Paristech, 2015. CloudCompare (Version 2.6.1) [GPL software] GSL.SP.2004.239.01.04.
[WWW Document]. URL. http://www.cloudcompare.org/ (accessed 07.06.15.). Jones, R.R., Pearce, M.A., Jacquemyn, C., Watson, F.E., 2016. Robust Best-fit Planes
Eisenbeiss, H., 2009. UAV Photogrammetry. PhD thesis. ETH Zurich. from Geospatial Data, 12, pp. 1e8. http://dx.doi.org/10.1130/GES01247.1.
Etheridge, M.A., Branson, J.C., Stuart-Smith, P.G., 1987. The Bass, Gippsland and Jordan, B.R., 2015. A bird's-eye view of geology: the use of micro drones/UAVs in
Otway Basins, southeast Australia: a branched rift system formed by geologic fieldwork and education. GSA Today 25, 50e52. http://dx.doi.org/
10.1130/GSATG232GW.1.Figure. Ramsay, J.G., 1967. Folding and Fracturing of Rocks. The Blackburn Press, p. 562.
Lane, R., Beckitt, G., Duffett, M., 2007. 3D geological mapping and potential field Schmid, D.W., Dabrowski, M., Krotkiewski, M., 2008. Evolution of large amplitude
modelling of west Arnhem land, northern territory, vol. 2007. ASEG Ext. Abstr., 3D fold patterns: a FEM study. Phys. Earth Planet. Inter. 171, 400e408. http://
p. 1. http://dx.doi.org/10.1071/ASEG2007ab072 dx.doi.org/10.1016/j.pepi.2008.08.007.
Leberl, F., Irschara, A., Pock, T., Meixner, P., Gruber, M., Scholz, S., Wiechert, A., 2010. Scholz, C.H., 2002. The Mechanics of Earthquakes and Faulting. Cambridge Uni-
Point clouds: lidar versus 3D vision. Photogramm. Eng. Remote Sens. 76, versity Press, Cambridge, UK, p. 471.
1123e1134. Seers, T., Hodgetts, D., 2013a. Comparison of digital outcrop and conventional data
Ledez, D., 2003. Mode lisation d'objets naturels par formulation implicites (PhD collection approaches for the characterisation of naturally fractured reservoir
thesis). Institut National Polytechnique de Lorraine, 198 pages. analogues. Geol. Soc. Lond. Spec. Publ. 374 http://dx.doi.org/10.1144/SP374.13.
Lennox, P.G., Golding, S.D., 1989. Quartz veining in simply folded arenites, Cape Seers, T., Hodgetts, D., 2013b. Characterisation of fracture systems using subsurface
Liptrap, southeast Victoria, Australia. Aust. J. Earth Sci. 36, 243e261. http:// digital outcrop models. In: Advances in the Study of Fractured Reservoirs.
dx.doi.org/10.1080/08120098908729484. Geological Society of London, London.
Lindner, A.W., 1953. The geology of the coastline of Waratah Bay between Wal- Seers, T.D., Hodgetts, D., 2016. Probabilistic constraints on structural lineament best
kerville and Cap Liptrap. In: Proceedings of the Royal Society of Victoria, vol. 64, fit plane precision obtained through numerical analysis. J. Struct. Geol. 82,
pp. 77e92. 37e47. http://dx.doi.org/10.1016/j.jsg.2015.11.004.
Lockner, D.A., Byerlee, J.D., Ponomarev, A., Sidorin, A., 1992. Observations of qua- Sibson, R., 1996. Structural permeability of fluid-driven fault-fracture meshes.
sistatic fault growth from acoustic emissions. In: Evans, B., Wong, T. (Eds.), Fault J. Struct. Geol. 18, 1031e1042.
Mechanics and Transport Properties of Rocks. Academic Press, London, Sibson, R.H., 2012. Reverse fault rupturing: competition between non-optimal and
pp. 3e31. optimal fault orientations. Geol. Soc. Lond. Spec. Publ. 367, 39e50. http://
Lowe, D.G., 1999. Object Recognition FromLocal Scale-invariant Features. IEEE Int. dx.doi.org/10.1144/SP367.4.
Conf. Comput. Vis. http://dx.doi.org/10.1109/ICCV.1999.790410 Smith, G.C., 1988. Oil and gas. In: Douglas, J.G., Ferguson, J.A. (Eds.), Geology of
Lowe, D.G., 2004. Distinctive image features from scale-invariant keypoints. Int. J. Victoria, vol. 5. Geological Society of Australia Special Publication, pp. 514e531.
Comput. Vis. 60, 91e110. Stearns, D.W., Friedman, M., 1972. Reservoirs in fractured rock. Mem. Am. Assoc. Pet.
Lucieer, A., Jong, S.M.D., Turner, D., 2014. Mapping landslide displacements using Geol. 16, 82e106. http://archives.datapages.com/data/specpubs/fieldst4/data/
Structure from Motion (SfM) and image correlation of multi-temporal UAV a010/a010/0001/0050/0082.htm.
photography. Prog. Phys. Geogr. 38, 97e116. http://dx.doi.org/10.1177/ € cker, C., Eltner, A., Karrasch, P., 2015. Measuring gullies by synergetic application
Sto
0309133313515293. of UAV and close range photogrammetry d a case study from Andalusia, Spain.
Mancktelow, N., 1999. Finite-element modelling of single-layer folding in elasto- Catena 132, 1e11. http://dx.doi.org/10.1016/j.catena.2015.04.004.
viscous materials: the effect of initial perturbation geometry. J. Struct. Geol. Strecha, C., Zoller, R., Rutishauser, S., Brot, B., Schneider-Zapp, K., Chovancova, V.,
21, 161e177. Krull, M., Glassey, L., 2015. Quality assessment of 3D reconstruction using
Maxelon, M., 2004. Developing a Three-dimensional Structural Model of the Lower fisheye and perspective sensors. ISPRS Ann. Photogramm. Remote Sens. Spat.
Lepontine Nappes (PhD thesis). ETH Zurich. http://dx.doi.org/10.3929/ethz-a- Inf. Sci. 215e222. http://dx.doi.org/10.5194/isprsannals-II-3-W4-215-2015. II-3/
004833149. W4.
McCaffrey, K.J.W., Hodgetts, D., Howell, J., Hunt, D., Imber, J., Jones, R.R., Tomasso, M., Talent, J.A., 1965. The stratigraphic and diastrophic evolution of central and eastern
Thurmond, J., Viseur, S., 2010. Virtual fieldtrips for petroleum geoscientists. Victoria in middle Palaeozoic tims. In: Proceedings of the Royal Society of
Geol. Soc. Lond. Pet. Geol. Conf. Ser. 7, 19e26. http://dx.doi.org/10.1144/ Victoria, vol. 79, pp. 179e195.
0070019. Thiele, S.T., Micklethwaite, S., Bourke, P., Verrall, M., Kovesi, P., 2015. Insights into
McCaffrey, K.J.W., Jones, R.R., Holdsworth, R.E., Wilson, R.W., Clegg, P., Imber, J., the mechanics of en-e chelon sigmoidal vein formation using ultra-high reso-
Holliman, N., Trinks, I., 2005. Unlocking the Spatial Dimension: Digital Tech- lution photogrammetry and computed tomography. J. Struct. Geol. 77, 27e44.
nologies and the Future of Geoscience Fieldwork, vol. 162, pp. 927e938. http:// http://dx.doi.org/10.1016/j.jsg.2015.05.006.
dx.doi.org/10.1144/0016-764905-017. Towse, D., 1980. A review of the mechanics and occurrence of natural fractures in
Mcinerney, P., Guillen, A., Courrioux, G., Calcagno, P., Lees, T., 2005. Building 3D rock as applied to th-e development of the tight Western gas sands. Lawrence
geological models directly from the data? a new approach applied to broken Livermore Lab, 366.
Hill, Australia. In: Digital Mapping Techniques 2005. Trinks, I., Clegg, P., McCaffrey, K., Jones, R., Hobbs, R., Holdsworth, B., Holliman, N.,
Micklethwaite, S., Turner, D., Vasuki, Y., Kovesi, P., Lucieer, A., 2012. Mapping from Imber, J., Waggott, S., Wilson, R., 2005. Mapping and analysing virtual outcrops.
an Armchair: rapid, high-resolution mapping using UAV and computer vision Vis. Geosci. 10, 13e19. http://dx.doi.org/10.1007/s10069-005-0026-9.
technology. In: Proceedings of Structural Geology and Resources 2012. Kal- Tveite, H., 2015. Line Direction Histogram e QGIS Plugin [WWW Document]. URL.
goorlie, Western Australia, pp. 130e133. http://arken.umb.no/~havatv/gis/qgisplugins/LineDirectionHistogram/
Mitra, S., 2002. Fold-accommodation faults. Am. Assoc. Pet. Geol. Bull. 86, 671e693. (accessed 7.11.15).
http://dx.doi.org/10.1306/61EEDB7A-173E-11D7-8645000102C1865D. Valenta, C.L., 2014. Multi-generational Fault Relationships and History, Cape
Müller, J., Ga €rtner-Roer, I., Thee, P., Ginzler, C., 2014. Accuracy assessment of Paterson to Inverloch, Western Gippsland Basin, Victoria. Honours thesis.
airborne photogrammetrically derived high-resolution digital elevation models Monash University, Clayton.
in a high mountain environment. ISPRS J. Photogramm. Remote Sens. 98, VandenBerg, A.H.M., Willman, C.E., Maher, S., Simons, B.A., Cayley, R.A., Taylor, D.H.,
58e69. http://dx.doi.org/10.1016/j.isprsjprs.2014.09.015. Morand, V.J., Moore, D.H., Radojkovic, A., 2000. The tasman fold belt system in
Newhouse, W.H., 1942. Ore Deposits as Related to Structural Features. Princetown Victoria. Geol. Surv. Vic. Spec. Publ. 463 pages.
University Press, Pirenceton, New Jersey, p. 280. Vasuki, Y., Holden, E.-J., Kovesi, P., Micklethwaite, S., 2014. Semi-automatic mapping
Niethammer, U., James, M.R., Rothmund, S., Travelletti, J., Joswig, M., 2012. UAV- of geological Structures using UAV-based photogrammetric data: an image
based remote sensing of the Super-Sauze landslide: evaluation and results. Eng. analysis approach. Comput. Geosci. 69, 22e32. http://dx.doi.org/10.1016/
Geol. 128, 2e11. http://dx.doi.org/10.1016/j.enggeo.2011.03.012. j.cageo.2014.04.012.
Nishar, A., Richards, S., Breen, D., Robertson, J., Breen, B., 2016. Thermal infrared Vollgger, S.A., Cruden, A.R., Aille res, L., Cowan, E.J., 2015. Regional dome evolution
imaging of geothermal environments and by an unmanned aerial vehicle and its control on ore-grade distribution: insights from 3D implicit modelling of
(UAV): a case study of the Wairakei e Tauhara geothermal field, Taupo, New the Navachab gold deposit, Namibia. Ore Geol. Rev. 69, 268e284.
Zealand. Renew. Energy 86, 1256e1264. http://dx.doi.org/10.1016/ Webb, J.A., 1991. Geological history of Victoria. In: Cochrane, G.W., Quick, G.W.,
j.renene.2015.09.042. Spencer-Jones, D. (Eds.), Introducing Victorian Geology. Victorian Division.
Norvick, M.S., Smith, M.A., Power, M.R., 2001. The plate tectonic evolution of eastern Geological Society of Australia, pp. 97e168.
Australasia guided by the stratigraphy of the Gippsland Basin. In: Hill, K.C., Wilcox, J.B., Colwell, J.B., Constantine, A.E., 1992. New ideas on Gippslan Basin
Bernecker, T. (Eds.), PESA Eastern Australasian Basins Symposium. PESA, Mel- regional tectonics. In: Barton, C.M., Hill, K., Abele, C., Foster, J., Kempton, N.
bourn, pp. 15e23. (Eds.), Energy, Economics and Environment: Gippsland Basin Symposium.
O'Connor, B., 1978. The Sedimentary and Tectonic Structures of the Lower Devonian Australian Institute of Mining & Metallurgy, pp. 93e100.
Liptrap Formation (PhD thesis). University of Melbourne, Victoria. Willcox, J.B., Colwell, J.B., Constantine, A.E., 1992. New ideas on Gippsland Basin
Paterson, M.S., Wong, T., 2005. Experimental Rock Deformation e the Brittle Field, regional tectonics. In: Proceedings to Gipplsan Basin Symposium, Melbourne,
second ed. Springer-Verlag Berlin Heidelberg, Berlin, p. 348. http://dx.doi.org/ 22-23 June 1992. The Australasian Institute of Mining and Metallurgy,
10.1007/b137431. pp. 93e110.
Perineck, D., Cockshell, C.D., Finlayson, D.M., Hill, K.A., 1994. The Otway Basin: early Willcox, J.B., Stagg, H., 1990. Australia's southern margin: a product of oblique
Cretaceous rifting to miocene strike-slip. In: Petroleum Exploration Society of extension. Tectonophysics 173, 269e281. http://dx.doi.org/10.1016/0040-
Australia, Otway Basin Symposium, pp. 27e33. 1951(90)90223-U.
Price, R.A., 1965. Flathead map area, British Columbia. Geol. Soc. Can. Mem. 336, 221. Woodcock, N.H., 1977. Specification of fabric shapes using an eigenvalue method.
QGIS Development team, 2015. QGIS Geographic Information System. Open Source Bull. Geol. Soc. Am. 88, 1231e1236. http://dx.doi.org/10.1130/0016-7606(1977)
Geospatial Foundation Project [WWW Document]. URL. http://qgis.osgeo.org 88<1231:SOFSUA>2.0.CO;2.
(accessed 07.06.15.).

Voll Gger 2016

Uploaded by

Copyright:

Available Formats

Voll Gger 2016

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Voll Gger 2016

Uploaded by

Copyright:

Available Formats

Journal of Structural Geology 85 (2016) 168e187

Contents lists available at ScienceDirect

Journal of Structural Geology

1. Introduction Extension fractures can be precursors to faults (Blenkinsop,

OLforward is the percentage of the desired image overlap in the ﬂight

(small aperture) to ensure a large depth of ﬁeld. Lenses with ﬁxed

4. Acquisition of aerial images and ground control points

Prior to the acquisition of aerial photographs by UAV, an A4

The nadir aerial photographs of the Fold Stack locality acquired

Finally, the point cloud, orthophotograph and digital elevation orthophotograph.

orthophotograph, hence smaller-scale features were not mapped in

5.3. Computation of a 3D implicit structural trend model

3D implicit modelling interpolates spatial data by computing a

6. Results and interpretation

6.1. D1-related folding and faulting

Eastewest shortening related to the Devonian Tabberabberan

6.4. Cape Paterson fracture populations

Only three main fracture set orientations were mapped within

You might also like